Compare commits

..

4 Commits

Author SHA1 Message Date
copilot-swe-agent[bot]
316b4144d3 Initial plan 2026-06-22 22:23:18 +00:00
Owen Mansel-Chan
f0576046b1 Merge pull request #22027 from owen-mc/go/improve-tests
Go: Improve two tests
2026-06-22 17:19:40 +01:00
Owen Mansel-Chan
07cf89568f Test CFG for function epilogue (read-result nodes and calls to defered functions) 2026-06-20 22:04:45 +01:00
Owen Mansel-Chan
42ebe56023 Make all lines in logging tests reachable 2026-06-20 22:04:43 +01:00
328 changed files with 6860 additions and 11063 deletions

View File

@@ -13,7 +13,7 @@ func logSomething(entry *logrus.Entry) {
entry.Traceln(text) // $ logger=text entry.Traceln(text) // $ logger=text
} }
func logrusCalls() { func logrusCalls(selector int) {
err := errors.New("Error") err := errors.New("Error")
var fields logrus.Fields = nil var fields logrus.Fields = nil
var fn logrus.LogFunction = nil var fn logrus.LogFunction = nil
@@ -28,10 +28,14 @@ func logrusCalls() {
logSomething(tmp) logSomething(tmp)
logrus.Error(text) // $ logger=text logrus.Error(text) // $ logger=text
logrus.Fatalf(fmt, text) // $ logger=fmt logger=text
logrus.Panicln(text) // $ logger=text
logrus.Infof(fmt, text) // $ logger=fmt logger=text logrus.Infof(fmt, text) // $ logger=fmt logger=text
if selector == 0 {
logrus.Fatalf(fmt, text) // $ logger=fmt logger=text
} else if selector == 1 {
logrus.Panicln(text) // $ logger=text
} else if selector == 2 {
logrus.FatalFn(fn) // $ logger=fn logrus.FatalFn(fn) // $ logger=fn
}
// components corresponding to the format specifier "%T" are not considered vulnerable // components corresponding to the format specifier "%T" are not considered vulnerable
logrus.Infof("%s: found type %T", text, v) // $ logger="%s: found type %T" logger=text type-logger=v logrus.Infof("%s: found type %T", text, v) // $ logger="%s: found type %T" logger=text type-logger=v

View File

@@ -8,6 +8,6 @@ var v []byte
func main() { func main() {
glogTest(len(v)) glogTest(len(v))
stdlib() stdlib(len(v))
slogTest() slogTest()
} }

View File

@@ -4,37 +4,69 @@ import (
"log" "log"
) )
func stdlib() { func stdlib(selector int) {
var logger log.Logger var logger log.Logger
logger.SetPrefix("prefix: ") logger.SetPrefix("prefix: ")
switch selector {
case 0:
logger.Fatal(text) // $ logger=text logger.Fatal(text) // $ logger=text
case 1:
logger.Fatalf(fmt, text) // $ logger=fmt logger=text logger.Fatalf(fmt, text) // $ logger=fmt logger=text
case 2:
logger.Fatalln(text) // $ logger=text logger.Fatalln(text) // $ logger=text
case 3:
logger.Panic(text) // $ logger=text logger.Panic(text) // $ logger=text
case 4:
logger.Panicf(fmt, text) // $ logger=fmt logger=text logger.Panicf(fmt, text) // $ logger=fmt logger=text
case 5:
logger.Panicln(text) // $ logger=text logger.Panicln(text) // $ logger=text
case 6:
logger.Print(text) // $ logger=text logger.Print(text) // $ logger=text
case 7:
logger.Printf(fmt, text) // $ logger=fmt logger=text logger.Printf(fmt, text) // $ logger=fmt logger=text
case 8:
logger.Println(text) // $ logger=text logger.Println(text) // $ logger=text
}
// components corresponding to the format specifier "%T" are not considered vulnerable // components corresponding to the format specifier "%T" are not considered vulnerable
switch selector {
case 9:
logger.Fatalf("%s: found type %T", text, v) // $ logger="%s: found type %T" logger=text type-logger=v logger.Fatalf("%s: found type %T", text, v) // $ logger="%s: found type %T" logger=text type-logger=v
case 10:
logger.Panicf("%s: found type %T", text, v) // $ logger="%s: found type %T" logger=text type-logger=v logger.Panicf("%s: found type %T", text, v) // $ logger="%s: found type %T" logger=text type-logger=v
case 11:
logger.Printf("%s: found type %T", text, v) // $ logger="%s: found type %T" logger=text type-logger=v logger.Printf("%s: found type %T", text, v) // $ logger="%s: found type %T" logger=text type-logger=v
}
log.SetPrefix("prefix: ") log.SetPrefix("prefix: ")
switch selector {
case 12:
log.Fatal(text) // $ logger=text log.Fatal(text) // $ logger=text
case 13:
log.Fatalf(fmt, text) // $ logger=fmt logger=text log.Fatalf(fmt, text) // $ logger=fmt logger=text
case 14:
log.Fatalln(text) // $ logger=text log.Fatalln(text) // $ logger=text
case 15:
log.Panic(text) // $ logger=text log.Panic(text) // $ logger=text
case 16:
log.Panicf(fmt, text) // $ logger=fmt logger=text log.Panicf(fmt, text) // $ logger=fmt logger=text
case 17:
log.Panicln(text) // $ logger=text log.Panicln(text) // $ logger=text
case 18:
log.Print(text) // $ logger=text log.Print(text) // $ logger=text
case 19:
log.Printf(fmt, text) // $ logger=fmt logger=text log.Printf(fmt, text) // $ logger=fmt logger=text
case 20:
log.Println(text) // $ logger=text log.Println(text) // $ logger=text
}
// components corresponding to the format specifier "%T" are not considered vulnerable // components corresponding to the format specifier "%T" are not considered vulnerable
switch selector {
case 21:
log.Fatalf("%s: found type %T", text, v) // $ logger="%s: found type %T" logger=text type-logger=v log.Fatalf("%s: found type %T", text, v) // $ logger="%s: found type %T" logger=text type-logger=v
case 22:
log.Panicf("%s: found type %T", text, v) // $ logger="%s: found type %T" logger=text type-logger=v log.Panicf("%s: found type %T", text, v) // $ logger="%s: found type %T" logger=text type-logger=v
case 23:
log.Printf("%s: found type %T", text, v) // $ logger="%s: found type %T" logger=text type-logger=v log.Printf("%s: found type %T", text, v) // $ logger="%s: found type %T" logger=text type-logger=v
} }
}

View File

@@ -34,6 +34,265 @@
| DuplicateSwitchCase.go:16:1:16:14 | function declaration | DuplicateSwitchCase.go:0:0:0:0 | exit | | DuplicateSwitchCase.go:16:1:16:14 | function declaration | DuplicateSwitchCase.go:0:0:0:0 | exit |
| DuplicateSwitchCase.go:16:6:16:9 | skip | DuplicateSwitchCase.go:16:1:16:14 | function declaration | | DuplicateSwitchCase.go:16:6:16:9 | skip | DuplicateSwitchCase.go:16:1:16:14 | function declaration |
| DuplicateSwitchCase.go:16:13:16:14 | skip | DuplicateSwitchCase.go:16:1:16:14 | exit | | DuplicateSwitchCase.go:16:13:16:14 | skip | DuplicateSwitchCase.go:16:1:16:14 | exit |
| epilogues.go:0:0:0:0 | entry | epilogues.go:3:1:3:12 | skip |
| epilogues.go:3:1:3:12 | skip | epilogues.go:8:1:10:1 | skip |
| epilogues.go:8:1:10:1 | skip | epilogues.go:12:21:12:23 | skip |
| epilogues.go:12:1:14:1 | entry | epilogues.go:12:7:12:7 | argument corresponding to l |
| epilogues.go:12:1:14:1 | function declaration | epilogues.go:16:20:16:27 | skip |
| epilogues.go:12:7:12:7 | argument corresponding to l | epilogues.go:12:7:12:7 | initialization of l |
| epilogues.go:12:7:12:7 | initialization of l | epilogues.go:12:25:12:27 | argument corresponding to msg |
| epilogues.go:12:21:12:23 | skip | epilogues.go:12:1:14:1 | function declaration |
| epilogues.go:12:25:12:27 | argument corresponding to msg | epilogues.go:12:25:12:27 | initialization of msg |
| epilogues.go:12:25:12:27 | initialization of msg | epilogues.go:12:37:12:40 | argument corresponding to code |
| epilogues.go:12:37:12:40 | argument corresponding to code | epilogues.go:12:37:12:40 | initialization of code |
| epilogues.go:12:37:12:40 | initialization of code | epilogues.go:13:2:13:12 | selection of Println |
| epilogues.go:13:2:13:12 | selection of Println | epilogues.go:13:14:13:14 | l |
| epilogues.go:13:2:13:33 | call to Println | epilogues.go:12:1:14:1 | exit |
| epilogues.go:13:14:13:14 | implicit dereference | epilogues.go:12:1:14:1 | exit |
| epilogues.go:13:14:13:14 | implicit dereference | epilogues.go:13:14:13:21 | selection of prefix |
| epilogues.go:13:14:13:14 | l | epilogues.go:13:14:13:14 | implicit dereference |
| epilogues.go:13:14:13:21 | selection of prefix | epilogues.go:13:24:13:26 | msg |
| epilogues.go:13:24:13:26 | msg | epilogues.go:13:29:13:32 | code |
| epilogues.go:13:29:13:32 | code | epilogues.go:13:2:13:33 | call to Println |
| epilogues.go:16:1:18:1 | entry | epilogues.go:16:7:16:7 | argument corresponding to l |
| epilogues.go:16:1:18:1 | function declaration | epilogues.go:23:6:23:15 | skip |
| epilogues.go:16:7:16:7 | argument corresponding to l | epilogues.go:16:7:16:7 | initialization of l |
| epilogues.go:16:7:16:7 | initialization of l | epilogues.go:16:29:16:31 | argument corresponding to msg |
| epilogues.go:16:20:16:27 | skip | epilogues.go:16:1:18:1 | function declaration |
| epilogues.go:16:29:16:31 | argument corresponding to msg | epilogues.go:16:29:16:31 | initialization of msg |
| epilogues.go:16:29:16:31 | initialization of msg | epilogues.go:17:2:17:12 | selection of Println |
| epilogues.go:17:2:17:12 | selection of Println | epilogues.go:17:14:17:14 | l |
| epilogues.go:17:2:17:27 | call to Println | epilogues.go:16:1:18:1 | exit |
| epilogues.go:17:14:17:14 | l | epilogues.go:17:14:17:21 | selection of prefix |
| epilogues.go:17:14:17:21 | selection of prefix | epilogues.go:17:24:17:26 | msg |
| epilogues.go:17:24:17:26 | msg | epilogues.go:17:2:17:27 | call to Println |
| epilogues.go:23:1:27:1 | entry | epilogues.go:24:5:24:5 | skip |
| epilogues.go:23:1:27:1 | function declaration | epilogues.go:31:6:31:13 | skip |
| epilogues.go:23:6:23:15 | skip | epilogues.go:23:1:27:1 | function declaration |
| epilogues.go:24:5:24:5 | assignment to r | epilogues.go:24:21:24:21 | r |
| epilogues.go:24:5:24:5 | skip | epilogues.go:24:10:24:16 | recover |
| epilogues.go:24:10:24:16 | recover | epilogues.go:24:10:24:18 | call to recover |
| epilogues.go:24:10:24:18 | call to recover | epilogues.go:24:5:24:5 | assignment to r |
| epilogues.go:24:21:24:21 | r | epilogues.go:24:26:24:28 | nil |
| epilogues.go:24:21:24:28 | ...!=... | epilogues.go:23:1:27:1 | exit |
| epilogues.go:24:21:24:28 | ...!=... | epilogues.go:24:21:24:28 | ...!=... is false |
| epilogues.go:24:21:24:28 | ...!=... | epilogues.go:24:21:24:28 | ...!=... is true |
| epilogues.go:24:21:24:28 | ...!=... is false | epilogues.go:23:1:27:1 | exit |
| epilogues.go:24:21:24:28 | ...!=... is true | epilogues.go:25:3:25:13 | selection of Println |
| epilogues.go:24:26:24:28 | nil | epilogues.go:24:21:24:28 | ...!=... |
| epilogues.go:25:3:25:13 | selection of Println | epilogues.go:25:15:25:26 | "recovered:" |
| epilogues.go:25:3:25:30 | call to Println | epilogues.go:23:1:27:1 | exit |
| epilogues.go:25:15:25:26 | "recovered:" | epilogues.go:25:29:25:29 | r |
| epilogues.go:25:29:25:29 | r | epilogues.go:25:3:25:30 | call to Println |
| epilogues.go:31:1:33:1 | entry | epilogues.go:31:15:31:15 | argument corresponding to x |
| epilogues.go:31:1:33:1 | function declaration | epilogues.go:36:6:36:12 | skip |
| epilogues.go:31:6:31:13 | skip | epilogues.go:31:1:33:1 | function declaration |
| epilogues.go:31:15:31:15 | argument corresponding to x | epilogues.go:31:15:31:15 | initialization of x |
| epilogues.go:31:15:31:15 | initialization of x | epilogues.go:32:9:32:9 | x |
| epilogues.go:32:2:32:13 | return statement | epilogues.go:31:1:33:1 | exit |
| epilogues.go:32:9:32:9 | x | epilogues.go:32:13:32:13 | 2 |
| epilogues.go:32:9:32:13 | ...*... | epilogues.go:32:2:32:13 | return statement |
| epilogues.go:32:13:32:13 | 2 | epilogues.go:32:9:32:13 | ...*... |
| epilogues.go:36:1:38:1 | entry | epilogues.go:37:2:37:12 | selection of Println |
| epilogues.go:36:1:38:1 | function declaration | epilogues.go:42:6:42:18 | skip |
| epilogues.go:36:6:36:12 | skip | epilogues.go:36:1:38:1 | function declaration |
| epilogues.go:37:2:37:12 | selection of Println | epilogues.go:37:14:37:19 | "void" |
| epilogues.go:37:2:37:20 | call to Println | epilogues.go:36:1:38:1 | exit |
| epilogues.go:37:14:37:19 | "void" | epilogues.go:37:2:37:20 | call to Println |
| epilogues.go:42:1:48:1 | entry | epilogues.go:42:20:42:20 | argument corresponding to x |
| epilogues.go:42:1:48:1 | function declaration | epilogues.go:51:6:51:21 | skip |
| epilogues.go:42:6:42:18 | skip | epilogues.go:42:1:48:1 | function declaration |
| epilogues.go:42:20:42:20 | argument corresponding to x | epilogues.go:42:20:42:20 | initialization of x |
| epilogues.go:42:20:42:20 | initialization of x | epilogues.go:42:28:42:33 | zero value for result |
| epilogues.go:42:28:42:33 | implicit read of result | epilogues.go:42:40:42:42 | implicit read of err |
| epilogues.go:42:28:42:33 | initialization of result | epilogues.go:42:40:42:42 | zero value for err |
| epilogues.go:42:28:42:33 | zero value for result | epilogues.go:42:28:42:33 | initialization of result |
| epilogues.go:42:40:42:42 | implicit read of err | epilogues.go:42:1:48:1 | exit |
| epilogues.go:42:40:42:42 | initialization of err | epilogues.go:43:5:43:5 | x |
| epilogues.go:42:40:42:42 | zero value for err | epilogues.go:42:40:42:42 | initialization of err |
| epilogues.go:43:5:43:5 | x | epilogues.go:43:9:43:9 | 0 |
| epilogues.go:43:5:43:9 | ...<... | epilogues.go:43:5:43:9 | ...<... is false |
| epilogues.go:43:5:43:9 | ...<... | epilogues.go:43:5:43:9 | ...<... is true |
| epilogues.go:43:5:43:9 | ...<... is false | epilogues.go:47:9:47:9 | x |
| epilogues.go:43:5:43:9 | ...<... is true | epilogues.go:44:3:44:8 | skip |
| epilogues.go:43:9:43:9 | 0 | epilogues.go:43:5:43:9 | ...<... |
| epilogues.go:44:3:44:8 | assignment to result | epilogues.go:45:3:45:8 | return statement |
| epilogues.go:44:3:44:8 | skip | epilogues.go:44:13:44:13 | x |
| epilogues.go:44:12:44:13 | -... | epilogues.go:44:3:44:8 | assignment to result |
| epilogues.go:44:13:44:13 | x | epilogues.go:44:12:44:13 | -... |
| epilogues.go:45:3:45:8 | return statement | epilogues.go:42:28:42:33 | implicit read of result |
| epilogues.go:47:2:47:14 | return statement | epilogues.go:42:28:42:33 | implicit read of result |
| epilogues.go:47:9:47:9 | implicit write of result | epilogues.go:47:12:47:14 | nil |
| epilogues.go:47:9:47:9 | x | epilogues.go:47:9:47:9 | implicit write of result |
| epilogues.go:47:12:47:14 | implicit write of err | epilogues.go:47:2:47:14 | return statement |
| epilogues.go:47:12:47:14 | nil | epilogues.go:47:12:47:14 | implicit write of err |
| epilogues.go:51:1:54:1 | entry | epilogues.go:51:23:51:23 | argument corresponding to x |
| epilogues.go:51:1:54:1 | function declaration | epilogues.go:59:6:59:25 | skip |
| epilogues.go:51:6:51:21 | skip | epilogues.go:51:1:54:1 | function declaration |
| epilogues.go:51:23:51:23 | argument corresponding to x | epilogues.go:51:23:51:23 | initialization of x |
| epilogues.go:51:23:51:23 | initialization of x | epilogues.go:51:31:51:31 | zero value for n |
| epilogues.go:51:31:51:31 | implicit read of n | epilogues.go:51:1:54:1 | exit |
| epilogues.go:51:31:51:31 | initialization of n | epilogues.go:52:2:52:2 | skip |
| epilogues.go:51:31:51:31 | zero value for n | epilogues.go:51:31:51:31 | initialization of n |
| epilogues.go:52:2:52:2 | assignment to n | epilogues.go:53:2:53:7 | return statement |
| epilogues.go:52:2:52:2 | skip | epilogues.go:52:6:52:6 | x |
| epilogues.go:52:6:52:6 | x | epilogues.go:52:10:52:10 | 1 |
| epilogues.go:52:6:52:10 | ...+... | epilogues.go:52:2:52:2 | assignment to n |
| epilogues.go:52:10:52:10 | 1 | epilogues.go:52:6:52:10 | ...+... |
| epilogues.go:53:2:53:7 | return statement | epilogues.go:51:31:51:31 | implicit read of n |
| epilogues.go:59:1:62:1 | entry | epilogues.go:59:27:59:27 | argument corresponding to l |
| epilogues.go:59:1:62:1 | function declaration | epilogues.go:66:6:66:26 | skip |
| epilogues.go:59:6:59:25 | skip | epilogues.go:59:1:62:1 | function declaration |
| epilogues.go:59:27:59:27 | argument corresponding to l | epilogues.go:59:27:59:27 | initialization of l |
| epilogues.go:59:27:59:27 | initialization of l | epilogues.go:59:41:59:45 | argument corresponding to items |
| epilogues.go:59:41:59:45 | argument corresponding to items | epilogues.go:59:41:59:45 | initialization of items |
| epilogues.go:59:41:59:45 | initialization of items | epilogues.go:60:8:60:8 | l |
| epilogues.go:60:2:60:33 | defer statement | epilogues.go:61:2:61:12 | selection of Println |
| epilogues.go:60:8:60:8 | l | epilogues.go:60:8:60:12 | selection of log |
| epilogues.go:60:8:60:12 | selection of log | epilogues.go:60:14:60:20 | "count" |
| epilogues.go:60:8:60:33 | call to log | epilogues.go:59:1:62:1 | exit |
| epilogues.go:60:14:60:20 | "count" | epilogues.go:60:23:60:25 | len |
| epilogues.go:60:23:60:25 | len | epilogues.go:60:27:60:31 | items |
| epilogues.go:60:23:60:32 | call to len | epilogues.go:60:2:60:33 | defer statement |
| epilogues.go:60:27:60:31 | items | epilogues.go:60:23:60:32 | call to len |
| epilogues.go:61:2:61:12 | selection of Println | epilogues.go:61:14:61:25 | "processing" |
| epilogues.go:61:2:61:38 | call to Println | epilogues.go:60:8:60:33 | call to log |
| epilogues.go:61:14:61:25 | "processing" | epilogues.go:61:28:61:30 | len |
| epilogues.go:61:28:61:30 | len | epilogues.go:61:32:61:36 | items |
| epilogues.go:61:28:61:37 | call to len | epilogues.go:61:2:61:38 | call to Println |
| epilogues.go:61:32:61:36 | items | epilogues.go:61:28:61:37 | call to len |
| epilogues.go:66:1:71:1 | entry | epilogues.go:66:28:66:33 | argument corresponding to prefix |
| epilogues.go:66:1:71:1 | function declaration | epilogues.go:77:6:77:20 | skip |
| epilogues.go:66:6:66:26 | skip | epilogues.go:66:1:71:1 | function declaration |
| epilogues.go:66:28:66:33 | argument corresponding to prefix | epilogues.go:66:28:66:33 | initialization of prefix |
| epilogues.go:66:28:66:33 | initialization of prefix | epilogues.go:67:2:67:2 | skip |
| epilogues.go:67:2:67:2 | assignment to l | epilogues.go:68:8:68:8 | l |
| epilogues.go:67:2:67:2 | skip | epilogues.go:67:7:67:31 | struct literal |
| epilogues.go:67:7:67:31 | struct literal | epilogues.go:67:25:67:30 | prefix |
| epilogues.go:67:17:67:30 | init of key-value pair | epilogues.go:67:2:67:2 | assignment to l |
| epilogues.go:67:25:67:30 | prefix | epilogues.go:67:17:67:30 | init of key-value pair |
| epilogues.go:68:2:68:24 | defer statement | epilogues.go:69:10:69:10 | l |
| epilogues.go:68:8:68:8 | l | epilogues.go:68:8:68:17 | selection of logValue |
| epilogues.go:68:8:68:17 | selection of logValue | epilogues.go:68:19:68:23 | "bye" |
| epilogues.go:68:8:68:24 | call to logValue | epilogues.go:66:1:71:1 | exit |
| epilogues.go:68:19:68:23 | "bye" | epilogues.go:68:2:68:24 | defer statement |
| epilogues.go:69:2:69:25 | defer statement | epilogues.go:70:2:70:12 | selection of Println |
| epilogues.go:69:8:69:15 | selection of log | epilogues.go:69:17:69:21 | "ptr" |
| epilogues.go:69:8:69:25 | call to log | epilogues.go:68:8:68:24 | call to logValue |
| epilogues.go:69:9:69:10 | &... | epilogues.go:69:8:69:15 | selection of log |
| epilogues.go:69:10:69:10 | l | epilogues.go:69:9:69:10 | &... |
| epilogues.go:69:17:69:21 | "ptr" | epilogues.go:69:24:69:24 | 7 |
| epilogues.go:69:24:69:24 | 7 | epilogues.go:69:2:69:25 | defer statement |
| epilogues.go:70:2:70:12 | selection of Println | epilogues.go:70:14:70:19 | "body" |
| epilogues.go:70:2:70:20 | call to Println | epilogues.go:69:8:69:25 | call to log |
| epilogues.go:70:14:70:19 | "body" | epilogues.go:70:2:70:20 | call to Println |
| epilogues.go:77:1:82:1 | entry | epilogues.go:77:22:77:22 | argument corresponding to x |
| epilogues.go:77:1:82:1 | function declaration | epilogues.go:87:6:87:20 | skip |
| epilogues.go:77:6:77:20 | skip | epilogues.go:77:1:82:1 | function declaration |
| epilogues.go:77:22:77:22 | argument corresponding to x | epilogues.go:77:22:77:22 | initialization of x |
| epilogues.go:77:22:77:22 | initialization of x | epilogues.go:78:8:80:2 | function literal |
| epilogues.go:78:2:80:15 | defer statement | epilogues.go:81:2:81:12 | selection of Println |
| epilogues.go:78:8:80:2 | entry | epilogues.go:78:13:78:17 | argument corresponding to label |
| epilogues.go:78:8:80:2 | function literal | epilogues.go:80:4:80:9 | "done" |
| epilogues.go:78:8:80:15 | function call | epilogues.go:77:1:82:1 | exit |
| epilogues.go:78:13:78:17 | argument corresponding to label | epilogues.go:78:13:78:17 | initialization of label |
| epilogues.go:78:13:78:17 | initialization of label | epilogues.go:78:27:78:27 | argument corresponding to n |
| epilogues.go:78:27:78:27 | argument corresponding to n | epilogues.go:78:27:78:27 | initialization of n |
| epilogues.go:78:27:78:27 | initialization of n | epilogues.go:79:3:79:13 | selection of Println |
| epilogues.go:79:3:79:13 | selection of Println | epilogues.go:79:15:79:19 | label |
| epilogues.go:79:3:79:23 | call to Println | epilogues.go:78:8:80:2 | exit |
| epilogues.go:79:15:79:19 | label | epilogues.go:79:22:79:22 | n |
| epilogues.go:79:22:79:22 | n | epilogues.go:79:3:79:23 | call to Println |
| epilogues.go:80:4:80:9 | "done" | epilogues.go:80:12:80:12 | x |
| epilogues.go:80:12:80:12 | x | epilogues.go:80:14:80:14 | 1 |
| epilogues.go:80:12:80:14 | ...+... | epilogues.go:78:2:80:15 | defer statement |
| epilogues.go:80:14:80:14 | 1 | epilogues.go:80:12:80:14 | ...+... |
| epilogues.go:81:2:81:12 | selection of Println | epilogues.go:81:14:81:19 | "body" |
| epilogues.go:81:2:81:23 | call to Println | epilogues.go:78:8:80:15 | function call |
| epilogues.go:81:14:81:19 | "body" | epilogues.go:81:22:81:22 | x |
| epilogues.go:81:22:81:22 | x | epilogues.go:81:2:81:23 | call to Println |
| epilogues.go:87:1:98:1 | entry | epilogues.go:87:22:87:22 | argument corresponding to x |
| epilogues.go:87:1:98:1 | function declaration | epilogues.go:102:6:102:24 | skip |
| epilogues.go:87:6:87:20 | skip | epilogues.go:87:1:98:1 | function declaration |
| epilogues.go:87:22:87:22 | argument corresponding to x | epilogues.go:87:22:87:22 | initialization of x |
| epilogues.go:87:22:87:22 | initialization of x | epilogues.go:87:30:87:35 | zero value for result |
| epilogues.go:87:30:87:35 | implicit read of result | epilogues.go:87:1:98:1 | exit |
| epilogues.go:87:30:87:35 | initialization of result | epilogues.go:88:8:92:2 | function literal |
| epilogues.go:87:30:87:35 | zero value for result | epilogues.go:87:30:87:35 | initialization of result |
| epilogues.go:88:2:92:4 | defer statement | epilogues.go:93:5:93:5 | x |
| epilogues.go:88:8:92:2 | entry | epilogues.go:89:6:89:6 | skip |
| epilogues.go:88:8:92:2 | function literal | epilogues.go:88:2:92:4 | defer statement |
| epilogues.go:88:8:92:4 | function call | epilogues.go:87:1:98:1 | exit |
| epilogues.go:88:8:92:4 | function call | epilogues.go:87:30:87:35 | implicit read of result |
| epilogues.go:89:6:89:6 | assignment to r | epilogues.go:89:22:89:22 | r |
| epilogues.go:89:6:89:6 | skip | epilogues.go:89:11:89:17 | recover |
| epilogues.go:89:11:89:17 | recover | epilogues.go:89:11:89:19 | call to recover |
| epilogues.go:89:11:89:19 | call to recover | epilogues.go:89:6:89:6 | assignment to r |
| epilogues.go:89:22:89:22 | r | epilogues.go:89:27:89:29 | nil |
| epilogues.go:89:22:89:29 | ...!=... | epilogues.go:88:8:92:2 | exit |
| epilogues.go:89:22:89:29 | ...!=... | epilogues.go:89:22:89:29 | ...!=... is false |
| epilogues.go:89:22:89:29 | ...!=... | epilogues.go:89:22:89:29 | ...!=... is true |
| epilogues.go:89:22:89:29 | ...!=... is false | epilogues.go:88:8:92:2 | exit |
| epilogues.go:89:22:89:29 | ...!=... is true | epilogues.go:90:4:90:9 | skip |
| epilogues.go:89:27:89:29 | nil | epilogues.go:89:22:89:29 | ...!=... |
| epilogues.go:90:4:90:9 | assignment to result | epilogues.go:88:8:92:2 | exit |
| epilogues.go:90:4:90:9 | skip | epilogues.go:90:13:90:14 | -... |
| epilogues.go:90:13:90:14 | -... | epilogues.go:90:4:90:9 | assignment to result |
| epilogues.go:93:5:93:5 | x | epilogues.go:93:9:93:9 | 0 |
| epilogues.go:93:5:93:9 | ...<... | epilogues.go:93:5:93:9 | ...<... is false |
| epilogues.go:93:5:93:9 | ...<... | epilogues.go:93:5:93:9 | ...<... is true |
| epilogues.go:93:5:93:9 | ...<... is false | epilogues.go:96:2:96:7 | skip |
| epilogues.go:93:5:93:9 | ...<... is true | epilogues.go:94:3:94:7 | panic |
| epilogues.go:93:9:93:9 | 0 | epilogues.go:93:5:93:9 | ...<... |
| epilogues.go:94:3:94:7 | panic | epilogues.go:94:9:94:13 | "neg" |
| epilogues.go:94:3:94:14 | call to panic | epilogues.go:88:8:92:4 | function call |
| epilogues.go:94:9:94:13 | "neg" | epilogues.go:94:3:94:14 | call to panic |
| epilogues.go:96:2:96:7 | assignment to result | epilogues.go:97:9:97:14 | result |
| epilogues.go:96:2:96:7 | skip | epilogues.go:96:11:96:11 | x |
| epilogues.go:96:11:96:11 | x | epilogues.go:96:15:96:15 | x |
| epilogues.go:96:11:96:15 | ...*... | epilogues.go:96:2:96:7 | assignment to result |
| epilogues.go:96:15:96:15 | x | epilogues.go:96:11:96:15 | ...*... |
| epilogues.go:97:2:97:14 | return statement | epilogues.go:88:8:92:4 | function call |
| epilogues.go:97:9:97:14 | implicit write of result | epilogues.go:97:2:97:14 | return statement |
| epilogues.go:97:9:97:14 | result | epilogues.go:97:9:97:14 | implicit write of result |
| epilogues.go:102:1:110:1 | entry | epilogues.go:102:26:102:26 | argument corresponding to x |
| epilogues.go:102:1:110:1 | function declaration | epilogues.go:115:6:115:22 | skip |
| epilogues.go:102:6:102:24 | skip | epilogues.go:102:1:110:1 | function declaration |
| epilogues.go:102:26:102:26 | argument corresponding to x | epilogues.go:102:26:102:26 | initialization of x |
| epilogues.go:102:26:102:26 | initialization of x | epilogues.go:102:34:102:35 | zero value for ok |
| epilogues.go:102:34:102:35 | implicit read of ok | epilogues.go:102:43:102:43 | implicit read of n |
| epilogues.go:102:34:102:35 | initialization of ok | epilogues.go:102:43:102:43 | zero value for n |
| epilogues.go:102:34:102:35 | zero value for ok | epilogues.go:102:34:102:35 | initialization of ok |
| epilogues.go:102:43:102:43 | implicit read of n | epilogues.go:102:1:110:1 | exit |
| epilogues.go:102:43:102:43 | initialization of n | epilogues.go:103:8:103:17 | epiRecover |
| epilogues.go:102:43:102:43 | zero value for n | epilogues.go:102:43:102:43 | initialization of n |
| epilogues.go:103:2:103:19 | defer statement | epilogues.go:104:5:104:5 | x |
| epilogues.go:103:8:103:17 | epiRecover | epilogues.go:103:2:103:19 | defer statement |
| epilogues.go:103:8:103:19 | call to epiRecover | epilogues.go:102:1:110:1 | exit |
| epilogues.go:103:8:103:19 | call to epiRecover | epilogues.go:102:34:102:35 | implicit read of ok |
| epilogues.go:104:5:104:5 | x | epilogues.go:104:10:104:10 | 0 |
| epilogues.go:104:5:104:10 | ...==... | epilogues.go:104:5:104:10 | ...==... is false |
| epilogues.go:104:5:104:10 | ...==... | epilogues.go:104:5:104:10 | ...==... is true |
| epilogues.go:104:5:104:10 | ...==... is false | epilogues.go:107:2:107:2 | skip |
| epilogues.go:104:5:104:10 | ...==... is true | epilogues.go:105:3:105:8 | return statement |
| epilogues.go:104:10:104:10 | 0 | epilogues.go:104:5:104:10 | ...==... |
| epilogues.go:105:3:105:8 | return statement | epilogues.go:103:8:103:19 | call to epiRecover |
| epilogues.go:107:2:107:2 | assignment to n | epilogues.go:108:2:108:3 | skip |
| epilogues.go:107:2:107:2 | skip | epilogues.go:107:6:107:6 | x |
| epilogues.go:107:6:107:6 | x | epilogues.go:107:2:107:2 | assignment to n |
| epilogues.go:108:2:108:3 | assignment to ok | epilogues.go:109:2:109:7 | return statement |
| epilogues.go:108:2:108:3 | skip | epilogues.go:108:7:108:10 | true |
| epilogues.go:108:7:108:10 | true | epilogues.go:108:2:108:3 | assignment to ok |
| epilogues.go:109:2:109:7 | return statement | epilogues.go:103:8:103:19 | call to epiRecover |
| epilogues.go:115:1:118:1 | entry | epilogues.go:116:8:116:17 | epiRecover |
| epilogues.go:115:1:118:1 | function declaration | epilogues.go:0:0:0:0 | exit |
| epilogues.go:115:6:115:22 | skip | epilogues.go:115:1:118:1 | function declaration |
| epilogues.go:116:2:116:19 | defer statement | epilogues.go:117:2:117:6 | panic |
| epilogues.go:116:8:116:17 | epiRecover | epilogues.go:116:2:116:19 | defer statement |
| epilogues.go:116:8:116:19 | call to epiRecover | epilogues.go:115:1:118:1 | exit |
| epilogues.go:117:2:117:6 | panic | epilogues.go:117:8:117:13 | "boom" |
| epilogues.go:117:2:117:14 | call to panic | epilogues.go:116:8:116:19 | call to epiRecover |
| epilogues.go:117:8:117:13 | "boom" | epilogues.go:117:2:117:14 | call to panic |
| equalitytests.go:0:0:0:0 | entry | equalitytests.go:3:1:5:1 | skip | | equalitytests.go:0:0:0:0 | entry | equalitytests.go:3:1:5:1 | skip |
| equalitytests.go:3:1:5:1 | skip | equalitytests.go:7:1:9:1 | skip | | equalitytests.go:3:1:5:1 | skip | equalitytests.go:7:1:9:1 | skip |
| equalitytests.go:7:1:9:1 | skip | equalitytests.go:11:6:11:18 | skip | | equalitytests.go:7:1:9:1 | skip | equalitytests.go:11:6:11:18 | skip |

View File

@@ -1,3 +1,4 @@
| epilogues.go:115:6:115:22 | epiRecoverUnnamed | github.com/github/codeql-go/ql/test/library-tests/semmle/go/controlflow/ControlFlowGraph.epiRecoverUnnamed |
| file://:0:0:0:0 | Exit | os.Exit | | file://:0:0:0:0 | Exit | os.Exit |
| file://:0:0:0:0 | Fatal | log.Fatal | | file://:0:0:0:0 | Fatal | log.Fatal |
| file://:0:0:0:0 | Fatal | log.Logger.Fatal | | file://:0:0:0:0 | Fatal | log.Logger.Fatal |

View File

@@ -0,0 +1,118 @@
package main
import "fmt"
// epiLogger has methods with both pointer and value receivers, used to check
// that the receiver and arguments of a deferred call are evaluated at the
// `defer` statement rather than in the function epilogue.
type epiLogger struct {
prefix string
}
func (l *epiLogger) log(msg string, code int) {
fmt.Println(l.prefix, msg, code)
}
func (l epiLogger) logValue(msg string) {
fmt.Println(l.prefix, msg)
}
// epiRecover recovers from a panic. It is used as a deferred function so we can
// check that control flow returns to the result-read nodes and the normal exit
// node after recovering.
func epiRecover() {
if r := recover(); r != nil {
fmt.Println("recovered:", r)
}
}
// epiPlain has no named result variable and a single `return` with a child
// expression.
func epiPlain(x int) int {
return x * 2
}
// epiVoid has no named result variable and no `return` statement at all.
func epiVoid() {
fmt.Println("void")
}
// epiNamedMixed has named result variables and a mix of a bare `return` (no
// child expressions) and a `return` with child expressions.
func epiNamedMixed(x int) (result int, err error) {
if x < 0 {
result = -x
return
}
return x, nil
}
// epiNamedBareOnly has a named result variable and only a bare `return`.
func epiNamedBareOnly(x int) (n int) {
n = x + 1
return
}
// epiDeferReceiverArgs has a deferred call with a (pointer) receiver and
// arguments that are expressions, so we can check the receiver `l` and the
// arguments `"count"` and `len(items)` are evaluated at the `defer` statement.
func epiDeferReceiverArgs(l *epiLogger, items []int) {
defer l.log("count", len(items))
fmt.Println("processing", len(items))
}
// epiDeferValueReceiver has deferred calls with a value receiver and an
// address-of receiver, both with arguments evaluated at the `defer` statement.
func epiDeferValueReceiver(prefix string) {
l := epiLogger{prefix: prefix}
defer l.logValue("bye")
defer (&l).log("ptr", 7)
fmt.Println("body")
}
// epiDeferFuncLit has a deferred function literal with parameters, so we can
// check that the arguments `"done"` and `x+1` are evaluated at the `defer`
// statement and that control flow enters the function literal body when it is
// invoked at the function epilogue.
func epiDeferFuncLit(x int) {
defer func(label string, n int) {
fmt.Println(label, n)
}("done", x+1)
fmt.Println("body", x)
}
// epiRecoverNamed has a named result variable and a deferred closure containing
// `recover()`. After recovering on the panic path, control flow should return
// to the result-read nodes and the normal exit node.
func epiRecoverNamed(x int) (result int) {
defer func() {
if r := recover(); r != nil {
result = -1
}
}()
if x < 0 {
panic("neg")
}
result = x * x
return result
}
// epiRecoverNamedBare has named result variables, a deferred function
// containing `recover()`, and only bare `return` statements.
func epiRecoverNamedBare(x int) (ok bool, n int) {
defer epiRecover()
if x == 0 {
return
}
n = x
ok = true
return
}
// epiRecoverUnnamed has no named result variables and a deferred function
// containing `recover()`; after recovering, control flow should reach the
// normal exit node directly (there are no result-read nodes).
func epiRecoverUnnamed() {
defer epiRecover()
panic("boom")
}

View File

@@ -1,2 +0,0 @@
import semmle.python.controlflow.internal.AstNodeImpl
import ControlFlow::Consistency

View File

@@ -9,7 +9,6 @@ private import semmle.python.dataflow.new.internal.DataFlowImplSpecific
private import semmle.python.dataflow.new.internal.DataFlowDispatch private import semmle.python.dataflow.new.internal.DataFlowDispatch
private import semmle.python.dataflow.new.internal.TaintTrackingImplSpecific private import semmle.python.dataflow.new.internal.TaintTrackingImplSpecific
private import codeql.dataflow.internal.DataFlowImplConsistency private import codeql.dataflow.internal.DataFlowImplConsistency
private import semmle.python.controlflow.internal.Cfg as Cfg
private module Input implements InputSig<Location, PythonDataFlow> { private module Input implements InputSig<Location, PythonDataFlow> {
private import Private private import Private
@@ -75,7 +74,7 @@ private module Input implements InputSig<Location, PythonDataFlow> {
// resolve to multiple functions), but we only make _one_ ArgumentNode for each // resolve to multiple functions), but we only make _one_ ArgumentNode for each
// argument in the CallNode, we end up violating this consistency check in those // argument in the CallNode, we end up violating this consistency check in those
// cases. (see `getCallArg` in DataFlowDispatch.qll) // cases. (see `getCallArg` in DataFlowDispatch.qll)
exists(DataFlowCall other, Cfg::CallNode cfgCall | other != call | exists(DataFlowCall other, CallNode cfgCall | other != call |
call.getNode() = cfgCall and call.getNode() = cfgCall and
other.getNode() = cfgCall and other.getNode() = cfgCall and
isArgumentNode(arg, call, _) and isArgumentNode(arg, call, _) and
@@ -91,16 +90,16 @@ private module Input implements InputSig<Location, PythonDataFlow> {
// allow it instead. // allow it instead.
( (
call.getScope() = attr.getScope() and call.getScope() = attr.getScope() and
any(CfgNode n | n.asCfgNode() = call.getNode().(Cfg::CallNode).getFunction()) any(CfgNode n | n.asCfgNode() = call.getNode().(CallNode).getFunction()).getALocalSource() =
.getALocalSource() = attr attr
or or
not exists(call.getScope().(Function).getDefinition()) and not exists(call.getScope().(Function).getDefinition()) and
call.getScope().getScope+() = attr.getScope() call.getScope().getScope+() = attr.getScope()
) and ) and
( (
other.getScope() = attr.getScope() and other.getScope() = attr.getScope() and
any(CfgNode n | n.asCfgNode() = other.getNode().(Cfg::CallNode).getFunction()) any(CfgNode n | n.asCfgNode() = other.getNode().(CallNode).getFunction()).getALocalSource() =
.getALocalSource() = attr attr
or or
not exists(other.getScope().(Function).getDefinition()) and not exists(other.getScope().(Function).getDefinition()) and
other.getScope().getScope+() = attr.getScope() other.getScope().getScope+() = attr.getScope()

View File

@@ -1,4 +0,0 @@
---
category: minorAnalysis
---
* A new Python control flow graph implementation has been added under `semmle.python.controlflow.internal.Cfg` (backed by `AstNodeImpl.qll`), built on the shared `codeql.controlflow.ControlFlowGraph` library. It is not yet used by the dataflow library or any production query; the legacy CFG in `semmle/python/Flow.qll` remains the default. The new library is exposed for tests and for upcoming migrations.

View File

@@ -1,4 +0,0 @@
---
category: minorAnalysis
---
* A new SSA adapter has been added under `semmle.python.dataflow.new.internal.SsaImpl`, built on the shared `codeql.ssa.Ssa` library and the new shared CFG (`semmle.python.controlflow.internal.Cfg`). It is not yet used by the dataflow library or any production query; the legacy ESSA SSA in `semmle/python/essa/*` remains the default. The new SSA adapter is exposed for tests and for the upcoming dataflow migration.

View File

@@ -1,4 +0,0 @@
---
category: breaking
---
* The deprecated `AstNode.getAFlowNode()` and `Function.getAReturnValueFlowNode()` predicates now return nodes from the new shared CFG (`Cfg::ControlFlowNode`) rather than from the legacy CFG (`ControlFlowNode`). Callers that still rely on these deprecated APIs and feed the result into legacy-CFG-aware predicates will no longer type-check; migrate to `n.getNode() = e` (or, for return values, the explicit `Return` pattern shown in the deprecation message) to get nodes from the dataflow library's current CFG.

View File

@@ -1,4 +0,0 @@
---
category: minorAnalysis
---
* The new (shared-CFG-based) Python control flow graph now visits parameter and return type annotations as CFG nodes for function definitions, matching the legacy CFG. This restores annotation-based type tracking through framework models such as FastAPI's `Depends()`, Pydantic request models, Starlette `WebSocket` handlers, and any other models that flow a class reference through `Parameter.getAnnotation()` to identify instances of the annotated class.

View File

@@ -1,45 +0,0 @@
/**
* @name Print CFG (New)
* @description Produces a representation of a file's Control Flow Graph
* using the new shared control flow library.
* This query is used by the VS Code extension.
* @id python/print-cfg
* @kind graph
* @tags ide-contextual-queries/print-cfg
*/
private import python as Py
import semmle.python.controlflow.internal.AstNodeImpl
external string selectedSourceFile();
private predicate selectedSourceFileAlias = selectedSourceFile/0;
external int selectedSourceLine();
private predicate selectedSourceLineAlias = selectedSourceLine/0;
external int selectedSourceColumn();
private predicate selectedSourceColumnAlias = selectedSourceColumn/0;
module ViewCfgQueryInput implements ControlFlow::ViewCfgQueryInputSig<Py::File> {
predicate selectedSourceFile = selectedSourceFileAlias/0;
predicate selectedSourceLine = selectedSourceLineAlias/0;
predicate selectedSourceColumn = selectedSourceColumnAlias/0;
predicate cfgScopeSpan(
Ast::Callable callable, Py::File file, int startLine, int startColumn, int endLine,
int endColumn
) {
exists(Py::Scope scope |
scope = callable.asScope() and
file = scope.getLocation().getFile() and
scope.getLocation().hasLocationInfo(_, startLine, startColumn, endLine, endColumn)
)
}
}
import ControlFlow::ViewCfgQuery<Py::File, ViewCfgQueryInput>

View File

@@ -6,9 +6,8 @@
* directed and labeled; they specify how the components represented by nodes relate to each other. * directed and labeled; they specify how the components represented by nodes relate to each other.
*/ */
// Importing python under the `PY` namespace to avoid pulling in `CallNode` from `Flow.qll` (via `import python`) and thereby having a naming conflict with `API::CallNode`. // Importing python under the `py` namespace to avoid importing `CallNode` from `Flow.qll` and thereby having a naming conflict with `API::CallNode`.
private import python as PY private import python as PY
private import semmle.python.controlflow.internal.Cfg as Cfg
import semmle.python.dataflow.new.DataFlow import semmle.python.dataflow.new.DataFlow
private import semmle.python.internal.CachedStages private import semmle.python.internal.CachedStages
@@ -283,7 +282,7 @@ module API {
index = this.getIndex() and index = this.getIndex() and
( (
// subscripting // subscripting
exists(Cfg::SubscriptNode subscript | exists(PY::SubscriptNode subscript |
subscript.getObject() = this.getAValueReachableFromSource().asCfgNode() and subscript.getObject() = this.getAValueReachableFromSource().asCfgNode() and
subscript.getIndex() = index.asSink().asCfgNode() subscript.getIndex() = index.asSink().asCfgNode()
| |
@@ -291,7 +290,7 @@ module API {
subscript = result.asSource().asCfgNode() subscript = result.asSource().asCfgNode()
or or
// writing // writing
subscript.(Cfg::DefinitionNode).getValue() = result.asSink().asCfgNode() subscript.(PY::DefinitionNode).getValue() = result.asSink().asCfgNode()
) )
or or
// dictionary literals // dictionary literals
@@ -685,7 +684,7 @@ module API {
* Ignores relative imports, such as `from ..foo.bar import baz`. * Ignores relative imports, such as `from ..foo.bar import baz`.
*/ */
private predicate imports(DataFlow::CfgNode imp, string name) { private predicate imports(DataFlow::CfgNode imp, string name) {
exists(Cfg::ImportExprNode iexpr | exists(PY::ImportExprNode iexpr |
imp.getNode() = iexpr and imp.getNode() = iexpr and
not iexpr.getNode().isRelative() and not iexpr.getNode().isRelative() and
name = iexpr.getNode().getImportedModuleName() name = iexpr.getNode().getImportedModuleName()
@@ -776,7 +775,7 @@ module API {
// list literals, from `x` to `[x]` // list literals, from `x` to `[x]`
// TODO: once convenient, this should be done at a higher level than the AST, // TODO: once convenient, this should be done at a higher level than the AST,
// at least at the CFG layer, to take splitting into account. // at least at the CFG layer, to take splitting into account.
// Also consider `Cfg::SequenceNode` for generality. // Also consider `SequenceNode for generality.
exists(PY::List list | list = pred.(DataFlow::ExprNode).getNode().getNode() | exists(PY::List list | list = pred.(DataFlow::ExprNode).getNode().getNode() |
rhs.(DataFlow::ExprNode).getNode().getNode() = list.getAnElt() and rhs.(DataFlow::ExprNode).getNode().getNode() = list.getAnElt() and
lbl = Label::subscript() lbl = Label::subscript()
@@ -806,7 +805,7 @@ module API {
subscript = trackUseNode(src).getSubscript(index) subscript = trackUseNode(src).getSubscript(index)
| |
// from `x` to a definition of `x[...]` // from `x` to a definition of `x[...]`
rhs.asCfgNode() = subscript.asCfgNode().(Cfg::DefinitionNode).getValue() and rhs.asCfgNode() = subscript.asCfgNode().(PY::DefinitionNode).getValue() and
lbl = Label::subscript() lbl = Label::subscript()
or or
// from `x` to `"key"` in `x["key"]` // from `x` to `"key"` in `x["key"]`

View File

@@ -3,7 +3,6 @@ module;
import python import python
private import semmle.python.internal.CachedStages private import semmle.python.internal.CachedStages
private import semmle.python.controlflow.internal.Cfg as Cfg
/** A syntactic node (Class, Function, Module, Expr, Stmt or Comprehension) corresponding to a flow node */ /** A syntactic node (Class, Function, Module, Expr, Stmt or Comprehension) corresponding to a flow node */
abstract class AstNode extends AstNode_ { abstract class AstNode extends AstNode_ {
@@ -20,16 +19,17 @@ abstract class AstNode extends AstNode_ {
/** /**
* DEPRECATED: use `ControlFlowNode.getNode()` from the other direction instead; * DEPRECATED: use `ControlFlowNode.getNode()` from the other direction instead;
* that is, replace `e.getAFlowNode() = n` with `n.getNode() = e`. This API is * that is, replace `e.getAFlowNode() = n` with `n.getNode() = e`. This API is
* being removed to untangle the AST and CFG hierarchies. * being removed to untangle the AST and CFG hierarchies in preparation for
* migrating the dataflow library off the legacy CFG.
* *
* Gets a flow node corresponding directly to this node, from the new * Gets a flow node corresponding directly to this node.
* (shared) CFG. NOTE: For some statements and other purely syntactic * NOTE: For some statements and other purely syntactic elements,
* elements, there may not be a `ControlFlowNode`. * there may not be a `ControlFlowNode`.
*/ */
cached cached
deprecated Cfg::ControlFlowNode getAFlowNode() { deprecated ControlFlowNode getAFlowNode() {
Stages::AST::ref() and Stages::AST::ref() and
result.getNode() = this py_flow_bb_node(result, this, _, _)
} }
/** Gets the location for this AST node */ /** Gets the location for this AST node */

View File

@@ -5,7 +5,6 @@
*/ */
private import python private import python
private import semmle.python.controlflow.internal.Cfg as Cfg
private import semmle.python.dataflow.new.DataFlow private import semmle.python.dataflow.new.DataFlow
private import semmle.python.dataflow.new.internal.DataFlowImplSpecific private import semmle.python.dataflow.new.internal.DataFlowImplSpecific
private import semmle.python.dataflow.new.RemoteFlowSources private import semmle.python.dataflow.new.RemoteFlowSources
@@ -215,7 +214,7 @@ module Path {
SafeAccessCheck() { this = DataFlow::BarrierGuard<safeAccessCheck/3>::getABarrierNode() } SafeAccessCheck() { this = DataFlow::BarrierGuard<safeAccessCheck/3>::getABarrierNode() }
} }
private predicate safeAccessCheck(DataFlow::GuardNode g, Cfg::ControlFlowNode node, boolean branch) { private predicate safeAccessCheck(DataFlow::GuardNode g, ControlFlowNode node, boolean branch) {
g.(SafeAccessCheck::Range).checks(node, branch) g.(SafeAccessCheck::Range).checks(node, branch)
} }
@@ -224,7 +223,7 @@ module Path {
/** A data-flow node that checks that a path is safe to access in some way, for example by having a controlled prefix. */ /** A data-flow node that checks that a path is safe to access in some way, for example by having a controlled prefix. */
abstract class Range extends DataFlow::GuardNode { abstract class Range extends DataFlow::GuardNode {
/** Holds if this guard validates `node` upon evaluating to `branch`. */ /** Holds if this guard validates `node` upon evaluating to `branch`. */
abstract predicate checks(Cfg::ControlFlowNode node, boolean branch); abstract predicate checks(ControlFlowNode node, boolean branch);
} }
} }
} }

View File

@@ -3,7 +3,6 @@ module;
private import python private import python
private import semmle.python.internal.CachedStages private import semmle.python.internal.CachedStages
private import semmle.python.controlflow.internal.Cfg as Cfg
/** An expression */ /** An expression */
class Expr extends Expr_, AstNode { class Expr extends Expr_, AstNode {
@@ -71,7 +70,7 @@ class Attribute extends Attribute_ {
/* syntax: Expr.name */ /* syntax: Expr.name */
override Expr getASubExpression() { result = this.getObject() } override Expr getASubExpression() { result = this.getObject() }
deprecated override Cfg::AttrNode getAFlowNode() { result = super.getAFlowNode() } deprecated override AttrNode getAFlowNode() { result = super.getAFlowNode() }
/** Gets the name of this attribute. That is the `name` in `obj.name` */ /** Gets the name of this attribute. That is the `name` in `obj.name` */
string getName() { result = Attribute_.super.getAttr() } string getName() { result = Attribute_.super.getAttr() }
@@ -100,7 +99,7 @@ class Subscript extends Subscript_ {
Expr getObject() { result = Subscript_.super.getValue() } Expr getObject() { result = Subscript_.super.getValue() }
deprecated override Cfg::SubscriptNode getAFlowNode() { result = super.getAFlowNode() } deprecated override SubscriptNode getAFlowNode() { result = super.getAFlowNode() }
} }
/** A call expression, such as `func(...)` */ /** A call expression, such as `func(...)` */
@@ -116,7 +115,7 @@ class Call extends Call_ {
override string toString() { result = this.getFunc().toString() + "()" } override string toString() { result = this.getFunc().toString() + "()" }
deprecated override Cfg::CallNode getAFlowNode() { result = super.getAFlowNode() } deprecated override CallNode getAFlowNode() { result = super.getAFlowNode() }
/** Gets a tuple (*) argument of this call. */ /** Gets a tuple (*) argument of this call. */
Expr getStarargs() { result = this.getAPositionalArg().(Starred).getValue() } Expr getStarargs() { result = this.getAPositionalArg().(Starred).getValue() }
@@ -204,7 +203,7 @@ class IfExp extends IfExp_ {
result = this.getTest() or result = this.getBody() or result = this.getOrelse() result = this.getTest() or result = this.getBody() or result = this.getOrelse()
} }
deprecated override Cfg::IfExprNode getAFlowNode() { result = super.getAFlowNode() } deprecated override IfExprNode getAFlowNode() { result = super.getAFlowNode() }
} }
/** A starred expression, such as the `*rest` in the assignment `first, *rest = seq` */ /** A starred expression, such as the `*rest` in the assignment `first, *rest = seq` */
@@ -414,7 +413,7 @@ class PlaceHolder extends PlaceHolder_ {
override string toString() { result = "$" + this.getId() } override string toString() { result = "$" + this.getId() }
deprecated override Cfg::NameNode getAFlowNode() { result = super.getAFlowNode() } deprecated override NameNode getAFlowNode() { result = super.getAFlowNode() }
} }
/** A tuple expression such as `( 1, 3, 5, 7, 9 )` */ /** A tuple expression such as `( 1, 3, 5, 7, 9 )` */
@@ -481,7 +480,7 @@ class Name extends Name_ {
override string toString() { result = this.getId() } override string toString() { result = this.getId() }
deprecated override Cfg::NameNode getAFlowNode() { result = super.getAFlowNode() } deprecated override NameNode getAFlowNode() { result = super.getAFlowNode() }
override predicate isArtificial() { override predicate isArtificial() {
/* Artificial variable names in comprehensions all start with "." */ /* Artificial variable names in comprehensions all start with "." */
@@ -588,7 +587,7 @@ abstract class NameConstant extends Name, ImmutableLiteral {
override predicate isConstant() { any() } override predicate isConstant() { any() }
deprecated override Cfg::NameConstantNode getAFlowNode() { result = Name.super.getAFlowNode() } deprecated override NameConstantNode getAFlowNode() { result = Name.super.getAFlowNode() }
override predicate isArtificial() { none() } override predicate isArtificial() { none() }
} }

View File

@@ -1,7 +1,7 @@
overlay[local] overlay[local]
module; module;
import python as Py import python
private import semmle.python.internal.CachedStages private import semmle.python.internal.CachedStages
private import codeql.controlflow.BasicBlock as BB private import codeql.controlflow.BasicBlock as BB
@@ -17,7 +17,7 @@ private import codeql.controlflow.BasicBlock as BB
*/ */
private predicate augstore(ControlFlowNode load, ControlFlowNode store) { private predicate augstore(ControlFlowNode load, ControlFlowNode store) {
exists(Py::Expr load_store | exists(Py::AugAssign aa | aa.getTarget() = load_store) | exists(Expr load_store | exists(AugAssign aa | aa.getTarget() = load_store) |
toAst(load) = load_store and toAst(load) = load_store and
toAst(store) = load_store and toAst(store) = load_store and
load.strictlyDominates(store) load.strictlyDominates(store)
@@ -25,7 +25,7 @@ private predicate augstore(ControlFlowNode load, ControlFlowNode store) {
} }
/** A non-dispatched getNode() to avoid negative recursion issues */ /** A non-dispatched getNode() to avoid negative recursion issues */
private Py::AstNode toAst(ControlFlowNode n) { py_flow_bb_node(n, result, _, _) } private AstNode toAst(ControlFlowNode n) { py_flow_bb_node(n, result, _, _) }
/** /**
* A control flow node. Control flow nodes have a many-to-one relation with syntactic nodes, * A control flow node. Control flow nodes have a many-to-one relation with syntactic nodes,
@@ -35,19 +35,19 @@ private Py::AstNode toAst(ControlFlowNode n) { py_flow_bb_node(n, result, _, _)
class ControlFlowNode extends @py_flow_node { class ControlFlowNode extends @py_flow_node {
/** Whether this control flow node is a load (including those in augmented assignments) */ /** Whether this control flow node is a load (including those in augmented assignments) */
predicate isLoad() { predicate isLoad() {
exists(Py::Expr e | e = toAst(this) | py_expr_contexts(_, 3, e) and not augstore(_, this)) exists(Expr e | e = toAst(this) | py_expr_contexts(_, 3, e) and not augstore(_, this))
} }
/** Whether this control flow node is a store (including those in augmented assignments) */ /** Whether this control flow node is a store (including those in augmented assignments) */
predicate isStore() { predicate isStore() {
exists(Py::Expr e | e = toAst(this) | py_expr_contexts(_, 5, e) or augstore(_, this)) exists(Expr e | e = toAst(this) | py_expr_contexts(_, 5, e) or augstore(_, this))
} }
/** Whether this control flow node is a delete */ /** Whether this control flow node is a delete */
predicate isDelete() { exists(Py::Expr e | e = toAst(this) | py_expr_contexts(_, 2, e)) } predicate isDelete() { exists(Expr e | e = toAst(this) | py_expr_contexts(_, 2, e)) }
/** Whether this control flow node is a parameter */ /** Whether this control flow node is a parameter */
predicate isParameter() { exists(Py::Expr e | e = toAst(this) | py_expr_contexts(_, 4, e)) } predicate isParameter() { exists(Expr e | e = toAst(this) | py_expr_contexts(_, 4, e)) }
/** Whether this control flow node is a store in an augmented assignment */ /** Whether this control flow node is a store in an augmented assignment */
predicate isAugStore() { augstore(_, this) } predicate isAugStore() { augstore(_, this) }
@@ -57,61 +57,61 @@ class ControlFlowNode extends @py_flow_node {
/** Whether this flow node corresponds to a literal */ /** Whether this flow node corresponds to a literal */
predicate isLiteral() { predicate isLiteral() {
toAst(this) instanceof Py::Bytes toAst(this) instanceof Bytes
or or
toAst(this) instanceof Py::Dict toAst(this) instanceof Dict
or or
toAst(this) instanceof Py::DictComp toAst(this) instanceof DictComp
or or
toAst(this) instanceof Py::Set toAst(this) instanceof Set
or or
toAst(this) instanceof Py::SetComp toAst(this) instanceof SetComp
or or
toAst(this) instanceof Py::Ellipsis toAst(this) instanceof Ellipsis
or or
toAst(this) instanceof Py::GeneratorExp toAst(this) instanceof GeneratorExp
or or
toAst(this) instanceof Py::Lambda toAst(this) instanceof Lambda
or or
toAst(this) instanceof Py::ListComp toAst(this) instanceof ListComp
or or
toAst(this) instanceof Py::List toAst(this) instanceof List
or or
toAst(this) instanceof Py::Num toAst(this) instanceof Num
or or
toAst(this) instanceof Py::Tuple toAst(this) instanceof Tuple
or or
toAst(this) instanceof Py::Unicode toAst(this) instanceof Unicode
or or
toAst(this) instanceof Py::NameConstant toAst(this) instanceof NameConstant
} }
/** Whether this flow node corresponds to an attribute expression */ /** Whether this flow node corresponds to an attribute expression */
predicate isAttribute() { toAst(this) instanceof Py::Attribute } predicate isAttribute() { toAst(this) instanceof Attribute }
/** Whether this flow node corresponds to an subscript expression */ /** Whether this flow node corresponds to an subscript expression */
predicate isSubscript() { toAst(this) instanceof Py::Subscript } predicate isSubscript() { toAst(this) instanceof Subscript }
/** Whether this flow node corresponds to an import member */ /** Whether this flow node corresponds to an import member */
predicate isImportMember() { toAst(this) instanceof Py::ImportMember } predicate isImportMember() { toAst(this) instanceof ImportMember }
/** Whether this flow node corresponds to a call */ /** Whether this flow node corresponds to a call */
predicate isCall() { toAst(this) instanceof Py::Call } predicate isCall() { toAst(this) instanceof Call }
/** Whether this flow node is the first in a module */ /** Whether this flow node is the first in a module */
predicate isModuleEntry() { this.isEntryNode() and toAst(this) instanceof Py::Module } predicate isModuleEntry() { this.isEntryNode() and toAst(this) instanceof Module }
/** Whether this flow node corresponds to an import */ /** Whether this flow node corresponds to an import */
predicate isImport() { toAst(this) instanceof Py::ImportExpr } predicate isImport() { toAst(this) instanceof ImportExpr }
/** Whether this flow node corresponds to a conditional expression */ /** Whether this flow node corresponds to a conditional expression */
predicate isIfExp() { toAst(this) instanceof Py::IfExp } predicate isIfExp() { toAst(this) instanceof IfExp }
/** Whether this flow node corresponds to a function definition expression */ /** Whether this flow node corresponds to a function definition expression */
predicate isFunction() { toAst(this) instanceof Py::FunctionExpr } predicate isFunction() { toAst(this) instanceof FunctionExpr }
/** Whether this flow node corresponds to a class definition expression */ /** Whether this flow node corresponds to a class definition expression */
predicate isClass() { toAst(this) instanceof Py::ClassExpr } predicate isClass() { toAst(this) instanceof ClassExpr }
/** Gets a predecessor of this flow node */ /** Gets a predecessor of this flow node */
ControlFlowNode getAPredecessor() { this = result.getASuccessor() } ControlFlowNode getAPredecessor() { this = result.getASuccessor() }
@@ -123,25 +123,25 @@ class ControlFlowNode extends @py_flow_node {
ControlFlowNode getImmediateDominator() { py_idoms(this, result) } ControlFlowNode getImmediateDominator() { py_idoms(this, result) }
/** Gets the syntactic element corresponding to this flow node */ /** Gets the syntactic element corresponding to this flow node */
Py::AstNode getNode() { py_flow_bb_node(this, result, _, _) } AstNode getNode() { py_flow_bb_node(this, result, _, _) }
/** Gets a textual representation of this element. */ /** Gets a textual representation of this element. */
cached cached
string toString() { string toString() {
Stages::AST::ref() and Stages::AST::ref() and
// Since modules can have ambigous names, entry nodes can too, if we do not collate them. // Since modules can have ambigous names, entry nodes can too, if we do not collate them.
exists(Py::Scope s | s.getEntryNode() = this | exists(Scope s | s.getEntryNode() = this |
result = "Entry node for " + concat( | | s.toString(), ",") result = "Entry node for " + concat( | | s.toString(), ",")
) )
or or
exists(Py::Scope s | s.getANormalExit() = this | result = "Exit node for " + s.toString()) exists(Scope s | s.getANormalExit() = this | result = "Exit node for " + s.toString())
or or
not exists(Py::Scope s | s.getEntryNode() = this or s.getANormalExit() = this) and not exists(Scope s | s.getEntryNode() = this or s.getANormalExit() = this) and
result = "ControlFlowNode for " + this.getNode().toString() result = "ControlFlowNode for " + this.getNode().toString()
} }
/** Gets the location of this ControlFlowNode */ /** Gets the location of this ControlFlowNode */
Py::Location getLocation() { result = this.getNode().getLocation() } Location getLocation() { result = this.getNode().getLocation() }
/** Whether this flow node is the first in its scope */ /** Whether this flow node is the first in its scope */
predicate isEntryNode() { py_scope_flow(this, _, -1) } predicate isEntryNode() { py_scope_flow(this, _, -1) }
@@ -151,9 +151,9 @@ class ControlFlowNode extends @py_flow_node {
/** Gets the scope containing this flow node */ /** Gets the scope containing this flow node */
cached cached
Py::Scope getScope() { Scope getScope() {
Stages::AST::ref() and Stages::AST::ref() and
if this.getNode() instanceof Py::Scope if this.getNode() instanceof Scope
then then
/* Entry or exit node */ /* Entry or exit node */
result = this.getNode() result = this.getNode()
@@ -161,7 +161,7 @@ class ControlFlowNode extends @py_flow_node {
} }
/** Gets the enclosing module */ /** Gets the enclosing module */
Py::Module getEnclosingModule() { result = this.getScope().getEnclosingModule() } Module getEnclosingModule() { result = this.getScope().getEnclosingModule() }
/** Gets a successor for this node if the relevant condition is True. */ /** Gets a successor for this node if the relevant condition is True. */
ControlFlowNode getATrueSuccessor() { ControlFlowNode getATrueSuccessor() {
@@ -188,7 +188,7 @@ class ControlFlowNode extends @py_flow_node {
} }
/** Whether the scope may be exited as a result of this node raising an exception */ /** Whether the scope may be exited as a result of this node raising an exception */
predicate isExceptionalExit(Py::Scope s) { py_scope_flow(this, s, 1) } predicate isExceptionalExit(Scope s) { py_scope_flow(this, s, 1) }
/** Whether this node is a normal (non-exceptional) exit */ /** Whether this node is a normal (non-exceptional) exit */
predicate isNormalExit() { py_scope_flow(this, _, 0) or py_scope_flow(this, _, 2) } predicate isNormalExit() { py_scope_flow(this, _, 0) or py_scope_flow(this, _, 2) }
@@ -236,7 +236,7 @@ class ControlFlowNode extends @py_flow_node {
/* join-ordering helper for `getAChild() */ /* join-ordering helper for `getAChild() */
pragma[noinline] pragma[noinline]
private ControlFlowNode getExprChild(BasicBlock dom) { private ControlFlowNode getExprChild(BasicBlock dom) {
this.getNode().(Py::Expr).getAChildNode() = result.getNode() and this.getNode().(Expr).getAChildNode() = result.getNode() and
result.getBasicBlock().dominates(dom) and result.getBasicBlock().dominates(dom) and
not this instanceof UnaryExprNode not this instanceof UnaryExprNode
} }
@@ -249,16 +249,16 @@ class ControlFlowNode extends @py_flow_node {
*/ */
private class AnyNode extends ControlFlowNode { private class AnyNode extends ControlFlowNode {
override Py::AstNode getNode() { result = super.getNode() } override AstNode getNode() { result = super.getNode() }
} }
/** A control flow node corresponding to a call expression, such as `func(...)` */ /** A control flow node corresponding to a call expression, such as `func(...)` */
class CallNode extends ControlFlowNode { class CallNode extends ControlFlowNode {
CallNode() { toAst(this) instanceof Py::Call } CallNode() { toAst(this) instanceof Call }
/** Gets the flow node corresponding to the function expression for the call corresponding to this flow node */ /** Gets the flow node corresponding to the function expression for the call corresponding to this flow node */
ControlFlowNode getFunction() { ControlFlowNode getFunction() {
exists(Py::Call c | exists(Call c |
this.getNode() = c and this.getNode() = c and
c.getFunc() = result.getNode() and c.getFunc() = result.getNode() and
result.getBasicBlock().dominates(this.getBasicBlock()) result.getBasicBlock().dominates(this.getBasicBlock())
@@ -267,7 +267,7 @@ class CallNode extends ControlFlowNode {
/** Gets the flow node corresponding to the n'th positional argument of the call corresponding to this flow node */ /** Gets the flow node corresponding to the n'th positional argument of the call corresponding to this flow node */
ControlFlowNode getArg(int n) { ControlFlowNode getArg(int n) {
exists(Py::Call c | exists(Call c |
this.getNode() = c and this.getNode() = c and
c.getArg(n) = result.getNode() and c.getArg(n) = result.getNode() and
result.getBasicBlock().dominates(this.getBasicBlock()) result.getBasicBlock().dominates(this.getBasicBlock())
@@ -276,7 +276,7 @@ class CallNode extends ControlFlowNode {
/** Gets the flow node corresponding to the named argument of the call corresponding to this flow node */ /** Gets the flow node corresponding to the named argument of the call corresponding to this flow node */
ControlFlowNode getArgByName(string name) { ControlFlowNode getArgByName(string name) {
exists(Py::Call c, Py::Keyword k | exists(Call c, Keyword k |
this.getNode() = c and this.getNode() = c and
k = c.getANamedArg() and k = c.getANamedArg() and
k.getValue() = result.getNode() and k.getValue() = result.getNode() and
@@ -292,7 +292,7 @@ class CallNode extends ControlFlowNode {
result = this.getArgByName(_) result = this.getArgByName(_)
} }
override Py::Call getNode() { result = super.getNode() } override Call getNode() { result = super.getNode() }
predicate isDecoratorCall() { predicate isDecoratorCall() {
this.isClassDecoratorCall() this.isClassDecoratorCall()
@@ -301,11 +301,11 @@ class CallNode extends ControlFlowNode {
} }
predicate isClassDecoratorCall() { predicate isClassDecoratorCall() {
exists(Py::ClassExpr cls | this.getNode() = cls.getADecoratorCall()) exists(ClassExpr cls | this.getNode() = cls.getADecoratorCall())
} }
predicate isFunctionDecoratorCall() { predicate isFunctionDecoratorCall() {
exists(Py::FunctionExpr func | this.getNode() = func.getADecoratorCall()) exists(FunctionExpr func | this.getNode() = func.getADecoratorCall())
} }
/** Gets the first tuple (*) argument of this call, if any. */ /** Gets the first tuple (*) argument of this call, if any. */
@@ -323,11 +323,11 @@ class CallNode extends ControlFlowNode {
/** A control flow corresponding to an attribute expression, such as `value.attr` */ /** A control flow corresponding to an attribute expression, such as `value.attr` */
class AttrNode extends ControlFlowNode { class AttrNode extends ControlFlowNode {
AttrNode() { toAst(this) instanceof Py::Attribute } AttrNode() { toAst(this) instanceof Attribute }
/** Gets the flow node corresponding to the object of the attribute expression corresponding to this flow node */ /** Gets the flow node corresponding to the object of the attribute expression corresponding to this flow node */
ControlFlowNode getObject() { ControlFlowNode getObject() {
exists(Py::Attribute a | exists(Attribute a |
this.getNode() = a and this.getNode() = a and
a.getObject() = result.getNode() and a.getObject() = result.getNode() and
result.getBasicBlock().dominates(this.getBasicBlock()) result.getBasicBlock().dominates(this.getBasicBlock())
@@ -339,7 +339,7 @@ class AttrNode extends ControlFlowNode {
* with the matching name * with the matching name
*/ */
ControlFlowNode getObject(string name) { ControlFlowNode getObject(string name) {
exists(Py::Attribute a | exists(Attribute a |
this.getNode() = a and this.getNode() = a and
a.getObject() = result.getNode() and a.getObject() = result.getNode() and
a.getName() = name and a.getName() = name and
@@ -348,57 +348,57 @@ class AttrNode extends ControlFlowNode {
} }
/** Gets the attribute name of the attribute expression corresponding to this flow node */ /** Gets the attribute name of the attribute expression corresponding to this flow node */
string getName() { exists(Py::Attribute a | this.getNode() = a and a.getName() = result) } string getName() { exists(Attribute a | this.getNode() = a and a.getName() = result) }
override Py::Attribute getNode() { result = super.getNode() } override Attribute getNode() { result = super.getNode() }
} }
/** A control flow node corresponding to a `from ... import ...` expression */ /** A control flow node corresponding to a `from ... import ...` expression */
class ImportMemberNode extends ControlFlowNode { class ImportMemberNode extends ControlFlowNode {
ImportMemberNode() { toAst(this) instanceof Py::ImportMember } ImportMemberNode() { toAst(this) instanceof ImportMember }
/** /**
* Gets the flow node corresponding to the module in the import-member expression corresponding to this flow node, * Gets the flow node corresponding to the module in the import-member expression corresponding to this flow node,
* with the matching name * with the matching name
*/ */
ControlFlowNode getModule(string name) { ControlFlowNode getModule(string name) {
exists(Py::ImportMember i | this.getNode() = i and i.getModule() = result.getNode() | exists(ImportMember i | this.getNode() = i and i.getModule() = result.getNode() |
i.getName() = name and i.getName() = name and
result.getBasicBlock().dominates(this.getBasicBlock()) result.getBasicBlock().dominates(this.getBasicBlock())
) )
} }
override Py::ImportMember getNode() { result = super.getNode() } override ImportMember getNode() { result = super.getNode() }
} }
/** A control flow node corresponding to an artificial expression representing an import */ /** A control flow node corresponding to an artificial expression representing an import */
class ImportExprNode extends ControlFlowNode { class ImportExprNode extends ControlFlowNode {
ImportExprNode() { toAst(this) instanceof Py::ImportExpr } ImportExprNode() { toAst(this) instanceof ImportExpr }
override Py::ImportExpr getNode() { result = super.getNode() } override ImportExpr getNode() { result = super.getNode() }
} }
/** A control flow node corresponding to a `from ... import *` statement */ /** A control flow node corresponding to a `from ... import *` statement */
class ImportStarNode extends ControlFlowNode { class ImportStarNode extends ControlFlowNode {
ImportStarNode() { toAst(this) instanceof Py::ImportStar } ImportStarNode() { toAst(this) instanceof ImportStar }
/** Gets the flow node corresponding to the module in the import-star corresponding to this flow node */ /** Gets the flow node corresponding to the module in the import-star corresponding to this flow node */
ControlFlowNode getModule() { ControlFlowNode getModule() {
exists(Py::ImportStar i | this.getNode() = i and i.getModuleExpr() = result.getNode() | exists(ImportStar i | this.getNode() = i and i.getModuleExpr() = result.getNode() |
result.getBasicBlock().dominates(this.getBasicBlock()) result.getBasicBlock().dominates(this.getBasicBlock())
) )
} }
override Py::ImportStar getNode() { result = super.getNode() } override ImportStar getNode() { result = super.getNode() }
} }
/** A control flow node corresponding to a subscript expression, such as `value[slice]` */ /** A control flow node corresponding to a subscript expression, such as `value[slice]` */
class SubscriptNode extends ControlFlowNode { class SubscriptNode extends ControlFlowNode {
SubscriptNode() { toAst(this) instanceof Py::Subscript } SubscriptNode() { toAst(this) instanceof Subscript }
/** flow node corresponding to the value of the sequence in a subscript operation */ /** flow node corresponding to the value of the sequence in a subscript operation */
ControlFlowNode getObject() { ControlFlowNode getObject() {
exists(Py::Subscript s | exists(Subscript s |
this.getNode() = s and this.getNode() = s and
s.getObject() = result.getNode() and s.getObject() = result.getNode() and
result.getBasicBlock().dominates(this.getBasicBlock()) result.getBasicBlock().dominates(this.getBasicBlock())
@@ -407,23 +407,23 @@ class SubscriptNode extends ControlFlowNode {
/** flow node corresponding to the index in a subscript operation */ /** flow node corresponding to the index in a subscript operation */
ControlFlowNode getIndex() { ControlFlowNode getIndex() {
exists(Py::Subscript s | exists(Subscript s |
this.getNode() = s and this.getNode() = s and
s.getIndex() = result.getNode() and s.getIndex() = result.getNode() and
result.getBasicBlock().dominates(this.getBasicBlock()) result.getBasicBlock().dominates(this.getBasicBlock())
) )
} }
override Py::Subscript getNode() { result = super.getNode() } override Subscript getNode() { result = super.getNode() }
} }
/** A control flow node corresponding to a comparison operation, such as `x<y` */ /** A control flow node corresponding to a comparison operation, such as `x<y` */
class CompareNode extends ControlFlowNode { class CompareNode extends ControlFlowNode {
CompareNode() { toAst(this) instanceof Py::Compare } CompareNode() { toAst(this) instanceof Compare }
/** Whether left and right are a pair of operands for this comparison */ /** Whether left and right are a pair of operands for this comparison */
predicate operands(ControlFlowNode left, Py::Cmpop op, ControlFlowNode right) { predicate operands(ControlFlowNode left, Cmpop op, ControlFlowNode right) {
exists(Py::Compare c, Py::Expr eleft, Py::Expr eright | exists(Compare c, Expr eleft, Expr eright |
this.getNode() = c and left.getNode() = eleft and right.getNode() = eright this.getNode() = c and left.getNode() = eleft and right.getNode() = eright
| |
eleft = c.getLeft() and eright = c.getComparator(0) and op = c.getOp(0) eleft = c.getLeft() and eright = c.getComparator(0) and op = c.getOp(0)
@@ -436,26 +436,26 @@ class CompareNode extends ControlFlowNode {
right.getBasicBlock().dominates(this.getBasicBlock()) right.getBasicBlock().dominates(this.getBasicBlock())
} }
override Py::Compare getNode() { result = super.getNode() } override Compare getNode() { result = super.getNode() }
} }
/** A control flow node corresponding to a conditional expression such as, `body if test else orelse` */ /** A control flow node corresponding to a conditional expression such as, `body if test else orelse` */
class IfExprNode extends ControlFlowNode { class IfExprNode extends ControlFlowNode {
IfExprNode() { toAst(this) instanceof Py::IfExp } IfExprNode() { toAst(this) instanceof IfExp }
/** flow node corresponding to one of the operands of an if-expression */ /** flow node corresponding to one of the operands of an if-expression */
ControlFlowNode getAnOperand() { result = this.getAPredecessor() } ControlFlowNode getAnOperand() { result = this.getAPredecessor() }
override Py::IfExp getNode() { result = super.getNode() } override IfExp getNode() { result = super.getNode() }
} }
/** A control flow node corresponding to an assignment expression such as `lhs := rhs`. */ /** A control flow node corresponding to an assignment expression such as `lhs := rhs`. */
class AssignmentExprNode extends ControlFlowNode { class AssignmentExprNode extends ControlFlowNode {
AssignmentExprNode() { toAst(this) instanceof Py::AssignExpr } AssignmentExprNode() { toAst(this) instanceof AssignExpr }
/** Gets the flow node corresponding to the left-hand side of the assignment expression */ /** Gets the flow node corresponding to the left-hand side of the assignment expression */
ControlFlowNode getTarget() { ControlFlowNode getTarget() {
exists(Py::AssignExpr a | exists(AssignExpr a |
this.getNode() = a and this.getNode() = a and
a.getTarget() = result.getNode() and a.getTarget() = result.getNode() and
result.getBasicBlock().dominates(this.getBasicBlock()) result.getBasicBlock().dominates(this.getBasicBlock())
@@ -464,27 +464,27 @@ class AssignmentExprNode extends ControlFlowNode {
/** Gets the flow node corresponding to the right-hand side of the assignment expression */ /** Gets the flow node corresponding to the right-hand side of the assignment expression */
ControlFlowNode getValue() { ControlFlowNode getValue() {
exists(Py::AssignExpr a | exists(AssignExpr a |
this.getNode() = a and this.getNode() = a and
a.getValue() = result.getNode() and a.getValue() = result.getNode() and
result.getBasicBlock().dominates(this.getBasicBlock()) result.getBasicBlock().dominates(this.getBasicBlock())
) )
} }
override Py::AssignExpr getNode() { result = super.getNode() } override AssignExpr getNode() { result = super.getNode() }
} }
/** A control flow node corresponding to a binary expression, such as `x + y` */ /** A control flow node corresponding to a binary expression, such as `x + y` */
class BinaryExprNode extends ControlFlowNode { class BinaryExprNode extends ControlFlowNode {
BinaryExprNode() { toAst(this) instanceof Py::BinaryExpr } BinaryExprNode() { toAst(this) instanceof BinaryExpr }
/** flow node corresponding to one of the operands of a binary expression */ /** flow node corresponding to one of the operands of a binary expression */
ControlFlowNode getAnOperand() { result = this.getLeft() or result = this.getRight() } ControlFlowNode getAnOperand() { result = this.getLeft() or result = this.getRight() }
override Py::BinaryExpr getNode() { result = super.getNode() } override BinaryExpr getNode() { result = super.getNode() }
ControlFlowNode getLeft() { ControlFlowNode getLeft() {
exists(Py::BinaryExpr b | exists(BinaryExpr b |
this.getNode() = b and this.getNode() = b and
result.getNode() = b.getLeft() and result.getNode() = b.getLeft() and
result.getBasicBlock().dominates(this.getBasicBlock()) result.getBasicBlock().dominates(this.getBasicBlock())
@@ -492,7 +492,7 @@ class BinaryExprNode extends ControlFlowNode {
} }
ControlFlowNode getRight() { ControlFlowNode getRight() {
exists(Py::BinaryExpr b | exists(BinaryExpr b |
this.getNode() = b and this.getNode() = b and
result.getNode() = b.getRight() and result.getNode() = b.getRight() and
result.getBasicBlock().dominates(this.getBasicBlock()) result.getBasicBlock().dominates(this.getBasicBlock())
@@ -500,11 +500,11 @@ class BinaryExprNode extends ControlFlowNode {
} }
/** Gets the operator of this binary expression node. */ /** Gets the operator of this binary expression node. */
Py::Operator getOp() { result = this.getNode().getOp() } Operator getOp() { result = this.getNode().getOp() }
/** Whether left and right are a pair of operands for this binary expression */ /** Whether left and right are a pair of operands for this binary expression */
predicate operands(ControlFlowNode left, Py::Operator op, ControlFlowNode right) { predicate operands(ControlFlowNode left, Operator op, ControlFlowNode right) {
exists(Py::BinaryExpr b, Py::Expr eleft, Py::Expr eright | exists(BinaryExpr b, Expr eleft, Expr eright |
this.getNode() = b and left.getNode() = eleft and right.getNode() = eright this.getNode() = b and left.getNode() = eleft and right.getNode() = eright
| |
eleft = b.getLeft() and eright = b.getRight() and op = b.getOp() eleft = b.getLeft() and eright = b.getRight() and op = b.getOp()
@@ -516,20 +516,20 @@ class BinaryExprNode extends ControlFlowNode {
/** A control flow node corresponding to a boolean shortcut (and/or) operation */ /** A control flow node corresponding to a boolean shortcut (and/or) operation */
class BoolExprNode extends ControlFlowNode { class BoolExprNode extends ControlFlowNode {
BoolExprNode() { toAst(this) instanceof Py::BoolExpr } BoolExprNode() { toAst(this) instanceof BoolExpr }
/** flow node corresponding to one of the operands of a boolean expression */ /** flow node corresponding to one of the operands of a boolean expression */
ControlFlowNode getAnOperand() { ControlFlowNode getAnOperand() {
exists(Py::BoolExpr b | this.getNode() = b and result.getNode() = b.getAValue()) and exists(BoolExpr b | this.getNode() = b and result.getNode() = b.getAValue()) and
this.getBasicBlock().dominates(result.getBasicBlock()) this.getBasicBlock().dominates(result.getBasicBlock())
} }
override Py::BoolExpr getNode() { result = super.getNode() } override BoolExpr getNode() { result = super.getNode() }
} }
/** A control flow node corresponding to a unary expression: (`+x`), (`-x`) or (`~x`) */ /** A control flow node corresponding to a unary expression: (`+x`), (`-x`) or (`~x`) */
class UnaryExprNode extends ControlFlowNode { class UnaryExprNode extends ControlFlowNode {
UnaryExprNode() { toAst(this) instanceof Py::UnaryExpr } UnaryExprNode() { toAst(this) instanceof UnaryExpr }
/** /**
* Gets flow node corresponding to the operand of a unary expression. * Gets flow node corresponding to the operand of a unary expression.
@@ -540,7 +540,7 @@ class UnaryExprNode extends ControlFlowNode {
*/ */
ControlFlowNode getOperand() { result = this.getAPredecessor() } ControlFlowNode getOperand() { result = this.getAPredecessor() }
override Py::UnaryExpr getNode() { result = super.getNode() } override UnaryExpr getNode() { result = super.getNode() }
override ControlFlowNode getAChild() { result = this.getAPredecessor() } override ControlFlowNode getAChild() { result = this.getAPredecessor() }
} }
@@ -555,22 +555,22 @@ class DefinitionNode extends ControlFlowNode {
cached cached
DefinitionNode() { DefinitionNode() {
Stages::AST::ref() and Stages::AST::ref() and
exists(Py::Assign a | this.getNode() = a.getATarget()) exists(Assign a | this.getNode() = a.getATarget())
or or
exists(Py::AssignExpr a | this.getNode() = a.getTarget()) exists(AssignExpr a | this.getNode() = a.getTarget())
or or
exists(Py::AnnAssign a | this.getNode() = a.getTarget() and exists(a.getValue())) exists(AnnAssign a | this.getNode() = a.getTarget() and exists(a.getValue()))
or or
exists(Py::Alias a | this.getNode() = a.getAsname()) exists(Alias a | this.getNode() = a.getAsname())
or or
augstore(_, this) augstore(_, this)
or or
// `x, y = 1, 2` where LHS is a combination of list or tuples // `x, y = 1, 2` where LHS is a combination of list or tuples
exists(Py::Assign a | this.getNode() = list_or_tuple_nested_element(a.getATarget())) exists(Assign a | this.getNode() = list_or_tuple_nested_element(a.getATarget()))
or or
exists(Py::For for | this.getNode() = for.getTarget()) exists(For for | this.getNode() = for.getTarget())
or or
exists(Py::Parameter param | this.getNode() = param.asName() and exists(param.getDefault())) exists(Parameter param | this.getNode() = param.asName() and exists(param.getDefault()))
} }
/** flow node corresponding to the value assigned for the definition corresponding to this flow node */ /** flow node corresponding to the value assigned for the definition corresponding to this flow node */
@@ -584,16 +584,16 @@ class DefinitionNode extends ControlFlowNode {
// since the default value for a parameter is evaluated in the same basic block as // since the default value for a parameter is evaluated in the same basic block as
// the function definition, but the parameter belongs to the basic block of the function, // the function definition, but the parameter belongs to the basic block of the function,
// there is no dominance relationship between the two. // there is no dominance relationship between the two.
exists(Py::Parameter param | this.getNode() = param.asName()) exists(Parameter param | this.getNode() = param.asName())
) )
} }
} }
private Py::Expr list_or_tuple_nested_element(Py::Expr list_or_tuple) { private Expr list_or_tuple_nested_element(Expr list_or_tuple) {
exists(Py::Expr elt | exists(Expr elt |
elt = list_or_tuple.(Py::Tuple).getAnElt() elt = list_or_tuple.(Tuple).getAnElt()
or or
elt = list_or_tuple.(Py::List).getAnElt() elt = list_or_tuple.(List).getAnElt()
| |
result = elt result = elt
or or
@@ -603,12 +603,12 @@ private Py::Expr list_or_tuple_nested_element(Py::Expr list_or_tuple) {
/** /**
* A control flow node corresponding to a deletion statement, such as `del x`. * A control flow node corresponding to a deletion statement, such as `del x`.
* There can be multiple `DeletionNode`s for each `Py::Delete` such that each * There can be multiple `DeletionNode`s for each `Delete` such that each
* target has own `DeletionNode`. The CFG for `del a, x.y` looks like: * target has own `DeletionNode`. The CFG for `del a, x.y` looks like:
* `NameNode('a') -> DeletionNode -> NameNode('b') -> AttrNode('y') -> DeletionNode`. * `NameNode('a') -> DeletionNode -> NameNode('b') -> AttrNode('y') -> DeletionNode`.
*/ */
class DeletionNode extends ControlFlowNode { class DeletionNode extends ControlFlowNode {
DeletionNode() { toAst(this) instanceof Py::Delete } DeletionNode() { toAst(this) instanceof Delete }
/** Gets the unique target of this deletion node. */ /** Gets the unique target of this deletion node. */
ControlFlowNode getTarget() { result.getASuccessor() = this } ControlFlowNode getTarget() { result.getASuccessor() = this }
@@ -617,9 +617,9 @@ class DeletionNode extends ControlFlowNode {
/** A control flow node corresponding to a sequence (tuple or list) literal */ /** A control flow node corresponding to a sequence (tuple or list) literal */
abstract class SequenceNode extends ControlFlowNode { abstract class SequenceNode extends ControlFlowNode {
SequenceNode() { SequenceNode() {
toAst(this) instanceof Py::Tuple toAst(this) instanceof Tuple
or or
toAst(this) instanceof Py::List toAst(this) instanceof List
} }
/** Gets the control flow node for an element of this sequence */ /** Gets the control flow node for an element of this sequence */
@@ -632,11 +632,11 @@ abstract class SequenceNode extends ControlFlowNode {
/** A control flow node corresponding to a tuple expression such as `( 1, 3, 5, 7, 9 )` */ /** A control flow node corresponding to a tuple expression such as `( 1, 3, 5, 7, 9 )` */
class TupleNode extends SequenceNode { class TupleNode extends SequenceNode {
TupleNode() { toAst(this) instanceof Py::Tuple } TupleNode() { toAst(this) instanceof Tuple }
override ControlFlowNode getElement(int n) { override ControlFlowNode getElement(int n) {
Stages::AST::ref() and Stages::AST::ref() and
exists(Py::Tuple t | this.getNode() = t and result.getNode() = t.getElt(n)) and exists(Tuple t | this.getNode() = t and result.getNode() = t.getElt(n)) and
( (
result.getBasicBlock().dominates(this.getBasicBlock()) result.getBasicBlock().dominates(this.getBasicBlock())
or or
@@ -647,10 +647,10 @@ class TupleNode extends SequenceNode {
/** A control flow node corresponding to a list expression, such as `[ 1, 3, 5, 7, 9 ]` */ /** A control flow node corresponding to a list expression, such as `[ 1, 3, 5, 7, 9 ]` */
class ListNode extends SequenceNode { class ListNode extends SequenceNode {
ListNode() { toAst(this) instanceof Py::List } ListNode() { toAst(this) instanceof List }
override ControlFlowNode getElement(int n) { override ControlFlowNode getElement(int n) {
exists(Py::List l | this.getNode() = l and result.getNode() = l.getElt(n)) and exists(List l | this.getNode() = l and result.getNode() = l.getElt(n)) and
( (
result.getBasicBlock().dominates(this.getBasicBlock()) result.getBasicBlock().dominates(this.getBasicBlock())
or or
@@ -661,10 +661,10 @@ class ListNode extends SequenceNode {
/** A control flow node corresponding to a set expression, such as `{ 1, 3, 5, 7, 9 }` */ /** A control flow node corresponding to a set expression, such as `{ 1, 3, 5, 7, 9 }` */
class SetNode extends ControlFlowNode { class SetNode extends ControlFlowNode {
SetNode() { toAst(this) instanceof Py::Set } SetNode() { toAst(this) instanceof Set }
ControlFlowNode getAnElement() { ControlFlowNode getAnElement() {
exists(Py::Set s | this.getNode() = s and result.getNode() = s.getElt(_)) and exists(Set s | this.getNode() = s and result.getNode() = s.getElt(_)) and
( (
result.getBasicBlock().dominates(this.getBasicBlock()) result.getBasicBlock().dominates(this.getBasicBlock())
or or
@@ -675,20 +675,20 @@ class SetNode extends ControlFlowNode {
/** A control flow node corresponding to a dictionary literal, such as `{ 'a': 1, 'b': 2 }` */ /** A control flow node corresponding to a dictionary literal, such as `{ 'a': 1, 'b': 2 }` */
class DictNode extends ControlFlowNode { class DictNode extends ControlFlowNode {
DictNode() { toAst(this) instanceof Py::Dict } DictNode() { toAst(this) instanceof Dict }
/** /**
* Gets a key of this dictionary literal node, for those items that have keys * Gets a key of this dictionary literal node, for those items that have keys
* E.g, in {'a':1, **b} this returns only 'a' * E.g, in {'a':1, **b} this returns only 'a'
*/ */
ControlFlowNode getAKey() { ControlFlowNode getAKey() {
exists(Py::Dict d | this.getNode() = d and result.getNode() = d.getAKey()) and exists(Dict d | this.getNode() = d and result.getNode() = d.getAKey()) and
result.getBasicBlock().dominates(this.getBasicBlock()) result.getBasicBlock().dominates(this.getBasicBlock())
} }
/** Gets a value of this dictionary literal node */ /** Gets a value of this dictionary literal node */
ControlFlowNode getAValue() { ControlFlowNode getAValue() {
exists(Py::Dict d | this.getNode() = d and result.getNode() = d.getAValue()) and exists(Dict d | this.getNode() = d and result.getNode() = d.getAValue()) and
result.getBasicBlock().dominates(this.getBasicBlock()) result.getBasicBlock().dominates(this.getBasicBlock())
} }
} }
@@ -712,23 +712,21 @@ class IterableNode extends ControlFlowNode {
} }
} }
private Py::AstNode assigned_value(Py::Expr lhs) { private AstNode assigned_value(Expr lhs) {
/* lhs = result */ /* lhs = result */
exists(Py::Assign a | a.getATarget() = lhs and result = a.getValue()) exists(Assign a | a.getATarget() = lhs and result = a.getValue())
or or
/* lhs := result */ /* lhs := result */
exists(Py::AssignExpr a | a.getTarget() = lhs and result = a.getValue()) exists(AssignExpr a | a.getTarget() = lhs and result = a.getValue())
or or
/* lhs : annotation = result */ /* lhs : annotation = result */
exists(Py::AnnAssign a | a.getTarget() = lhs and result = a.getValue()) exists(AnnAssign a | a.getTarget() = lhs and result = a.getValue())
or or
/* import result as lhs */ /* import result as lhs */
exists(Py::Alias a | a.getAsname() = lhs and result = a.getValue()) exists(Alias a | a.getAsname() = lhs and result = a.getValue())
or or
/* lhs += x => result = (lhs + x) */ /* lhs += x => result = (lhs + x) */
exists(Py::AugAssign a, Py::BinaryExpr b | exists(AugAssign a, BinaryExpr b | b = a.getOperation() and result = b and lhs = b.getLeft())
b = a.getOperation() and result = b and lhs = b.getLeft()
)
or or
/* /*
* ..., lhs, ... = ..., result, ... * ..., lhs, ... = ..., result, ...
@@ -736,31 +734,31 @@ private Py::AstNode assigned_value(Py::Expr lhs) {
* ..., (..., lhs, ...), ... = ..., (..., result, ...), ... * ..., (..., lhs, ...), ... = ..., (..., result, ...), ...
*/ */
exists(Py::Assign a | nested_sequence_assign(a.getATarget(), a.getValue(), lhs, result)) exists(Assign a | nested_sequence_assign(a.getATarget(), a.getValue(), lhs, result))
or or
/* for lhs in seq: => `result` is the `for` node, representing the `iter(next(seq))` operation. */ /* for lhs in seq: => `result` is the `for` node, representing the `iter(next(seq))` operation. */
result.(Py::For).getTarget() = lhs result.(For).getTarget() = lhs
or or
exists(Py::Parameter param | lhs = param.asName() and result = param.getDefault()) exists(Parameter param | lhs = param.asName() and result = param.getDefault())
} }
predicate nested_sequence_assign( predicate nested_sequence_assign(
Py::Expr left_parent, Py::Expr right_parent, Py::Expr left_result, Py::Expr right_result Expr left_parent, Expr right_parent, Expr left_result, Expr right_result
) { ) {
exists(Py::Assign a | exists(Assign a |
a.getATarget().getASubExpression*() = left_parent and a.getATarget().getASubExpression*() = left_parent and
a.getValue().getASubExpression*() = right_parent a.getValue().getASubExpression*() = right_parent
) and ) and
exists(int i, Py::Expr left_elem, Py::Expr right_elem | exists(int i, Expr left_elem, Expr right_elem |
( (
left_elem = left_parent.(Py::Tuple).getElt(i) left_elem = left_parent.(Tuple).getElt(i)
or or
left_elem = left_parent.(Py::List).getElt(i) left_elem = left_parent.(List).getElt(i)
) and ) and
( (
right_elem = right_parent.(Py::Tuple).getElt(i) right_elem = right_parent.(Tuple).getElt(i)
or or
right_elem = right_parent.(Py::List).getElt(i) right_elem = right_parent.(List).getElt(i)
) )
| |
left_result = left_elem and right_result = right_elem left_result = left_elem and right_result = right_elem
@@ -771,9 +769,9 @@ predicate nested_sequence_assign(
/** A flow node for a `for` statement. */ /** A flow node for a `for` statement. */
class ForNode extends ControlFlowNode { class ForNode extends ControlFlowNode {
ForNode() { toAst(this) instanceof Py::For } ForNode() { toAst(this) instanceof For }
override Py::For getNode() { result = super.getNode() } override For getNode() { result = super.getNode() }
/** Holds if this `for` statement causes iteration over `sequence` storing each step of the iteration in `target` */ /** Holds if this `for` statement causes iteration over `sequence` storing each step of the iteration in `target` */
predicate iterates(ControlFlowNode target, ControlFlowNode sequence) { predicate iterates(ControlFlowNode target, ControlFlowNode sequence) {
@@ -784,7 +782,7 @@ class ForNode extends ControlFlowNode {
/** Gets the sequence node for this `for` statement. */ /** Gets the sequence node for this `for` statement. */
ControlFlowNode getSequence() { ControlFlowNode getSequence() {
exists(Py::For for | exists(For for |
toAst(this) = for and toAst(this) = for and
for.getIter() = result.getNode() for.getIter() = result.getNode()
| |
@@ -794,7 +792,7 @@ class ForNode extends ControlFlowNode {
/** A possible `target` for this `for` statement, not accounting for loop unrolling */ /** A possible `target` for this `for` statement, not accounting for loop unrolling */
private ControlFlowNode possibleTarget() { private ControlFlowNode possibleTarget() {
exists(Py::For for | exists(For for |
toAst(this) = for and toAst(this) = for and
for.getTarget() = result.getNode() and for.getTarget() = result.getNode() and
this.getBasicBlock().dominates(result.getBasicBlock()) this.getBasicBlock().dominates(result.getBasicBlock())
@@ -811,11 +809,11 @@ class ForNode extends ControlFlowNode {
/** A flow node for a `raise` statement */ /** A flow node for a `raise` statement */
class RaiseStmtNode extends ControlFlowNode { class RaiseStmtNode extends ControlFlowNode {
RaiseStmtNode() { toAst(this) instanceof Py::Raise } RaiseStmtNode() { toAst(this) instanceof Raise }
/** Gets the control flow node for the exception raised by this raise statement */ /** Gets the control flow node for the exception raised by this raise statement */
ControlFlowNode getException() { ControlFlowNode getException() {
exists(Py::Raise r | exists(Raise r |
r = toAst(this) and r = toAst(this) and
r.getException() = toAst(result) and r.getException() = toAst(result) and
result.getBasicBlock().dominates(this.getBasicBlock()) result.getBasicBlock().dominates(this.getBasicBlock())
@@ -829,36 +827,36 @@ class RaiseStmtNode extends ControlFlowNode {
*/ */
class NameNode extends ControlFlowNode { class NameNode extends ControlFlowNode {
NameNode() { NameNode() {
exists(Py::Name n | py_flow_bb_node(this, n, _, _)) exists(Name n | py_flow_bb_node(this, n, _, _))
or or
exists(Py::PlaceHolder p | py_flow_bb_node(this, p, _, _)) exists(PlaceHolder p | py_flow_bb_node(this, p, _, _))
} }
/** Whether this flow node defines the variable `v`. */ /** Whether this flow node defines the variable `v`. */
predicate defines(Py::Variable v) { predicate defines(Variable v) {
exists(Py::Name d | this.getNode() = d and d.defines(v)) and exists(Name d | this.getNode() = d and d.defines(v)) and
not this.isLoad() not this.isLoad()
} }
/** Whether this flow node deletes the variable `v`. */ /** Whether this flow node deletes the variable `v`. */
predicate deletes(Py::Variable v) { exists(Py::Name d | this.getNode() = d and d.deletes(v)) } predicate deletes(Variable v) { exists(Name d | this.getNode() = d and d.deletes(v)) }
/** Whether this flow node uses the variable `v`. */ /** Whether this flow node uses the variable `v`. */
predicate uses(Py::Variable v) { predicate uses(Variable v) {
this.isLoad() and this.isLoad() and
exists(Py::Name u | this.getNode() = u and u.uses(v)) exists(Name u | this.getNode() = u and u.uses(v))
or or
exists(Py::PlaceHolder u | exists(PlaceHolder u |
this.getNode() = u and u.getVariable() = v and u.getCtx() instanceof Py::Load this.getNode() = u and u.getVariable() = v and u.getCtx() instanceof Load
) )
or or
Scopes::use_of_global_variable(this, v.getScope(), v.getId()) Scopes::use_of_global_variable(this, v.getScope(), v.getId())
} }
string getId() { string getId() {
result = this.getNode().(Py::Name).getId() result = this.getNode().(Name).getId()
or or
result = this.getNode().(Py::PlaceHolder).getId() result = this.getNode().(PlaceHolder).getId()
} }
/** Whether this is a use of a local variable. */ /** Whether this is a use of a local variable. */
@@ -870,39 +868,37 @@ class NameNode extends ControlFlowNode {
/** Whether this is a use of a global (including builtin) variable. */ /** Whether this is a use of a global (including builtin) variable. */
predicate isGlobal() { Scopes::use_of_global_variable(this, _, _) } predicate isGlobal() { Scopes::use_of_global_variable(this, _, _) }
predicate isSelf() { predicate isSelf() { exists(SsaVariable selfvar | selfvar.isSelf() and selfvar.getAUse() = this) }
exists(Py::SsaVariable selfvar | selfvar.isSelf() and selfvar.getAUse() = this)
}
} }
/** A control flow node corresponding to a named constant, one of `None`, `True` or `False`. */ /** A control flow node corresponding to a named constant, one of `None`, `True` or `False`. */
class NameConstantNode extends NameNode { class NameConstantNode extends NameNode {
NameConstantNode() { exists(Py::NameConstant n | py_flow_bb_node(this, n, _, _)) } NameConstantNode() { exists(NameConstant n | py_flow_bb_node(this, n, _, _)) }
/* /*
* We ought to override uses as well, but that has * We ought to override uses as well, but that has
* a serious performance impact. * a serious performance impact.
* deprecated predicate uses(Py::Variable v) { none() } * deprecated predicate uses(Variable v) { none() }
*/ */
} }
/** A control flow node corresponding to a starred expression, `*a`. */ /** A control flow node corresponding to a starred expression, `*a`. */
class StarredNode extends ControlFlowNode { class StarredNode extends ControlFlowNode {
StarredNode() { toAst(this) instanceof Py::Starred } StarredNode() { toAst(this) instanceof Starred }
ControlFlowNode getValue() { toAst(result) = toAst(this).(Py::Starred).getValue() } ControlFlowNode getValue() { toAst(result) = toAst(this).(Starred).getValue() }
} }
/** The ControlFlowNode for an 'except' statement. */ /** The ControlFlowNode for an 'except' statement. */
class ExceptFlowNode extends ControlFlowNode { class ExceptFlowNode extends ControlFlowNode {
ExceptFlowNode() { this.getNode() instanceof Py::ExceptStmt } ExceptFlowNode() { this.getNode() instanceof ExceptStmt }
/** /**
* Gets the type handled by this exception handler. * Gets the type handled by this exception handler.
* `Py::ExceptionType` in `except Py::ExceptionType as e:` * `ExceptionType` in `except ExceptionType as e:`
*/ */
ControlFlowNode getType() { ControlFlowNode getType() {
exists(Py::ExceptStmt ex | exists(ExceptStmt ex |
this.getBasicBlock().dominates(result.getBasicBlock()) and this.getBasicBlock().dominates(result.getBasicBlock()) and
ex = this.getNode() and ex = this.getNode() and
result.getNode() = ex.getType() result.getNode() = ex.getType()
@@ -911,10 +907,10 @@ class ExceptFlowNode extends ControlFlowNode {
/** /**
* Gets the name assigned to the handled exception, if any. * Gets the name assigned to the handled exception, if any.
* `e` in `except Py::ExceptionType as e:` * `e` in `except ExceptionType as e:`
*/ */
ControlFlowNode getName() { ControlFlowNode getName() {
exists(Py::ExceptStmt ex | exists(ExceptStmt ex |
this.getBasicBlock().dominates(result.getBasicBlock()) and this.getBasicBlock().dominates(result.getBasicBlock()) and
ex = this.getNode() and ex = this.getNode() and
result.getNode() = ex.getName() result.getNode() = ex.getName()
@@ -924,30 +920,30 @@ class ExceptFlowNode extends ControlFlowNode {
/** The ControlFlowNode for an 'except*' statement. */ /** The ControlFlowNode for an 'except*' statement. */
class ExceptGroupFlowNode extends ControlFlowNode { class ExceptGroupFlowNode extends ControlFlowNode {
ExceptGroupFlowNode() { this.getNode() instanceof Py::ExceptGroupStmt } ExceptGroupFlowNode() { this.getNode() instanceof ExceptGroupStmt }
/** /**
* Gets the type handled by this exception handler. * Gets the type handled by this exception handler.
* `Py::ExceptionType` in `except* Py::ExceptionType as e:` * `ExceptionType` in `except* ExceptionType as e:`
*/ */
ControlFlowNode getType() { ControlFlowNode getType() {
this.getBasicBlock().dominates(result.getBasicBlock()) and this.getBasicBlock().dominates(result.getBasicBlock()) and
result.getNode() = this.getNode().(Py::ExceptGroupStmt).getType() result.getNode() = this.getNode().(ExceptGroupStmt).getType()
} }
/** /**
* Gets the name assigned to the handled exception, if any. * Gets the name assigned to the handled exception, if any.
* `e` in `except* Py::ExceptionType as e:` * `e` in `except* ExceptionType as e:`
*/ */
ControlFlowNode getName() { ControlFlowNode getName() {
this.getBasicBlock().dominates(result.getBasicBlock()) and this.getBasicBlock().dominates(result.getBasicBlock()) and
result.getNode() = this.getNode().(Py::ExceptGroupStmt).getName() result.getNode() = this.getNode().(ExceptGroupStmt).getName()
} }
} }
private module Scopes { private module Scopes {
private predicate fast_local(NameNode n) { private predicate fast_local(NameNode n) {
exists(Py::FastLocalVariable v | exists(FastLocalVariable v |
n.uses(v) and n.uses(v) and
v.getScope() = n.getScope() v.getScope() = n.getScope()
) )
@@ -956,15 +952,15 @@ private module Scopes {
predicate local(NameNode n) { predicate local(NameNode n) {
fast_local(n) fast_local(n)
or or
exists(Py::SsaVariable var | exists(SsaVariable var |
var.getAUse() = n and var.getAUse() = n and
n.getScope() instanceof Py::Class and n.getScope() instanceof Class and
exists(var.getDefinition()) exists(var.getDefinition())
) )
} }
predicate non_local(NameNode n) { predicate non_local(NameNode n) {
exists(Py::FastLocalVariable flv | exists(FastLocalVariable flv |
flv.getALoad() = n.getNode() and flv.getALoad() = n.getNode() and
not flv.getScope() = n.getScope() not flv.getScope() = n.getScope()
) )
@@ -972,20 +968,20 @@ private module Scopes {
// magic is fine, but we get questionable join-ordering of it // magic is fine, but we get questionable join-ordering of it
pragma[nomagic] pragma[nomagic]
predicate use_of_global_variable(NameNode n, Py::Module scope, string name) { predicate use_of_global_variable(NameNode n, Module scope, string name) {
n.isLoad() and n.isLoad() and
not non_local(n) and not non_local(n) and
not exists(Py::SsaVariable var | var.getAUse() = n | not exists(SsaVariable var | var.getAUse() = n |
var.getVariable() instanceof Py::FastLocalVariable var.getVariable() instanceof FastLocalVariable
or or
n.getScope() instanceof Py::Class and n.getScope() instanceof Class and
not maybe_undefined(var) not maybe_undefined(var)
) and ) and
name = n.getId() and name = n.getId() and
scope = n.getEnclosingModule() scope = n.getEnclosingModule()
} }
private predicate maybe_undefined(Py::SsaVariable var) { private predicate maybe_undefined(SsaVariable var) {
not exists(var.getDefinition()) and not py_ssa_phi(var, _) not exists(var.getDefinition()) and not py_ssa_phi(var, _)
or or
var.getDefinition().isDelete() var.getDefinition().isDelete()
@@ -1062,13 +1058,13 @@ class BasicBlock extends @py_flow_node {
private predicate oneNodeBlock() { this.firstNode() = this.getLastNode() } private predicate oneNodeBlock() { this.firstNode() = this.getLastNode() }
private predicate startLocationInfo(string file, int line, int col) { private predicate startLocationInfo(string file, int line, int col) {
if this.firstNode().getNode() instanceof Py::Scope if this.firstNode().getNode() instanceof Scope
then this.firstNode().getASuccessor().getLocation().hasLocationInfo(file, line, col, _, _) then this.firstNode().getASuccessor().getLocation().hasLocationInfo(file, line, col, _, _)
else this.firstNode().getLocation().hasLocationInfo(file, line, col, _, _) else this.firstNode().getLocation().hasLocationInfo(file, line, col, _, _)
} }
private predicate endLocationInfo(int endl, int endc) { private predicate endLocationInfo(int endl, int endc) {
if this.getLastNode().getNode() instanceof Py::Scope and not this.oneNodeBlock() if this.getLastNode().getNode() instanceof Scope and not this.oneNodeBlock()
then this.getLastNode().getAPredecessor().getLocation().hasLocationInfo(_, _, _, endl, endc) then this.getLastNode().getAPredecessor().getLocation().hasLocationInfo(_, _, _, endl, endc)
else this.getLastNode().getLocation().hasLocationInfo(_, _, _, endl, endc) else this.getLastNode().getLocation().hasLocationInfo(_, _, _, endl, endc)
} }
@@ -1085,7 +1081,7 @@ class BasicBlock extends @py_flow_node {
/** Whether flow from this basic block reaches a normal exit from its scope */ /** Whether flow from this basic block reaches a normal exit from its scope */
predicate reachesExit() { predicate reachesExit() {
exists(Py::Scope s | s.getANormalExit().getBasicBlock() = this) exists(Scope s | s.getANormalExit().getBasicBlock() = this)
or or
this.getASuccessor().reachesExit() this.getASuccessor().reachesExit()
} }
@@ -1126,7 +1122,7 @@ class BasicBlock extends @py_flow_node {
/** Gets the scope of this block */ /** Gets the scope of this block */
pragma[nomagic] pragma[nomagic]
Py::Scope getScope() { Scope getScope() {
exists(ControlFlowNode n | n.getBasicBlock() = this | exists(ControlFlowNode n | n.getBasicBlock() = this |
/* Take care not to use an entry or exit node as that node's scope will be the outer scope */ /* Take care not to use an entry or exit node as that node's scope will be the outer scope */
not py_scope_flow(n, _, -1) and not py_scope_flow(n, _, -1) and
@@ -1149,17 +1145,17 @@ class BasicBlock extends @py_flow_node {
predicate reaches(BasicBlock other) { this = other or this.strictlyReaches(other) } predicate reaches(BasicBlock other) { this = other or this.strictlyReaches(other) }
/** /**
* Gets the `Py::ConditionBlock`, if any, that controls this block and * Gets the `ConditionBlock`, if any, that controls this block and
* does not control any other `Py::ConditionBlock`s that control this block. * does not control any other `ConditionBlock`s that control this block.
* That is the `Py::ConditionBlock` that is closest dominator. * That is the `ConditionBlock` that is closest dominator.
*/ */
Py::ConditionBlock getImmediatelyControllingBlock() { ConditionBlock getImmediatelyControllingBlock() {
result = this.nonControllingImmediateDominator*().getImmediateDominator() result = this.nonControllingImmediateDominator*().getImmediateDominator()
} }
private BasicBlock nonControllingImmediateDominator() { private BasicBlock nonControllingImmediateDominator() {
result = this.getImmediateDominator() and result = this.getImmediateDominator() and
not result.(Py::ConditionBlock).controls(this, _) not result.(ConditionBlock).controls(this, _)
} }
/** /**
@@ -1179,7 +1175,7 @@ private class ControlFlowNodeAlias = ControlFlowNode;
final private class FinalBasicBlock = BasicBlock; final private class FinalBasicBlock = BasicBlock;
module Cfg implements BB::CfgSig<Py::Location> { module Cfg implements BB::CfgSig<Location> {
private import codeql.controlflow.SuccessorType private import codeql.controlflow.SuccessorType
class ControlFlowNode = ControlFlowNodeAlias; class ControlFlowNode = ControlFlowNodeAlias;
@@ -1190,7 +1186,7 @@ module Cfg implements BB::CfgSig<Py::Location> {
// Using the location of the first node is simple // Using the location of the first node is simple
// and we just need a way to identify the basic block // and we just need a way to identify the basic block
// during debugging, so this will be serviceable. // during debugging, so this will be serviceable.
Py::Location getLocation() { result = super.getNode(0).getLocation() } Location getLocation() { result = super.getNode(0).getLocation() }
int length() { result = count(int i | exists(this.getNode(i))) } int length() { result = count(int i | exists(this.getNode(i))) }

View File

@@ -2,7 +2,6 @@ overlay[local]
module; module;
import python import python
private import semmle.python.controlflow.internal.Cfg as Cfg
/** /**
* A function, independent of defaults and binding. * A function, independent of defaults and binding.
@@ -158,12 +157,12 @@ class Function extends Function_, Scope, AstNode {
* DEPRECATED: bind a `Return` node explicitly instead, e.g. * DEPRECATED: bind a `Return` node explicitly instead, e.g.
* `exists(Return ret | ret.getScope() = this and n.getNode() = ret.getValue())`. * `exists(Return ret | ret.getScope() = this and n.getNode() = ret.getValue())`.
* This API is being phased out together with `AstNode.getAFlowNode()` to * This API is being phased out together with `AstNode.getAFlowNode()` to
* untangle the AST and CFG hierarchies. * untangle the AST and CFG hierarchies in preparation for migrating the
* dataflow library off the legacy CFG.
* *
* Gets a control flow node for a return value of this function, from the * Gets a control flow node for a return value of this function.
* new (shared) CFG.
*/ */
deprecated Cfg::ControlFlowNode getAReturnValueFlowNode() { deprecated ControlFlowNode getAReturnValueFlowNode() {
exists(Return ret | exists(Return ret |
ret.getScope() = this and ret.getScope() = this and
ret.getValue() = result.getNode() ret.getValue() = result.getNode()

View File

@@ -4,7 +4,6 @@ module;
import python import python
private import semmle.python.types.Builtins private import semmle.python.types.Builtins
private import semmle.python.internal.CachedStages private import semmle.python.internal.CachedStages
private import semmle.python.controlflow.internal.Cfg as Cfg
/** /**
* An alias in an import statement, the `mod as name` part of `import mod as name`. May be artificial; * An alias in an import statement, the `mod as name` part of `import mod as name`. May be artificial;
@@ -164,7 +163,7 @@ class ImportMember extends ImportMember_ {
result = this.getModule().(ImportExpr).getImportedModuleName() + "." + this.getName() result = this.getModule().(ImportExpr).getImportedModuleName() + "." + this.getName()
} }
deprecated override Cfg::ImportMemberNode getAFlowNode() { result = super.getAFlowNode() } deprecated override ImportMemberNode getAFlowNode() { result = super.getAFlowNode() }
} }
/** An import statement */ /** An import statement */

File diff suppressed because it is too large Load Diff

View File

@@ -1,12 +1,11 @@
/** Provides commonly used BarrierGuards. */ /** Provides commonly used BarrierGuards. */
private import python private import python
private import semmle.python.controlflow.internal.Cfg as Cfg
private import semmle.python.dataflow.new.DataFlow private import semmle.python.dataflow.new.DataFlow
private predicate constCompare(DataFlow::GuardNode g, Cfg::ControlFlowNode node, boolean branch) { private predicate constCompare(DataFlow::GuardNode g, ControlFlowNode node, boolean branch) {
exists(Cfg::CompareNode cn | cn = g | exists(CompareNode cn | cn = g |
exists(ImmutableLiteral const, Cmpop op, Cfg::ControlFlowNode c | exists(ImmutableLiteral const, Cmpop op, ControlFlowNode c |
c.getNode() = const and c.getNode() = const and
( (
op = any(Eq eq) and branch = true op = any(Eq eq) and branch = true
@@ -19,7 +18,7 @@ private predicate constCompare(DataFlow::GuardNode g, Cfg::ControlFlowNode node,
cn.operands(node, op, c) cn.operands(node, op, c)
) )
or or
exists(NameConstant const, Cmpop op, Cfg::ControlFlowNode c | exists(NameConstant const, Cmpop op, ControlFlowNode c |
c.getNode() = const and c.getNode() = const and
( (
op = any(Is is_) and branch = true op = any(Is is_) and branch = true
@@ -32,12 +31,12 @@ private predicate constCompare(DataFlow::GuardNode g, Cfg::ControlFlowNode node,
cn.operands(node, op, c) cn.operands(node, op, c)
) )
or or
exists(Cfg::IterableNode const_iterable, Cmpop op | exists(IterableNode const_iterable, Cmpop op |
op = any(In in_) and branch = true op = any(In in_) and branch = true
or or
op = any(NotIn ni) and branch = false op = any(NotIn ni) and branch = false
| |
forall(Cfg::ControlFlowNode elem | elem = const_iterable.getAnElement() | forall(ControlFlowNode elem | elem = const_iterable.getAnElement() |
elem.getNode() instanceof ImmutableLiteral elem.getNode() instanceof ImmutableLiteral
) and ) and
cn.operands(node, op, const_iterable) cn.operands(node, op, const_iterable)

View File

@@ -4,7 +4,6 @@
*/ */
private import python private import python
private import semmle.python.controlflow.internal.Cfg as Cfg
private import semmle.python.dataflow.new.DataFlow private import semmle.python.dataflow.new.DataFlow
// Need to import `semmle.python.Frameworks` since frameworks can extend `SensitiveDataSource::Range` // Need to import `semmle.python.Frameworks` since frameworks can extend `SensitiveDataSource::Range`
private import semmle.python.Frameworks private import semmle.python.Frameworks
@@ -106,7 +105,7 @@ private module SensitiveDataModeling {
or or
// to cover functions that we don't have the definition for, and where the // to cover functions that we don't have the definition for, and where the
// reference to the function has not already been marked as being sensitive // reference to the function has not already been marked as being sensitive
this.getFunction().asCfgNode().(Cfg::NameNode).getId() = sensitiveString(classification) this.getFunction().asCfgNode().(NameNode).getId() = sensitiveString(classification)
} }
override SensitiveDataClassification getClassification() { result = classification } override SensitiveDataClassification getClassification() { result = classification }
@@ -252,12 +251,12 @@ private module SensitiveDataModeling {
SensitiveDataClassification classification; SensitiveDataClassification classification;
SensitiveVariableAssignment() { SensitiveVariableAssignment() {
exists(Cfg::DefinitionNode def | exists(DefinitionNode def |
def.(Cfg::NameNode).getId() = sensitiveString(classification) and def.(NameNode).getId() = sensitiveString(classification) and
( (
this.asCfgNode() = def.getValue() this.asCfgNode() = def.getValue()
or or
this.asCfgNode() = def.getValue().(Cfg::ForNode).getSequence() this.asCfgNode() = def.getValue().(ForNode).getSequence()
) and ) and
not this.asExpr() instanceof FunctionExpr and not this.asExpr() instanceof FunctionExpr and
not this.asExpr() instanceof ClassExpr not this.asExpr() instanceof ClassExpr
@@ -294,7 +293,7 @@ private module SensitiveDataModeling {
SensitiveDataClassification classification; SensitiveDataClassification classification;
SensitiveSubscript() { SensitiveSubscript() {
this.asCfgNode().(Cfg::SubscriptNode).getIndex() = this.asCfgNode().(SubscriptNode).getIndex() =
sensitiveLookupStringConst(classification).asCfgNode() sensitiveLookupStringConst(classification).asCfgNode()
} }

View File

@@ -3,7 +3,6 @@ overlay[local]
module; module;
private import python private import python
private import semmle.python.controlflow.internal.Cfg as Cfg
import DataFlowUtil import DataFlowUtil
import DataFlowPublic import DataFlowPublic
private import DataFlowPrivate private import DataFlowPrivate
@@ -84,9 +83,9 @@ abstract class AttrWrite extends AttrRef {
* ```python * ```python
* object.attr = value * object.attr = value
* ``` * ```
* Also gives access to the `value` being written, by extending `Cfg::DefinitionNode`. * Also gives access to the `value` being written, by extending `DefinitionNode`.
*/ */
private class AttributeAssignmentNode extends Cfg::DefinitionNode, Cfg::AttrNode { } private class AttributeAssignmentNode extends DefinitionNode, AttrNode { }
/** A simple attribute assignment: `object.attr = value`. */ /** A simple attribute assignment: `object.attr = value`. */
private class AttributeAssignmentAsAttrWrite extends AttrWrite, CfgNode { private class AttributeAssignmentAsAttrWrite extends AttrWrite, CfgNode {
@@ -132,13 +131,13 @@ private class GlobalAttributeAssignmentAsAttrWrite extends AttrWrite, CfgNode {
override string getAttributeName() { result = node.getName() } override string getAttributeName() { result = node.getName() }
} }
/** Represents `Cfg::CallNode`s that may refer to calls to built-in functions or classes. */ /** Represents `CallNode`s that may refer to calls to built-in functions or classes. */
private class BuiltInCallNode extends Cfg::CallNode { private class BuiltInCallNode extends CallNode {
string name; string name;
BuiltInCallNode() { BuiltInCallNode() {
// TODO disallow instances where the name of the built-in may refer to an in-scope variable of that name. // TODO disallow instances where the name of the built-in may refer to an in-scope variable of that name.
exists(Cfg::NameNode id | exists(NameNode id |
name = Builtins::getBuiltinName() and name = Builtins::getBuiltinName() and
this.getFunction() = id and this.getFunction() = id and
id.getId() = name and id.getId() = name and
@@ -146,7 +145,7 @@ private class BuiltInCallNode extends Cfg::CallNode {
) )
} }
/** Gets the name of the built-in function that is called at this `Cfg::CallNode` */ /** Gets the name of the built-in function that is called at this `CallNode` */
string getBuiltinName() { result = name } string getBuiltinName() { result = name }
} }
@@ -158,20 +157,20 @@ private class BuiltinAttrCallNode extends BuiltInCallNode {
BuiltinAttrCallNode() { name in ["setattr", "getattr", "hasattr", "delattr"] } BuiltinAttrCallNode() { name in ["setattr", "getattr", "hasattr", "delattr"] }
/** Gets the control flow node for object on which the attribute is accessed. */ /** Gets the control flow node for object on which the attribute is accessed. */
Cfg::ControlFlowNode getObject() { result in [this.getArg(0), this.getArgByName("object")] } ControlFlowNode getObject() { result in [this.getArg(0), this.getArgByName("object")] }
/** /**
* Gets the control flow node for the value that is being written to the attribute. * Gets the control flow node for the value that is being written to the attribute.
* Only relevant for `setattr` calls. * Only relevant for `setattr` calls.
*/ */
Cfg::ControlFlowNode getValue() { ControlFlowNode getValue() {
// only valid for `setattr` // only valid for `setattr`
name = "setattr" and name = "setattr" and
result in [this.getArg(2), this.getArgByName("value")] result in [this.getArg(2), this.getArgByName("value")]
} }
/** Gets the control flow node that defines the name of the attribute being accessed. */ /** Gets the control flow node that defines the name of the attribute being accessed. */
Cfg::ControlFlowNode getName() { result in [this.getArg(1), this.getArgByName("name")] } ControlFlowNode getName() { result in [this.getArg(1), this.getArgByName("name")] }
} }
/** Represents calls to the built-in `setattr`. */ /** Represents calls to the built-in `setattr`. */
@@ -206,10 +205,10 @@ private class SetAttrCallAsAttrWrite extends AttrWrite, CfgNode {
* attr = value * attr = value
* ... * ...
* ``` * ```
* Instances of this class correspond to the `Cfg::NameNode` for `attr`, and also gives access to `value` by * Instances of this class correspond to the `NameNode` for `attr`, and also gives access to `value` by
* virtue of being a `Cfg::DefinitionNode`. * virtue of being a `DefinitionNode`.
*/ */
private class ClassAttributeAssignmentNode extends Cfg::DefinitionNode, Cfg::NameNode { private class ClassAttributeAssignmentNode extends DefinitionNode, NameNode {
ClassAttributeAssignmentNode() { this.getScope() = any(ClassExpr c).getInnerScope() } ClassAttributeAssignmentNode() { this.getScope() = any(ClassExpr c).getInnerScope() }
} }
@@ -249,7 +248,7 @@ abstract class AttrRead extends AttrRef, Node, LocalSourceNode {
/** A simple attribute read, e.g. `object.attr` */ /** A simple attribute read, e.g. `object.attr` */
private class AttributeReadAsAttrRead extends AttrRead, CfgNode { private class AttributeReadAsAttrRead extends AttrRead, CfgNode {
override Cfg::AttrNode node; override AttrNode node;
AttributeReadAsAttrRead() { node.isLoad() } AttributeReadAsAttrRead() { node.isLoad() }
@@ -286,7 +285,7 @@ private class GetAttrCallAsAttrRead extends AttrRead, CfgNode {
* is treated as if it is a read of the attribute `module.attr`, even if `module` is not imported directly. * is treated as if it is a read of the attribute `module.attr`, even if `module` is not imported directly.
*/ */
private class ModuleAttributeImportAsAttrRead extends AttrRead, CfgNode { private class ModuleAttributeImportAsAttrRead extends AttrRead, CfgNode {
override Cfg::ImportMemberNode node; override ImportMemberNode node;
override Node getObject() { result.asCfgNode() = node.getModule(_) } override Node getObject() { result.asCfgNode() = node.getModule(_) }

View File

@@ -3,7 +3,6 @@ overlay[local]
module; module;
private import python private import python
private import semmle.python.controlflow.internal.Cfg as Cfg
private import semmle.python.dataflow.new.DataFlow private import semmle.python.dataflow.new.DataFlow
private import semmle.python.dataflow.new.internal.ImportStar private import semmle.python.dataflow.new.internal.ImportStar
@@ -68,7 +67,7 @@ module Builtins {
DataFlow::CfgNode likelyBuiltin(string name) { DataFlow::CfgNode likelyBuiltin(string name) {
exists(Module m | exists(Module m |
result.getNode() = result.getNode() =
any(Cfg::NameNode n | any(NameNode n |
possible_builtin_accessed_in_module(n, name, m) and possible_builtin_accessed_in_module(n, name, m) and
not possible_builtin_defined_in_module(name, m) not possible_builtin_defined_in_module(name, m)
) )
@@ -88,7 +87,7 @@ module Builtins {
* Holds if `n` is an access of a global variable called `name` (which is also the name of a * Holds if `n` is an access of a global variable called `name` (which is also the name of a
* built-in) inside the module `m`. * built-in) inside the module `m`.
*/ */
private predicate possible_builtin_accessed_in_module(Cfg::NameNode n, string name, Module m) { private predicate possible_builtin_accessed_in_module(NameNode n, string name, Module m) {
n.isGlobal() and n.isGlobal() and
n.isLoad() and n.isLoad() and
name = n.getId() and name = n.getId() and

View File

@@ -25,7 +25,7 @@
* what callable this call might end up targeting. * what callable this call might end up targeting.
* *
* Specifically this means that we cannot use type-backtrackers from the function of a * Specifically this means that we cannot use type-backtrackers from the function of a
* `Cfg::CallNode`, since there is no `Cfg::CallNode` to backtrack from for `func` in the example * `CallNode`, since there is no `CallNode` to backtrack from for `func` in the example
* above. * above.
* *
* Note: This hasn't been 100% realized yet, so we don't currently expose a predicate to * Note: This hasn't been 100% realized yet, so we don't currently expose a predicate to
@@ -35,7 +35,6 @@ overlay[local?]
module; module;
private import python private import python
private import semmle.python.controlflow.internal.Cfg as Cfg
private import DataFlowPublic private import DataFlowPublic
private import DataFlowPrivate private import DataFlowPrivate
private import FlowSummaryImpl as FlowSummaryImpl private import FlowSummaryImpl as FlowSummaryImpl
@@ -163,7 +162,7 @@ newtype TArgumentPosition =
*/ */
TLambdaSelfArgumentPosition() or TLambdaSelfArgumentPosition() or
TPositionalArgumentPosition(int index) { TPositionalArgumentPosition(int index) {
exists(any(Cfg::CallNode c).getArg(index)) exists(any(CallNode c).getArg(index))
or or
// since synthetic calls within a summarized callable could use a unique argument // since synthetic calls within a summarized callable could use a unique argument
// position, we need to ensure we make these available (these are specified as // position, we need to ensure we make these available (these are specified as
@@ -175,7 +174,7 @@ newtype TArgumentPosition =
index = 0 index = 0
} or } or
TKeywordArgumentPosition(string name) { TKeywordArgumentPosition(string name) {
exists(any(Cfg::CallNode c).getArgByName(name)) exists(any(CallNode c).getArgByName(name))
or or
// see comment for TPositionalArgumentPosition // see comment for TPositionalArgumentPosition
FlowSummaryImpl::ParsePositions::isParsedKeywordParameterPosition(_, name) FlowSummaryImpl::ParsePositions::isParsedKeywordParameterPosition(_, name)
@@ -298,12 +297,10 @@ predicate hasPropertyDecorator(Function func) {
*/ */
overlay[local] overlay[local]
predicate hasContextmanagerDecorator(Function func) { predicate hasContextmanagerDecorator(Function func) {
exists(Cfg::ControlFlowNode contextmanager | exists(ControlFlowNode contextmanager |
contextmanager.(Cfg::NameNode).getId() = "contextmanager" and contextmanager.(NameNode).getId() = "contextmanager" and contextmanager.(NameNode).isGlobal()
contextmanager.(Cfg::NameNode).isGlobal()
or or
contextmanager.(Cfg::AttrNode).getObject("contextmanager").(Cfg::NameNode).getId() = contextmanager.(AttrNode).getObject("contextmanager").(NameNode).getId() = "contextlib"
"contextlib"
| |
func.getADecorator() = contextmanager.getNode() func.getADecorator() = contextmanager.getNode()
) )
@@ -319,10 +316,10 @@ predicate hasContextmanagerDecorator(Function func) {
*/ */
overlay[local] overlay[local]
private predicate hasOverloadDecorator(Function func) { private predicate hasOverloadDecorator(Function func) {
exists(Cfg::ControlFlowNode overload | exists(ControlFlowNode overload |
overload.(Cfg::NameNode).getId() = "overload" and overload.(Cfg::NameNode).isGlobal() overload.(NameNode).getId() = "overload" and overload.(NameNode).isGlobal()
or or
overload.(Cfg::AttrNode).getObject("overload").(Cfg::NameNode).isGlobal() overload.(AttrNode).getObject("overload").(NameNode).isGlobal()
| |
func.getADecorator() = overload.getNode() func.getADecorator() = overload.getNode()
) )
@@ -541,7 +538,7 @@ class LibraryCallableValue extends DataFlowCallable, TLibraryCallable {
// ============================================================================= // =============================================================================
/** Gets a call to `type`. */ /** Gets a call to `type`. */
private CallCfgNode getTypeCall() { private CallCfgNode getTypeCall() {
exists(Cfg::NameNode id | id.getId() = "type" and id.isGlobal() | exists(NameNode id | id.getId() = "type" and id.isGlobal() |
result.getFunction().asCfgNode() = id result.getFunction().asCfgNode() = id
) )
} }
@@ -553,7 +550,7 @@ private CallCfgNode getSuperCall() {
// link below), but otherwise only 2 edgecases. Overall it seems ok to ignore this complexity. // link below), but otherwise only 2 edgecases. Overall it seems ok to ignore this complexity.
// //
// https://github.com/python/cpython/blob/18b1782192f85bd26db89f5bc850f8bee4247c1a/Lib/unittest/mock.py#L48-L50 // https://github.com/python/cpython/blob/18b1782192f85bd26db89f5bc850f8bee4247c1a/Lib/unittest/mock.py#L48-L50
exists(Cfg::NameNode id | id.getId() = "super" and id.isGlobal() | exists(NameNode id | id.getId() = "super" and id.isGlobal() |
result.getFunction().asCfgNode() = id result.getFunction().asCfgNode() = id
) )
} }
@@ -1039,7 +1036,7 @@ private module MethodCalls {
*/ */
pragma[nomagic] pragma[nomagic]
private predicate directCall( private predicate directCall(
Cfg::CallNode call, Function target, string functionName, Class cls, AttrRead attr, Node self CallNode call, Function target, string functionName, Class cls, AttrRead attr, Node self
) { ) {
target = findFunctionAccordingToMroKnownStartingClass(cls, functionName) and target = findFunctionAccordingToMroKnownStartingClass(cls, functionName) and
directCall_join(call, functionName, cls, attr, self) directCall_join(call, functionName, cls, attr, self)
@@ -1048,7 +1045,7 @@ private module MethodCalls {
/** Extracted to give good join order */ /** Extracted to give good join order */
pragma[nomagic] pragma[nomagic]
private predicate directCall_join( private predicate directCall_join(
Cfg::CallNode call, string functionName, Class cls, AttrRead attr, Node self CallNode call, string functionName, Class cls, AttrRead attr, Node self
) { ) {
call.getFunction() = attrReadTracker(attr).asCfgNode() and call.getFunction() = attrReadTracker(attr).asCfgNode() and
attr.accesses(self, functionName) and attr.accesses(self, functionName) and
@@ -1065,7 +1062,7 @@ private module MethodCalls {
*/ */
pragma[nomagic] pragma[nomagic]
private predicate callWithinMethodImplicitSelfOrCls( private predicate callWithinMethodImplicitSelfOrCls(
Cfg::CallNode call, Function target, string functionName, Class classWithMethod, AttrRead attr, CallNode call, Function target, string functionName, Class classWithMethod, AttrRead attr,
Node self Node self
) { ) {
target = findFunctionAccordingToMro(getADirectSubclass*(classWithMethod), functionName) and target = findFunctionAccordingToMro(getADirectSubclass*(classWithMethod), functionName) and
@@ -1075,7 +1072,7 @@ private module MethodCalls {
/** Extracted to give good join order */ /** Extracted to give good join order */
pragma[nomagic] pragma[nomagic]
private predicate callWithinMethodImplicitSelfOrCls_join( private predicate callWithinMethodImplicitSelfOrCls_join(
Cfg::CallNode call, string functionName, Class classWithMethod, AttrRead attr, Node self CallNode call, string functionName, Class classWithMethod, AttrRead attr, Node self
) { ) {
call.getFunction() = attrReadTracker(attr).asCfgNode() and call.getFunction() = attrReadTracker(attr).asCfgNode() and
attr.accesses(self, functionName) and attr.accesses(self, functionName) and
@@ -1087,7 +1084,7 @@ private module MethodCalls {
* resolve the call to a known target (since the only super class might be the * resolve the call to a known target (since the only super class might be the
* builtin `object`, so we never have the implementation of `__new__` in the DB). * builtin `object`, so we never have the implementation of `__new__` in the DB).
*/ */
predicate fromSuperNewCall(Cfg::CallNode call, Class classUsedInSuper, AttrRead attr, Node self) { predicate fromSuperNewCall(CallNode call, Class classUsedInSuper, AttrRead attr, Node self) {
fromSuper_join(call, "__new__", classUsedInSuper, attr, self) and fromSuper_join(call, "__new__", classUsedInSuper, attr, self) and
self in [classTracker(_), clsArgumentTracker(_)] self in [classTracker(_), clsArgumentTracker(_)]
} }
@@ -1109,7 +1106,7 @@ private module MethodCalls {
*/ */
pragma[nomagic] pragma[nomagic]
predicate fromSuper( predicate fromSuper(
Cfg::CallNode call, Function target, string functionName, Class classUsedInSuper, AttrRead attr, CallNode call, Function target, string functionName, Class classUsedInSuper, AttrRead attr,
Node self Node self
) { ) {
target = findFunctionAccordingToMro(getNextClassInMro(classUsedInSuper), functionName) and target = findFunctionAccordingToMro(getNextClassInMro(classUsedInSuper), functionName) and
@@ -1119,7 +1116,7 @@ private module MethodCalls {
/** Extracted to give good join order */ /** Extracted to give good join order */
pragma[nomagic] pragma[nomagic]
private predicate fromSuper_join( private predicate fromSuper_join(
Cfg::CallNode call, string functionName, Class classUsedInSuper, AttrRead attr, Node self CallNode call, string functionName, Class classUsedInSuper, AttrRead attr, Node self
) { ) {
call.getFunction() = attrReadTracker(attr).asCfgNode() and call.getFunction() = attrReadTracker(attr).asCfgNode() and
( (
@@ -1138,7 +1135,7 @@ private module MethodCalls {
) )
} }
predicate resolveMethodCall(Cfg::CallNode call, Function target, CallType type, Node self) { predicate resolveMethodCall(CallNode call, Function target, CallType type, Node self) {
( (
directCall(call, target, _, _, _, self) directCall(call, target, _, _, _, self)
or or
@@ -1185,7 +1182,7 @@ import MethodCalls
* NOTE: We have this predicate mostly to be able to compare with old point-to * NOTE: We have this predicate mostly to be able to compare with old point-to
* call-graph resolution. So it could be removed in the future. * call-graph resolution. So it could be removed in the future.
*/ */
predicate resolveClassCall(Cfg::CallNode call, Class cls) { predicate resolveClassCall(CallNode call, Class cls) {
call.getFunction() = classTracker(cls).asCfgNode() call.getFunction() = classTracker(cls).asCfgNode()
or or
// `cls()` inside a classmethod (which also contains `type(self)()` inside a method) // `cls()` inside a classmethod (which also contains `type(self)()` inside a method)
@@ -1215,7 +1212,7 @@ Function invokedFunctionFromClassConstruction(Class cls, string funcName) {
* *
* See https://docs.python.org/3/reference/datamodel.html#object.__call__ * See https://docs.python.org/3/reference/datamodel.html#object.__call__
*/ */
predicate resolveClassInstanceCall(Cfg::CallNode call, Function target, Node self) { predicate resolveClassInstanceCall(CallNode call, Function target, Node self) {
exists(Class cls | exists(Class cls |
call.getFunction() = classInstanceTracker(cls).asCfgNode() and call.getFunction() = classInstanceTracker(cls).asCfgNode() and
target = findFunctionAccordingToMroKnownStartingClass(cls, "__call__") target = findFunctionAccordingToMroKnownStartingClass(cls, "__call__")
@@ -1234,7 +1231,7 @@ predicate resolveClassInstanceCall(Cfg::CallNode call, Function target, Node sel
* Holds if `call` is a call to the `target`, with call-type `type`. * Holds if `call` is a call to the `target`, with call-type `type`.
*/ */
cached cached
predicate resolveCall(Cfg::CallNode call, Function target, CallType type) { predicate resolveCall(CallNode call, Function target, CallType type) {
Stages::DataFlow::ref() and Stages::DataFlow::ref() and
( (
type instanceof CallTypePlainFunction and type instanceof CallTypePlainFunction and
@@ -1259,11 +1256,11 @@ predicate resolveCall(Cfg::CallNode call, Function target, CallType type) {
// ============================================================================= // =============================================================================
/** /**
* Holds if the argument of `call` at position `apos` is `arg`. This is just a helper * Holds if the argument of `call` at position `apos` is `arg`. This is just a helper
* predicate that maps ArgumentPositions to the arguments of the underlying `Cfg::CallNode`. * predicate that maps ArgumentPositions to the arguments of the underlying `CallNode`.
*/ */
overlay[local] overlay[local]
cached cached
predicate normalCallArg(Cfg::CallNode call, Node arg, ArgumentPosition apos) { predicate normalCallArg(CallNode call, Node arg, ArgumentPosition apos) {
exists(int index | exists(int index |
apos.isPositional(index) and apos.isPositional(index) and
arg.asCfgNode() = call.getArg(index) arg.asCfgNode() = call.getArg(index)
@@ -1278,7 +1275,7 @@ predicate normalCallArg(Cfg::CallNode call, Node arg, ArgumentPosition apos) {
exists(int index | exists(int index |
apos.isStarArgs(index) and apos.isStarArgs(index) and
arg.asCfgNode() = call.getStarArg() and arg.asCfgNode() = call.getStarArg() and
// since `Cfg::CallNode.getArg` doesn't include `*args`, we need to drop to the AST level // since `CallNode.getArg` doesn't include `*args`, we need to drop to the AST level
// to get the index. Notice that we only use the AST for getting the index, so we // to get the index. Notice that we only use the AST for getting the index, so we
// don't need to check for dominance in regards to splitting. // don't need to check for dominance in regards to splitting.
call.getStarArg().getNode() = call.getNode().getPositionalArg(index).(Starred).getValue() call.getStarArg().getNode() = call.getNode().getPositionalArg(index).(Starred).getValue()
@@ -1352,9 +1349,7 @@ predicate normalCallArg(Cfg::CallNode call, Node arg, ArgumentPosition apos) {
* translated into `l.clear()`, and we can still have use-use flow. * translated into `l.clear()`, and we can still have use-use flow.
*/ */
cached cached
predicate getCallArg( predicate getCallArg(CallNode call, Function target, CallType type, Node arg, ArgumentPosition apos) {
Cfg::CallNode call, Function target, CallType type, Node arg, ArgumentPosition apos
) {
Stages::DataFlow::ref() and Stages::DataFlow::ref() and
resolveCall(call, target, type) and resolveCall(call, target, type) and
( (
@@ -1447,12 +1442,10 @@ private predicate sameEnclosingCallable(Node node1, Node node2) {
// DataFlowCall // DataFlowCall
// ============================================================================= // =============================================================================
newtype TDataFlowCall = newtype TDataFlowCall =
TNormalCall(Cfg::CallNode call, Function target, CallType type) { TNormalCall(CallNode call, Function target, CallType type) { resolveCall(call, target, type) } or
call.injects(_) and resolveCall(call, target, type)
} or
/** A call to the generated function inside a comprehension */ /** A call to the generated function inside a comprehension */
TComprehensionCall(Comp c) or TComprehensionCall(Comp c) or
TPotentialLibraryCall(Cfg::CallNode call) { call.injects(_) } or TPotentialLibraryCall(CallNode call) or
/** A synthesized call inside a summarized callable */ /** A synthesized call inside a summarized callable */
TSummaryCall( TSummaryCall(
FlowSummaryImpl::Public::SummarizedCallable c, FlowSummaryImpl::Private::SummaryNode receiver FlowSummaryImpl::Public::SummarizedCallable c, FlowSummaryImpl::Private::SummaryNode receiver
@@ -1472,7 +1465,7 @@ abstract class DataFlowCall extends TDataFlowCall {
abstract ArgumentNode getArgument(ArgumentPosition apos); abstract ArgumentNode getArgument(ArgumentPosition apos);
/** Get the control flow node representing this call, if any. */ /** Get the control flow node representing this call, if any. */
abstract Cfg::ControlFlowNode getNode(); abstract ControlFlowNode getNode();
/** Gets the enclosing callable of this call. */ /** Gets the enclosing callable of this call. */
DataFlowCallable getEnclosingCallable() { result = getCallableScope(this.getScope()) } DataFlowCallable getEnclosingCallable() { result = getCallableScope(this.getScope()) }
@@ -1503,28 +1496,28 @@ abstract class ExtractedDataFlowCall extends DataFlowCall {
} }
/** /**
* A resolved call in source code with an underlying `Cfg::CallNode`. * A resolved call in source code with an underlying `CallNode`.
* *
* This is considered normal, compared with special calls such as `obj[0]` calling the * This is considered normal, compared with special calls such as `obj[0]` calling the
* `__getitem__` method on the object. However, this also includes calls that go to the * `__getitem__` method on the object. However, this also includes calls that go to the
* `__call__` special method. * `__call__` special method.
*/ */
class NormalCall extends ExtractedDataFlowCall, TNormalCall { class NormalCall extends ExtractedDataFlowCall, TNormalCall {
Cfg::CallNode call; CallNode call;
Function target; Function target;
CallType type; CallType type;
NormalCall() { this = TNormalCall(call, target, type) } NormalCall() { this = TNormalCall(call, target, type) }
override string toString() { override string toString() {
// note: if we used toString directly on the Cfg::CallNode we would get // note: if we used toString directly on the CallNode we would get
// `Cfg::ControlFlowNode for func()` // `ControlFlowNode for func()`
// but the `Cfg::ControlFlowNode` part is just clutter, so we go directly to the AST node // but the `ControlFlowNode` part is just clutter, so we go directly to the AST node
// instead. // instead.
result = call.getNode().toString() result = call.getNode().toString()
} }
override Cfg::ControlFlowNode getNode() { result = call } override ControlFlowNode getNode() { result = call }
override Scope getScope() { result = call.getScope() } override Scope getScope() { result = call.getScope() }
@@ -1552,7 +1545,7 @@ class ComprehensionCall extends ExtractedDataFlowCall, TComprehensionCall {
override string toString() { result = "comprehension call" } override string toString() { result = "comprehension call" }
override Cfg::ControlFlowNode getNode() { result.getNode() = c } override ControlFlowNode getNode() { result.getNode() = c }
override Scope getScope() { result = c.getScope() } override Scope getScope() { result = c.getScope() }
@@ -1575,14 +1568,14 @@ class ComprehensionCall extends ExtractedDataFlowCall, TComprehensionCall {
* in this class. * in this class.
*/ */
class PotentialLibraryCall extends ExtractedDataFlowCall, TPotentialLibraryCall { class PotentialLibraryCall extends ExtractedDataFlowCall, TPotentialLibraryCall {
Cfg::CallNode call; CallNode call;
PotentialLibraryCall() { this = TPotentialLibraryCall(call) } PotentialLibraryCall() { this = TPotentialLibraryCall(call) }
override string toString() { override string toString() {
// note: if we used toString directly on the Cfg::CallNode we would get // note: if we used toString directly on the CallNode we would get
// `Cfg::ControlFlowNode for func()` // `ControlFlowNode for func()`
// but the `Cfg::ControlFlowNode` part is just clutter, so we go directly to the AST node // but the `ControlFlowNode` part is just clutter, so we go directly to the AST node
// instead. // instead.
result = call.getNode().toString() result = call.getNode().toString()
} }
@@ -1599,10 +1592,10 @@ class PotentialLibraryCall extends ExtractedDataFlowCall, TPotentialLibraryCall
// potential self argument, from `foo.bar()` -- note that this could also just be a // potential self argument, from `foo.bar()` -- note that this could also just be a
// module reference, but we really don't have a good way of knowing :| // module reference, but we really don't have a good way of knowing :|
apos.isSelf() and apos.isSelf() and
result.asCfgNode() = call.getFunction().(Cfg::AttrNode).getObject() result.asCfgNode() = call.getFunction().(AttrNode).getObject()
} }
override Cfg::ControlFlowNode getNode() { result = call } override ControlFlowNode getNode() { result = call }
override Scope getScope() { result = call.getScope() } override Scope getScope() { result = call.getScope() }
} }
@@ -1634,7 +1627,7 @@ class SummaryCall extends DataFlowCall, TSummaryCall {
override ArgumentNode getArgument(ArgumentPosition apos) { none() } override ArgumentNode getArgument(ArgumentPosition apos) { none() }
override Cfg::ControlFlowNode getNode() { none() } override ControlFlowNode getNode() { none() }
override string toString() { result = "[summary] call to " + receiver + " in " + c } override string toString() { result = "[summary] call to " + receiver + " in " + c }
@@ -1776,12 +1769,12 @@ private class SummaryPostUpdateNode extends FlowSummaryNode, PostUpdateNodeImpl
* This is used for tracking flow through captured variables. * This is used for tracking flow through captured variables.
*/ */
class SynthCapturedVariablesArgumentNode extends Node, TSynthCapturedVariablesArgumentNode { class SynthCapturedVariablesArgumentNode extends Node, TSynthCapturedVariablesArgumentNode {
Cfg::ControlFlowNode callable; ControlFlowNode callable;
SynthCapturedVariablesArgumentNode() { this = TSynthCapturedVariablesArgumentNode(callable) } SynthCapturedVariablesArgumentNode() { this = TSynthCapturedVariablesArgumentNode(callable) }
/** Gets the `Cfg::CallNode` corresponding to this captured variables argument node. */ /** Gets the `CallNode` corresponding to this captured variables argument node. */
Cfg::CallNode getCallNode() { result.getFunction() = callable } CallNode getCallNode() { result.getFunction() = callable }
/** Gets the `CfgNode` that corresponds to this synthetic node. */ /** Gets the `CfgNode` that corresponds to this synthetic node. */
CfgNode getUnderlyingNode() { result.asCfgNode() = callable } CfgNode getUnderlyingNode() { result.asCfgNode() = callable }
@@ -1799,7 +1792,7 @@ class CapturedVariablesArgumentNodeAsArgumentNode extends ArgumentNode,
{ {
overlay[global] overlay[global]
override predicate argumentOf(DataFlowCall call, ArgumentPosition pos) { override predicate argumentOf(DataFlowCall call, ArgumentPosition pos) {
exists(Cfg::CallNode callNode | callNode = this.getCallNode() | exists(CallNode callNode | callNode = this.getCallNode() |
callNode = call.getNode() and callNode = call.getNode() and
exists(Function target | resolveCall(callNode, target, _) | exists(Function target | resolveCall(callNode, target, _) |
target = any(VariableCapture::CapturedVariable v).getACapturingScope() target = any(VariableCapture::CapturedVariable v).getACapturingScope()
@@ -1813,7 +1806,7 @@ class CapturedVariablesArgumentNodeAsArgumentNode extends ArgumentNode,
class SynthCapturedVariablesArgumentPostUpdateNode extends PostUpdateNodeImpl, class SynthCapturedVariablesArgumentPostUpdateNode extends PostUpdateNodeImpl,
TSynthCapturedVariablesArgumentPostUpdateNode TSynthCapturedVariablesArgumentPostUpdateNode
{ {
Cfg::ControlFlowNode callable; ControlFlowNode callable;
SynthCapturedVariablesArgumentPostUpdateNode() { SynthCapturedVariablesArgumentPostUpdateNode() {
this = TSynthCapturedVariablesArgumentPostUpdateNode(callable) this = TSynthCapturedVariablesArgumentPostUpdateNode(callable)

View File

@@ -2,9 +2,8 @@ overlay[local?]
module; module;
private import python private import python
private import semmle.python.controlflow.internal.Cfg as Cfg
private import DataFlowPublic private import DataFlowPublic
private import semmle.python.dataflow.new.internal.SsaImpl as SsaImpl private import semmle.python.essa.SsaCompute
private import semmle.python.dataflow.new.internal.ImportResolution private import semmle.python.dataflow.new.internal.ImportResolution
private import FlowSummaryImpl as FlowSummaryImpl private import FlowSummaryImpl as FlowSummaryImpl
private import semmle.python.frameworks.data.ModelsAsData private import semmle.python.frameworks.data.ModelsAsData
@@ -44,28 +43,13 @@ predicate isArgumentNode(ArgumentNode arg, DataFlowCall c, ArgumentPosition pos)
// Nodes // Nodes
//-------- //--------
overlay[local] overlay[local]
predicate isExpressionNode(Cfg::ControlFlowNode node) { predicate isExpressionNode(ControlFlowNode node) { node.getNode() instanceof Expr }
// Restrict to the `injects` representative so the dataflow layer creates
// exactly one `TCfgNode` per AST expression.
node.injects(_) and
(
node.getNode() instanceof Expr
or
// `Cfg::ForNode` wraps a `For` statement's iter position, but
// overrides `.getNode()` to return the `Py::For` statement (for
// legacy parity). The underlying AST is still an `Expr` (the iter
// expression); we want a dataflow node here so that for-loop
// content reads (`for y in l`) have a source expression node to
// read content from.
node instanceof Cfg::ForNode
)
}
// ============================================================================= // =============================================================================
// SyntheticPreUpdateNode // SyntheticPreUpdateNode
// ============================================================================= // =============================================================================
class SyntheticPreUpdateNode extends Node, TSyntheticPreUpdateNode { class SyntheticPreUpdateNode extends Node, TSyntheticPreUpdateNode {
Cfg::CallNode node; CallNode node;
SyntheticPreUpdateNode() { this = TSyntheticPreUpdateNode(node) } SyntheticPreUpdateNode() { this = TSyntheticPreUpdateNode(node) }
@@ -167,7 +151,7 @@ predicate synthStarArgsElementParameterNodeStoreStep(
* been passed in a `**kwargs` argument. * been passed in a `**kwargs` argument.
*/ */
class SynthDictSplatArgumentNode extends Node, TSynthDictSplatArgumentNode { class SynthDictSplatArgumentNode extends Node, TSynthDictSplatArgumentNode {
Cfg::CallNode node; CallNode node;
SynthDictSplatArgumentNode() { this = TSynthDictSplatArgumentNode(node) } SynthDictSplatArgumentNode() { this = TSynthDictSplatArgumentNode(node) }
@@ -181,7 +165,7 @@ class SynthDictSplatArgumentNode extends Node, TSynthDictSplatArgumentNode {
private predicate synthDictSplatArgumentNodeStoreStep( private predicate synthDictSplatArgumentNodeStoreStep(
ArgumentNode nodeFrom, DictionaryElementContent c, SynthDictSplatArgumentNode nodeTo ArgumentNode nodeFrom, DictionaryElementContent c, SynthDictSplatArgumentNode nodeTo
) { ) {
exists(string name, Cfg::CallNode call, ArgumentPosition keywordPos | exists(string name, CallNode call, ArgumentPosition keywordPos |
nodeTo = TSynthDictSplatArgumentNode(call) and nodeTo = TSynthDictSplatArgumentNode(call) and
getCallArg(call, _, _, nodeFrom, keywordPos) and getCallArg(call, _, _, nodeFrom, keywordPos) and
keywordPos.isKeyword(name) and keywordPos.isKeyword(name) and
@@ -305,7 +289,7 @@ abstract class PostUpdateNodeImpl extends Node {
* Synthetic post-update nodes for synthetic nodes need to be listed one by one. * Synthetic post-update nodes for synthetic nodes need to be listed one by one.
*/ */
class SyntheticPostUpdateNode extends PostUpdateNodeImpl, TSyntheticPostUpdateNode { class SyntheticPostUpdateNode extends PostUpdateNodeImpl, TSyntheticPostUpdateNode {
Cfg::ControlFlowNode node; ControlFlowNode node;
SyntheticPostUpdateNode() { this = TSyntheticPostUpdateNode(node) } SyntheticPostUpdateNode() { this = TSyntheticPostUpdateNode(node) }
@@ -349,42 +333,16 @@ module LocalFlow {
// `x = f(42)` // `x = f(42)`
// nodeFrom is `f(42)` // nodeFrom is `f(42)`
// nodeTo is `x` // nodeTo is `x`
// exists(AssignmentDefinition def |
// We use the CFG-level `DefinitionNode.getValue()` directly rather
// than going through SSA, because the new SSA library prunes write
// definitions that have no subsequent read in the same scope (e.g.
// a module-level `def f():` whose `f` is only read inside other
// functions). The CFG-level link is unconditional.
//
// The Name-target restriction mirrors legacy ESSA's
// `SsaDefinitions::assignment_definition`, which required
// `defn.(NameNode).defines(v)`. Subscript and attribute writes
// (`x[i] = 42`, `obj.attr = 42`) are intentionally excluded — their
// value flow is handled by the content-flow / `AttrWrite` machinery,
// not by a local-flow step *into* the Subscript/Attribute expression.
// Excluding them is essential for keeping augmented-assignment
// targets (`x[i] += 42`) classifiable as `LocalSourceNode` on the
// read side: the single canonical CFG node is both a load and a
// store, and any incoming local-flow step would disqualify it from
// being a local source.
exists(Cfg::DefinitionNode def |
nodeFrom.(CfgNode).getNode() = def.getValue() and nodeFrom.(CfgNode).getNode() = def.getValue() and
nodeTo.(CfgNode).getNode() = def and nodeTo.(CfgNode).getNode() = def.getDefiningNode()
def instanceof Cfg::NameNode and
// Parameter defaults are evaluated in the enclosing scope, while the
// parameter itself lives in the function's scope. The cross-scope
// edge is provided by `runtimeJumpStep` instead.
not exists(Py::Parameter param | def.getNode() = param.asName())
) )
or or
// With definition // With definition
// `with f(42) as x:` // `with f(42) as x:`
// nodeFrom is `f(42)` // nodeFrom is `f(42)`
// nodeTo is `x` // nodeTo is `x`
exists( exists(With with, ControlFlowNode contextManager, WithDefinition withDef, ControlFlowNode var |
With with, Cfg::ControlFlowNode contextManager, SsaImpl::WithDefinition withDef,
Cfg::ControlFlowNode var
|
var = withDef.getDefiningNode() var = withDef.getDefiningNode()
| |
nodeFrom.(CfgNode).getNode() = contextManager and nodeFrom.(CfgNode).getNode() = contextManager and
@@ -403,13 +361,13 @@ module LocalFlow {
predicate expressionFlowStep(Node nodeFrom, Node nodeTo) { predicate expressionFlowStep(Node nodeFrom, Node nodeTo) {
// If expressions // If expressions
nodeFrom.asCfgNode() = nodeTo.asCfgNode().(Cfg::IfExprNode).getAnOperand() nodeFrom.asCfgNode() = nodeTo.asCfgNode().(IfExprNode).getAnOperand()
or or
// Assignment expressions // Assignment expressions
nodeFrom.asCfgNode() = nodeTo.asCfgNode().(Cfg::AssignmentExprNode).getValue() nodeFrom.asCfgNode() = nodeTo.asCfgNode().(AssignmentExprNode).getValue()
or or
// boolean inline expressions such as `x or y` or `x and y` // boolean inline expressions such as `x or y` or `x and y`
nodeFrom.asCfgNode() = nodeTo.asCfgNode().(Cfg::BoolExprNode).getAnOperand() nodeFrom.asCfgNode() = nodeTo.asCfgNode().(BoolExprNode).getAnOperand()
or or
// Flow inside an unpacking assignment // Flow inside an unpacking assignment
iterableUnpackingFlowStep(nodeFrom, nodeTo) iterableUnpackingFlowStep(nodeFrom, nodeTo)
@@ -418,25 +376,12 @@ module LocalFlow {
matchFlowStep(nodeFrom, nodeTo) matchFlowStep(nodeFrom, nodeTo)
} }
predicate useToNextUse(Cfg::NameNode nodeFrom, Cfg::NameNode nodeTo) { predicate useToNextUse(NameNode nodeFrom, NameNode nodeTo) {
// The SSA-level adjacent-use predicate works on specific CFG variants AdjacentUses::adjacentUseUse(nodeFrom, nodeTo)
// (e.g. boolean-outcome `[true]`/`[false]` or emptiness `[empty]`/`[non-empty]`
// splits of the same AST node), but dataflow values are insensitive to
// those splits — there is at most one `CfgNode` per AST. Project both
// ends through `.getNode()` so all variants contribute their use-use
// edges to the canonical pair.
exists(Cfg::NameNode fromVariant, Cfg::NameNode toVariant |
SsaImpl::AdjacentUses::adjacentUseUse(fromVariant, toVariant) and
fromVariant.getNode() = nodeFrom.getNode() and
toVariant.getNode() = nodeTo.getNode()
)
} }
predicate defToFirstUse(SsaImpl::EssaVariable var, Cfg::NameNode nodeTo) { predicate defToFirstUse(EssaVariable var, NameNode nodeTo) {
exists(Cfg::NameNode toVariant | AdjacentUses::firstUse(var.getDefinition(), nodeTo)
SsaImpl::AdjacentUses::firstUse(var.getDefinition(), toVariant) and
toVariant.getNode() = nodeTo.getNode()
)
} }
predicate useUseFlowStep(Node nodeFrom, Node nodeTo) { predicate useUseFlowStep(Node nodeFrom, Node nodeTo) {
@@ -445,13 +390,12 @@ module LocalFlow {
// `x = f(y)` // `x = f(y)`
// nodeFrom is `y` on first line // nodeFrom is `y` on first line
// nodeTo is `y` on second line // nodeTo is `y` on second line
exists(SsaImpl::EssaDefinition def, Cfg::NameNode toVariant | exists(EssaDefinition def |
nodeFrom.(CfgNode).getNode() = def.(SsaImpl::EssaNodeDefinition).getDefiningNode() nodeFrom.(CfgNode).getNode() = def.(EssaNodeDefinition).getDefiningNode()
or or
nodeFrom.(ScopeEntryDefinitionNode).getDefinition() = def nodeFrom.(ScopeEntryDefinitionNode).getDefinition() = def
| |
SsaImpl::AdjacentUses::firstUse(def, toVariant) and AdjacentUses::firstUse(def, nodeTo.(CfgNode).getNode())
toVariant.getNode() = nodeTo.(CfgNode).getNode().getNode()
) )
or or
// Next use after use // Next use after use
@@ -613,9 +557,9 @@ predicate runtimeJumpStep(Node nodeFrom, Node nodeTo) {
// a parameter with a default value, since the parameter will be in the scope of the // a parameter with a default value, since the parameter will be in the scope of the
// function, while the default value itself will be in the scope that _defines_ the // function, while the default value itself will be in the scope that _defines_ the
// function. // function.
exists(SsaImpl::ParameterDefinition param | exists(ParameterDefinition param |
// note: we go to the _control-flow node_ of the parameter, and not the ESSA node of the parameter, since for type-tracking, the ESSA node is not a LocalSourceNode, so we would get in trouble. // note: we go to the _control-flow node_ of the parameter, and not the ESSA node of the parameter, since for type-tracking, the ESSA node is not a LocalSourceNode, so we would get in trouble.
nodeFrom.asCfgNode().getNode() = param.getParameter().(Parameter).getDefault() and nodeFrom.asCfgNode() = param.getDefault() and
nodeTo.asCfgNode() = param.getDefiningNode() nodeTo.asCfgNode() = param.getDefiningNode()
) )
or or
@@ -719,7 +663,7 @@ predicate neverSkipInPathGraph(Node n) {
// ``` // ```
// we would end up saying that the path MUST not skip the x in `y = x`, which is just // we would end up saying that the path MUST not skip the x in `y = x`, which is just
// annoying and doesn't help the path explanation become clearer. // annoying and doesn't help the path explanation become clearer.
n.asCfgNode() = any(SsaImpl::EssaNodeDefinition def).getDefiningNode() n.asCfgNode() = any(EssaNodeDefinition def).getDefiningNode()
} }
/** /**
@@ -930,7 +874,7 @@ predicate listStoreStep(CfgNode nodeFrom, ListElementContent c, CfgNode nodeTo)
// nodeFrom is `42`, cfg node // nodeFrom is `42`, cfg node
// nodeTo is the list, `[..., 42, ...]`, cfg node // nodeTo is the list, `[..., 42, ...]`, cfg node
// c denotes element of list // c denotes element of list
nodeTo.getNode().(Cfg::ListNode).getAnElement() = nodeFrom.getNode() and nodeTo.getNode().(ListNode).getAnElement() = nodeFrom.getNode() and
not nodeTo.getNode() instanceof UnpackingAssignmentSequenceTarget and not nodeTo.getNode() instanceof UnpackingAssignmentSequenceTarget and
// Suppress unused variable warning // Suppress unused variable warning
c = c c = c
@@ -943,7 +887,7 @@ predicate setStoreStep(CfgNode nodeFrom, SetElementContent c, CfgNode nodeTo) {
// nodeFrom is `42`, cfg node // nodeFrom is `42`, cfg node
// nodeTo is the set, `{..., 42, ...}`, cfg node // nodeTo is the set, `{..., 42, ...}`, cfg node
// c denotes element of list // c denotes element of list
nodeTo.getNode().(Cfg::SetNode).getAnElement() = nodeFrom.getNode() and nodeTo.getNode().(SetNode).getAnElement() = nodeFrom.getNode() and
// Suppress unused variable warning // Suppress unused variable warning
c = c c = c
} }
@@ -956,7 +900,7 @@ predicate tupleStoreStep(CfgNode nodeFrom, TupleElementContent c, CfgNode nodeTo
// nodeTo is the tuple, `(..., 42, ...)`, cfg node // nodeTo is the tuple, `(..., 42, ...)`, cfg node
// c denotes element of tuple and index of nodeFrom // c denotes element of tuple and index of nodeFrom
exists(int n | exists(int n |
nodeTo.getNode().(Cfg::TupleNode).getElement(n) = nodeFrom.getNode() and nodeTo.getNode().(TupleNode).getElement(n) = nodeFrom.getNode() and
not nodeTo.getNode() instanceof UnpackingAssignmentSequenceTarget and not nodeTo.getNode() instanceof UnpackingAssignmentSequenceTarget and
c.getIndex() = n c.getIndex() = n
) )
@@ -970,7 +914,7 @@ predicate dictStoreStep(CfgNode nodeFrom, DictionaryElementContent c, Node nodeT
// nodeTo is the dict, `{..., "key" = 42, ...}`, cfg node // nodeTo is the dict, `{..., "key" = 42, ...}`, cfg node
// c denotes element of dictionary and the key `"key"` // c denotes element of dictionary and the key `"key"`
exists(KeyValuePair item | exists(KeyValuePair item |
item = nodeTo.asCfgNode().(Cfg::DictNode).getNode().(Dict).getAnItem() and item = nodeTo.asCfgNode().(DictNode).getNode().(Dict).getAnItem() and
nodeFrom.getNode().getNode() = item.getValue() and nodeFrom.getNode().getNode() = item.getValue() and
c.getKey() = item.getKey().(StringLiteral).getS() c.getKey() = item.getKey().(StringLiteral).getS()
) )
@@ -985,9 +929,9 @@ predicate dictStoreStep(CfgNode nodeFrom, DictionaryElementContent c, Node nodeT
private predicate moreDictStoreSteps(CfgNode nodeFrom, DictionaryElementContent c, Node nodeTo) { private predicate moreDictStoreSteps(CfgNode nodeFrom, DictionaryElementContent c, Node nodeTo) {
// NOTE: It's important to add logic to the newtype definition of // NOTE: It's important to add logic to the newtype definition of
// DictionaryElementContent if you add new cases here. // DictionaryElementContent if you add new cases here.
exists(Cfg::SubscriptNode subscript | exists(SubscriptNode subscript |
nodeTo.(PostUpdateNode).getPreUpdateNode().asCfgNode() = subscript.getObject() and nodeTo.(PostUpdateNode).getPreUpdateNode().asCfgNode() = subscript.getObject() and
nodeFrom.asCfgNode() = subscript.(Cfg::DefinitionNode).getValue() and nodeFrom.asCfgNode() = subscript.(DefinitionNode).getValue() and
c.getKey() = subscript.getIndex().getNode().(StringLiteral).getText() c.getKey() = subscript.getIndex().getNode().(StringLiteral).getText()
) )
or or
@@ -1000,8 +944,8 @@ private predicate moreDictStoreSteps(CfgNode nodeFrom, DictionaryElementContent
} }
predicate dictClearStep(Node node, DictionaryElementContent c) { predicate dictClearStep(Node node, DictionaryElementContent c) {
exists(Cfg::SubscriptNode subscript | exists(SubscriptNode subscript |
subscript instanceof Cfg::DefinitionNode and subscript instanceof DefinitionNode and
node.asCfgNode() = subscript.getObject() and node.asCfgNode() = subscript.getObject() and
c.getKey() = subscript.getIndex().getNode().(StringLiteral).getText() c.getKey() = subscript.getIndex().getNode().(StringLiteral).getText()
) )
@@ -1080,7 +1024,7 @@ predicate subscriptReadStep(CfgNode nodeFrom, Content c, CfgNode nodeTo) {
// nodeFrom is `l`, cfg node // nodeFrom is `l`, cfg node
// nodeTo is `l[3]`, cfg node // nodeTo is `l[3]`, cfg node
// c is compatible with 3 // c is compatible with 3
nodeFrom.getNode() = nodeTo.getNode().(Cfg::SubscriptNode).getObject() and nodeFrom.getNode() = nodeTo.getNode().(SubscriptNode).getObject() and
( (
c instanceof ListElementContent c instanceof ListElementContent
or or
@@ -1089,10 +1033,10 @@ predicate subscriptReadStep(CfgNode nodeFrom, Content c, CfgNode nodeTo) {
c instanceof DictionaryElementAnyContent c instanceof DictionaryElementAnyContent
or or
c.(TupleElementContent).getIndex() = c.(TupleElementContent).getIndex() =
nodeTo.getNode().(Cfg::SubscriptNode).getIndex().getNode().(IntegerLiteral).getValue() nodeTo.getNode().(SubscriptNode).getIndex().getNode().(IntegerLiteral).getValue()
or or
c.(DictionaryElementContent).getKey() = c.(DictionaryElementContent).getKey() =
nodeTo.getNode().(Cfg::SubscriptNode).getIndex().getNode().(StringLiteral).getS() nodeTo.getNode().(SubscriptNode).getIndex().getNode().(StringLiteral).getS()
) )
} }
@@ -1147,7 +1091,7 @@ module Conversions {
predicate formatReadStep(Node nodeFrom, ContentSet c, Node nodeTo) { predicate formatReadStep(Node nodeFrom, ContentSet c, Node nodeTo) {
// % formatting // % formatting
exists(Cfg::BinaryExprNode fmt | fmt = nodeTo.asCfgNode() | exists(BinaryExprNode fmt | fmt = nodeTo.asCfgNode() |
fmt.getOp() instanceof Mod and fmt.getOp() instanceof Mod and
fmt.getRight() = nodeFrom.asCfgNode() fmt.getRight() = nodeFrom.asCfgNode()
) and ) and

View File

@@ -5,14 +5,11 @@ overlay[local]
module; module;
private import python private import python
private import semmle.python.controlflow.internal.Cfg as Cfg
private import semmle.python.controlflow.internal.AstNodeImpl as CfgImpl
private import codeql.controlflow.SuccessorType
private import DataFlowPrivate private import DataFlowPrivate
import semmle.python.dataflow.new.TypeTracking import semmle.python.dataflow.new.TypeTracking
import Attributes import Attributes
import LocalSources import LocalSources
private import semmle.python.dataflow.new.internal.SsaImpl as SsaImpl private import semmle.python.essa.SsaCompute
private import semmle.python.dataflow.new.internal.ImportStar private import semmle.python.dataflow.new.internal.ImportStar
private import semmle.python.frameworks.data.ModelsAsData private import semmle.python.frameworks.data.ModelsAsData
private import FlowSummaryImpl as FlowSummaryImpl private import FlowSummaryImpl as FlowSummaryImpl
@@ -30,18 +27,16 @@ private import semmle.python.frameworks.data.ModelsAsData
overlay[local] overlay[local]
newtype TNode = newtype TNode =
/** A node corresponding to a control flow node. */ /** A node corresponding to a control flow node. */
TCfgNode(Cfg::ControlFlowNode node) { TCfgNode(ControlFlowNode node) {
isExpressionNode(node) isExpressionNode(node)
or or
node.injects(_) and node.getNode() instanceof Pattern node.getNode() instanceof Pattern
} or } or
/** /**
* A node corresponding to a scope entry definition. That is, the value of a variable * A node corresponding to a scope entry definition. That is, the value of a variable
* as it enters a scope. * as it enters a scope.
*/ */
TScopeEntryDefinitionNode(SsaImpl::ScopeEntryDefinition def) { TScopeEntryDefinitionNode(ScopeEntryDefinition def) { not def.getScope() instanceof Module } or
not def.getScope() instanceof Module
} or
/** /**
* A synthetic node representing the value of an object before a state change. * A synthetic node representing the value of an object before a state change.
* *
@@ -52,15 +47,13 @@ newtype TNode =
// NOTE: since we can't rely on the call graph, but we want to have synthetic // NOTE: since we can't rely on the call graph, but we want to have synthetic
// pre-update nodes for class calls, we end up getting synthetic pre-update nodes for // pre-update nodes for class calls, we end up getting synthetic pre-update nodes for
// ALL calls :| // ALL calls :|
TSyntheticPreUpdateNode(Cfg::CallNode call) { call.injects(_) } or TSyntheticPreUpdateNode(CallNode call) or
/** /**
* A synthetic node representing the value of an object after a state change. * A synthetic node representing the value of an object after a state change.
* See QLDoc for `PostUpdateNode`. * See QLDoc for `PostUpdateNode`.
*/ */
TSyntheticPostUpdateNode(Cfg::ControlFlowNode node) { TSyntheticPostUpdateNode(ControlFlowNode node) {
node.injects(_) and exists(CallNode call |
(
exists(Cfg::CallNode call |
node = call.getArg(_) node = call.getArg(_)
or or
node = call.getArgByName(_) node = call.getArgByName(_)
@@ -69,12 +62,12 @@ newtype TNode =
node = call.getFunction() node = call.getFunction()
) )
or or
node = any(Cfg::AttrNode a).getObject() node = any(AttrNode a).getObject()
or or
node = any(Cfg::SubscriptNode s).getObject() node = any(SubscriptNode s).getObject()
or or
// self parameter when used implicitly in `super()` // self parameter when used implicitly in `super()`
exists(Class cls, Function func, SsaImpl::ParameterDefinition def | exists(Class cls, Function func, ParameterDefinition def |
func = cls.getAMethod() and func = cls.getAMethod() and
not isStaticmethod(func) and not isStaticmethod(func) and
// this matches what we do in ExtractedParameterNode // this matches what we do in ExtractedParameterNode
@@ -84,7 +77,6 @@ newtype TNode =
or or
// the iterable argument to the implicit comprehension function // the iterable argument to the implicit comprehension function
node.getNode() = any(Comp c).getIterable() node.getNode() = any(Comp c).getIterable()
)
} or } or
/** A node representing a global (module-level) variable in a specific module. */ /** A node representing a global (module-level) variable in a specific module. */
TModuleVariableNode(Module m, GlobalVariable v) { v.getScope() = m } or TModuleVariableNode(Module m, GlobalVariable v) { v.getScope() = m } or
@@ -120,9 +112,7 @@ newtype TNode =
exists(ParameterPosition ppos | ppos.isStarArgs(_) | exists(callable.getParameter(ppos))) exists(ParameterPosition ppos | ppos.isStarArgs(_) | exists(callable.getParameter(ppos)))
} or } or
/** A synthetic node to capture keyword arguments that are passed to a `**kwargs` parameter. */ /** A synthetic node to capture keyword arguments that are passed to a `**kwargs` parameter. */
TSynthDictSplatArgumentNode(Cfg::CallNode call) { TSynthDictSplatArgumentNode(CallNode call) { exists(call.getArgByName(_)) } or
call.injects(_) and exists(call.getArgByName(_))
} or
/** A synthetic node to allow flow to keyword parameters from a `**kwargs` argument. */ /** A synthetic node to allow flow to keyword parameters from a `**kwargs` argument. */
TSynthDictSplatParameterNode(DataFlowCallable callable) { TSynthDictSplatParameterNode(DataFlowCallable callable) {
exists(ParameterPosition ppos | ppos.isKeyword(_) | exists(callable.getParameter(ppos))) exists(ParameterPosition ppos | ppos.isKeyword(_) | exists(callable.getParameter(ppos)))
@@ -138,15 +128,15 @@ newtype TNode =
* A synthetic node representing the values of the variables captured * A synthetic node representing the values of the variables captured
* by the callable being called. * by the callable being called.
*/ */
TSynthCapturedVariablesArgumentNode(Cfg::ControlFlowNode callable) { TSynthCapturedVariablesArgumentNode(ControlFlowNode callable) {
callable.injects(_) and callable = any(Cfg::CallNode c).getFunction() callable = any(CallNode c).getFunction()
} or } or
/** /**
* A synthetic node representing the values of the variables captured * A synthetic node representing the values of the variables captured
* by the callable being called, after the output has been computed. * by the callable being called, after the output has been computed.
*/ */
TSynthCapturedVariablesArgumentPostUpdateNode(Cfg::ControlFlowNode callable) { TSynthCapturedVariablesArgumentPostUpdateNode(ControlFlowNode callable) {
callable.injects(_) and callable = any(Cfg::CallNode c).getFunction() callable = any(CallNode c).getFunction()
} or } or
/** A synthetic node representing the values of variables captured by a comprehension. */ /** A synthetic node representing the values of variables captured by a comprehension. */
TSynthCompCapturedVariablesArgumentNode(Comp comp) { TSynthCompCapturedVariablesArgumentNode(Comp comp) {
@@ -204,7 +194,7 @@ class Node extends TNode {
} }
/** Gets the control-flow node corresponding to this node, if any. */ /** Gets the control-flow node corresponding to this node, if any. */
Cfg::ControlFlowNode asCfgNode() { none() } ControlFlowNode asCfgNode() { none() }
/** Gets the expression corresponding to this node, if any. */ /** Gets the expression corresponding to this node, if any. */
Expr asExpr() { none() } Expr asExpr() { none() }
@@ -217,14 +207,14 @@ class Node extends TNode {
/** A data-flow node corresponding to a control-flow node. */ /** A data-flow node corresponding to a control-flow node. */
class CfgNode extends Node, TCfgNode { class CfgNode extends Node, TCfgNode {
Cfg::ControlFlowNode node; ControlFlowNode node;
CfgNode() { this = TCfgNode(node) } CfgNode() { this = TCfgNode(node) }
/** Gets the `Cfg::ControlFlowNode` represented by this data-flow node. */ /** Gets the `ControlFlowNode` represented by this data-flow node. */
Cfg::ControlFlowNode getNode() { result = node } ControlFlowNode getNode() { result = node }
override Cfg::ControlFlowNode asCfgNode() { result = node } override ControlFlowNode asCfgNode() { result = node }
/** Gets a textual representation of this element. */ /** Gets a textual representation of this element. */
override string toString() { result = node.toString() } override string toString() { result = node.toString() }
@@ -234,9 +224,9 @@ class CfgNode extends Node, TCfgNode {
override Location getLocation() { result = node.getLocation() } override Location getLocation() { result = node.getLocation() }
} }
/** A data-flow node corresponding to a `Cfg::CallNode` in the control-flow graph. */ /** A data-flow node corresponding to a `CallNode` in the control-flow graph. */
class CallCfgNode extends CfgNode, LocalSourceNode { class CallCfgNode extends CfgNode, LocalSourceNode {
override Cfg::CallNode node; override CallNode node;
/** /**
* Gets the data-flow node for the function component of the call corresponding to this data-flow * Gets the data-flow node for the function component of the call corresponding to this data-flow
@@ -317,15 +307,15 @@ ExprNode exprNode(DataFlowExpr e) { result.getNode().getNode() = e }
* as it enters a scope. * as it enters a scope.
*/ */
class ScopeEntryDefinitionNode extends Node, TScopeEntryDefinitionNode { class ScopeEntryDefinitionNode extends Node, TScopeEntryDefinitionNode {
SsaImpl::ScopeEntryDefinition def; ScopeEntryDefinition def;
ScopeEntryDefinitionNode() { this = TScopeEntryDefinitionNode(def) } ScopeEntryDefinitionNode() { this = TScopeEntryDefinitionNode(def) }
/** Gets the `SsaImpl::ScopeEntryDefinition` associated with this node. */ /** Gets the `ScopeEntryDefinition` associated with this node. */
SsaImpl::ScopeEntryDefinition getDefinition() { result = def } ScopeEntryDefinition getDefinition() { result = def }
/** Gets the source variable represented by this node. */ /** Gets the source variable represented by this node. */
SsaImpl::SsaSourceVariable getVariable() { result = def.getSourceVariable() } SsaSourceVariable getVariable() { result = def.getSourceVariable() }
override Location getLocation() { result = def.getLocation() } override Location getLocation() { result = def.getLocation() }
@@ -347,7 +337,7 @@ class ParameterNode extends Node instanceof ParameterNodeImpl {
/** A parameter node found in the source code (not in a summary). */ /** A parameter node found in the source code (not in a summary). */
class ExtractedParameterNode extends ParameterNodeImpl, CfgNode { class ExtractedParameterNode extends ParameterNodeImpl, CfgNode {
//, LocalSourceNode { //, LocalSourceNode {
SsaImpl::ParameterDefinition def; ParameterDefinition def;
ExtractedParameterNode() { node = def.getDefiningNode() } ExtractedParameterNode() { node = def.getDefiningNode() }
@@ -378,10 +368,10 @@ Node getCallArgApproximation() {
exists(Class c | result.asExpr() = c.getAMethod().getArg(0)) exists(Class c | result.asExpr() = c.getAMethod().getArg(0))
or or
// the object part of an attribute expression (which might be a bound method) // the object part of an attribute expression (which might be a bound method)
result.asCfgNode() = any(Cfg::AttrNode a).getObject() result.asCfgNode() = any(AttrNode a).getObject()
or or
// the function part of any call // the function part of any call
result.asCfgNode() = any(Cfg::CallNode c).getFunction() result.asCfgNode() = any(CallNode c).getFunction()
} }
/** Gets the extracted argument nodes that do not rely on `getCallArg`. */ /** Gets the extracted argument nodes that do not rely on `getCallArg`. */
@@ -390,7 +380,7 @@ private Node implicitArgumentNode() {
normalCallArg(_, result, _) normalCallArg(_, result, _)
or or
// and self arguments // and self arguments
result.asCfgNode() = any(Cfg::CallNode c).getFunction().(Cfg::AttrNode).getObject() result.asCfgNode() = any(CallNode c).getFunction().(AttrNode).getObject()
or or
// for comprehensions, we allow the synthetic `iterable` argument // for comprehensions, we allow the synthetic `iterable` argument
result.asExpr() = any(Comp c).getIterable() result.asExpr() = any(Comp c).getIterable()
@@ -499,20 +489,17 @@ class ModuleVariableNode extends Node, TModuleVariableNode {
not result.getScope() = mod not result.getScope() = mod
} }
/** Gets a CFG node that corresponds to an assignment of this global variable. */ /** Gets an `EssaNode` that corresponds to an assignment of this global variable. */
Node getAWrite() { Node getAWrite() {
exists(Cfg::NameNode n | any(EssaNodeDefinition def).definedBy(var, result.asCfgNode().(DefinitionNode))
n.defines(var) and
result.asCfgNode() = n
)
} }
/** Gets the possible values of the variable at the end of import time */ /** Gets the possible values of the variable at the end of import time */
CfgNode getADefiningWrite() { CfgNode getADefiningWrite() {
exists(SsaImpl::EssaVariable def | exists(SsaVariable def |
def = any(SsaImpl::EssaVariable ssa_var).getAnUltimateDefinition() and def = any(SsaVariable ssa_var).getAnUltimateDefinition() and
def.getDefinition().(SsaImpl::EssaNodeDefinition).getDefiningNode() = result.asCfgNode() and def.getDefinition() = result.asCfgNode() and
def.getSourceVariable().getVariable() = var def.getVariable() = var
) )
} }
@@ -529,7 +516,7 @@ private ModuleVariableNode import_star_read(Node n) {
overlay[global] overlay[global]
pragma[nomagic] pragma[nomagic]
private predicate resolved_import_star_module(Module m, string name, Node n) { private predicate resolved_import_star_module(Module m, string name, Node n) {
exists(Cfg::NameNode nn | nn = n.asCfgNode() | exists(NameNode nn | nn = n.asCfgNode() |
ImportStar::importStarResolvesTo(pragma[only_bind_into](nn), m) and ImportStar::importStarResolvesTo(pragma[only_bind_into](nn), m) and
nn.getId() = name nn.getId() = name
) )
@@ -587,88 +574,88 @@ class StarPatternElementNode extends Node, TStarPatternElementNode {
} }
/** /**
* A node that participates in a conditional split: a CFG node whose * Gets a node that controls whether other nodes are evaluated.
* evaluation outcome (true/false) is used to choose between two
* successor basic blocks. In the shared CFG, branching is detected
* via typed successor edges (boolean successor types) on the unique
* `injects` node for each AST expression.
* *
* Users typically obtain a `GuardNode` by casting from a more specific * In the base case, this is the last node of `conditionBlock`, and `flipped` is `false`.
* Cfg type: `g.(Cfg::CallNode)` for a call-based check, etc. * This definition accounts for (short circuting) `and`- and `or`-expressions, as the structure
* of basic blocks will reflect their semantics.
*
* However, in the program
* ```python
* if not is_safe(path):
* return
* ```
* the last node in the `ConditionBlock` is `not is_safe(path)`.
*
* We would like to consider also `is_safe(path)` a guard node, albeit with `flipped` being `true`.
* Thus we recurse through `not`-expressions.
*/ */
class GuardNode extends Cfg::ControlFlowNode { ControlFlowNode guardNode(ConditionBlock conditionBlock, boolean flipped) {
GuardNode() { // Base case: the last node truly does determine which successor is chosen
// This node has boolean successor edges (directly or via wrapping). result = conditionBlock.getLastNode() and
outcomeOfGuard(this, _, _) flipped = false
} or
// Recursive cases:
/** Holds if this guard controls block `b` upon evaluating to `branch`. */ // if a guard node is a `not`-expression,
predicate controlsBlock(Cfg::BasicBlock b, boolean branch) { // the operand is also a guard node, but with inverted polarity.
exists(CfgImpl::BasicBlock outcomeBB | exists(UnaryExprNode notNode |
outcomeOfGuard(this, outcomeBB, branch) and result = notNode.getOperand() and
outcomeBB.dominates(b) notNode.getNode().getOp() instanceof Not
|
notNode = guardNode(conditionBlock, flipped.booleanNot())
)
or
// if a guard node is compared to a boolean literal,
// the other operand is also a guard node,
// but with polarity depending on the literal (and on the comparison).
exists(CompareNode cmpNode, Cmpop op, ControlFlowNode b, boolean should_flip |
(
cmpNode.operands(result, op, b) or
cmpNode.operands(b, op, result)
) and
not result.getNode() instanceof BooleanLiteral and
(
// comparing to the boolean
(op instanceof Eq or op instanceof Is) and
// we should flip if the value compared against, here the value of `b`, is false
should_flip = b.getNode().(BooleanLiteral).booleanValue().booleanNot()
or
// comparing to the negation of the boolean
(op instanceof NotEq or op instanceof IsNot) and
// again, we should flip if the value compared against, here the value of `not b`, is false.
// That is, if the value of `b` is true.
should_flip = b.getNode().(BooleanLiteral).booleanValue()
)
|
// we flip `flipped` according to `should_flip` via the formula `flipped xor should_flip`.
flipped in [true, false] and
cmpNode = guardNode(conditionBlock, flipped.booleanXor(should_flip))
) )
}
} }
/** /**
* Holds if `outcomeBB` is the basic block entered when `guard` evaluates * A node that controls whether other nodes are evaluated.
* to `branch`.
* *
* For a direct guard `if g:`, the outcome BB starts at the after-value * The field `flipped` allows us to match `GuardNode`s underneath
* node for the matching branch. For wrapped guards like `not g` or * `not`-expressions and still choose the appropriate branch.
* `g == True`, we follow those wrappers up the AST to find the
* outermost expression that actually branches, with an appropriate
* polarity transform.
*/ */
private predicate outcomeOfGuard( class GuardNode extends ControlFlowNode {
Cfg::ControlFlowNode guard, CfgImpl::BasicBlock outcomeBB, boolean branch ConditionBlock conditionBlock;
) { boolean flipped;
// Base case: the guard has boolean successor edges.
// Only the canonical representative (injects) can act as a guard base. GuardNode() { this = guardNode(conditionBlock, flipped) }
guard.injects(_) and
exists(BooleanSuccessor t | /** Holds if this guard controls block `b` upon evaluating to `branch`. */
t.getValue() = branch and predicate controlsBlock(BasicBlock b, boolean branch) {
outcomeBB = guard.(CfgImpl::ControlFlowNode).getASuccessor(t).getBasicBlock() branch in [true, false] and
) conditionBlock.controls(b, branch.booleanXor(flipped))
or }
// Recursive: `not guard` — same outcome split as `guard`, flipped.
exists(Cfg::UnaryExprNode notNode, boolean notBranch |
notNode.injects(_) and
notNode.getOperand().getNode() = guard.getNode() and
notNode.getNode().getOp() instanceof Not and
outcomeOfGuard(notNode, outcomeBB, notBranch) and
branch = notBranch.booleanNot()
)
or
// Recursive: comparisons against a boolean literal.
exists(
Cfg::CompareNode cmpNode, Cmpop op, Cfg::ControlFlowNode otherOperand,
Cfg::ControlFlowNode guardOperand, boolean polarity, boolean cmpBranch
|
cmpNode.injects(_) and
guardOperand.getNode() = guard.getNode() and
(
cmpNode.operands(guardOperand, op, otherOperand) or
cmpNode.operands(otherOperand, op, guardOperand)
) and
not guard.getNode() instanceof BooleanLiteral and
(
(op instanceof Eq or op instanceof Is) and
polarity = otherOperand.getNode().(BooleanLiteral).booleanValue()
or
(op instanceof NotEq or op instanceof IsNot) and
polarity = otherOperand.getNode().(BooleanLiteral).booleanValue().booleanNot()
) and
outcomeOfGuard(cmpNode, outcomeBB, cmpBranch) and
branch = cmpBranch.booleanXor(polarity.booleanNot())
)
} }
/** /**
* Holds if the guard `g` validates `node` upon evaluating to `branch`. * Holds if the guard `g` validates `node` upon evaluating to `branch`.
*/ */
signature predicate guardChecksSig(GuardNode g, Cfg::ControlFlowNode node, boolean branch); signature predicate guardChecksSig(GuardNode g, ControlFlowNode node, boolean branch);
/** /**
* Provides a set of barrier nodes for a guard that validates a node. * Provides a set of barrier nodes for a guard that validates a node.
@@ -683,9 +670,7 @@ module BarrierGuard<guardChecksSig/3 guardChecks> {
result = ParameterizedBarrierGuard<Unit, extendedGuardChecks/4>::getABarrierNode(_) result = ParameterizedBarrierGuard<Unit, extendedGuardChecks/4>::getABarrierNode(_)
} }
private predicate extendedGuardChecks( private predicate extendedGuardChecks(GuardNode g, ControlFlowNode node, boolean branch, Unit u) {
GuardNode g, Cfg::ControlFlowNode node, boolean branch, Unit u
) {
guardChecks(g, node, branch) and guardChecks(g, node, branch) and
u = u u = u
} }
@@ -695,7 +680,7 @@ bindingset[this]
private signature class ParamSig; private signature class ParamSig;
private module WithParam<ParamSig P> { private module WithParam<ParamSig P> {
signature predicate guardChecksSig(GuardNode g, Cfg::ControlFlowNode node, boolean branch, P param); signature predicate guardChecksSig(GuardNode g, ControlFlowNode node, boolean branch, P param);
} }
/** /**
@@ -708,16 +693,10 @@ module ParameterizedBarrierGuard<ParamSig P, WithParam<P>::guardChecksSig/4 guar
/** Gets a node that is safely guarded by the given guard check with parameter `param`. */ /** Gets a node that is safely guarded by the given guard check with parameter `param`. */
overlay[global] overlay[global]
ExprNode getABarrierNode(P param) { ExprNode getABarrierNode(P param) {
exists(GuardNode g, SsaImpl::EssaDefinition def, Cfg::ControlFlowNode node, boolean branch | exists(GuardNode g, EssaDefinition def, ControlFlowNode node, boolean branch |
SsaImpl::AdjacentUses::useOfDef(def, node) and AdjacentUses::useOfDef(def, node) and
guardChecks(g, node, branch, param) and guardChecks(g, node, branch, param) and
SsaImpl::AdjacentUses::useOfDef(def, result.asCfgNode()) and AdjacentUses::useOfDef(def, result.asCfgNode()) and
// The protected use must be a different SSA position than the test
// position itself: `controlsBlock` is reflexive on dominance, and
// the test expression is an SSA-use position on the def-use chain.
// Without this guard, the test position would be returned as a
// barrier and block flow before it can reach genuine branch uses.
node != result.asCfgNode() and
g.controlsBlock(result.asCfgNode().getBasicBlock(), branch) g.controlsBlock(result.asCfgNode().getBasicBlock(), branch)
) )
} }
@@ -733,7 +712,7 @@ module ExternalBarrierGuard {
private import semmle.python.ApiGraphs private import semmle.python.ApiGraphs
overlay[global] overlay[global]
private predicate guardCheck(GuardNode g, Cfg::ControlFlowNode node, boolean branch, string kind) { private predicate guardCheck(GuardNode g, ControlFlowNode node, boolean branch, string kind) {
exists(API::CallNode call, API::Node parameter | exists(API::CallNode call, API::Node parameter |
parameter = call.getAParameter() and parameter = call.getAParameter() and
parameter = ModelOutput::getABarrierGuardNode(kind, branch) parameter = ModelOutput::getABarrierGuardNode(kind, branch)
@@ -769,10 +748,10 @@ newtype TContent =
TSetElementContent() or TSetElementContent() or
/** An element of a tuple at a specific index. */ /** An element of a tuple at a specific index. */
TTupleElementContent(int index) { TTupleElementContent(int index) {
exists(any(Cfg::TupleNode tn).getElement(index)) exists(any(TupleNode tn).getElement(index))
or or
// Arguments can overflow and end up in the starred parameter tuple. // Arguments can overflow and end up in the starred parameter tuple.
exists(any(Cfg::CallNode cn).getArg(index)) exists(any(CallNode cn).getArg(index))
or or
// since flow summaries might use tuples, we ensure that we at least have valid // since flow summaries might use tuples, we ensure that we at least have valid
// TTupleElementContent for the 0..7 (7 was picked to match `small_tuple` in // TTupleElementContent for the 0..7 (7 was picked to match `small_tuple` in
@@ -789,14 +768,10 @@ newtype TContent =
or or
// d["key"] = ... // d["key"] = ...
key = key =
any(Cfg::SubscriptNode sub | any(SubscriptNode sub | sub.isStore() | sub.getIndex().getNode().(StringLiteral).getText())
sub.isStore()
|
sub.getIndex().getNode().(StringLiteral).getText()
)
or or
// d.setdefault("key", ...) // d.setdefault("key", ...)
exists(Cfg::CallNode call | call.getFunction().(Cfg::AttrNode).getName() = "setdefault" | exists(CallNode call | call.getFunction().(AttrNode).getName() = "setdefault" |
key = call.getArg(0).getNode().(StringLiteral).getText() key = call.getArg(0).getNode().(StringLiteral).getText()
) )
} or } or

View File

@@ -5,18 +5,17 @@
*/ */
private import python private import python
private import semmle.python.controlflow.internal.Cfg as Cfg
private import semmle.python.dataflow.new.internal.SsaImpl as SsaImpl
private import semmle.python.dataflow.new.DataFlow private import semmle.python.dataflow.new.DataFlow
private import semmle.python.dataflow.new.internal.ImportStar private import semmle.python.dataflow.new.internal.ImportStar
private import semmle.python.dataflow.new.TypeTracking private import semmle.python.dataflow.new.TypeTracking
private import semmle.python.dataflow.new.internal.DataFlowPrivate private import semmle.python.dataflow.new.internal.DataFlowPrivate
/** /**
* Holds if the name of `var` refers to a submodule of a package and `init` is * Holds if `init` is a package's `__init__.py` and `var` is a global variable in
* the `__init__` module of that package. Locally inlined replacement for the * `init` whose name matches a submodule of the package.
* legacy `SsaSource::init_module_submodule_defn` so that this module has no *
* direct dependency on `semmle.python.essa.SsaDefinitions`. * Inlined from `SsaSource::init_module_submodule_defn` to avoid pulling
* `semmle.python.essa.SsaDefinitions` into the new dataflow stack.
*/ */
private predicate initModuleSubmoduleDefn(GlobalVariable var, Module init) { private predicate initModuleSubmoduleDefn(GlobalVariable var, Module init) {
init.isPackageInit() and init.isPackageInit() and
@@ -82,19 +81,13 @@ module ImportResolution {
* Holds if there is an ESSA step from `defFrom` to `defTo`, which should be allowed * Holds if there is an ESSA step from `defFrom` to `defTo`, which should be allowed
* for import resolution. * for import resolution.
*/ */
private predicate allowedEssaImportStep( private predicate allowedEssaImportStep(EssaDefinition defFrom, EssaDefinition defTo) {
SsaImpl::EssaDefinition defFrom, SsaImpl::EssaDefinition defTo
) {
// to handle definitions guarded by if-then-else // to handle definitions guarded by if-then-else
defFrom = defTo.(SsaImpl::PhiFunction).getAnInput() defFrom = defTo.(PhiFunction).getAnInput()
or or
// to handle uncertain writes such as `from X import *`, which create an // refined variable
// uncertain SSA definition for every name in the importing scope. The // example: https://github.com/nvbn/thefuck/blob/ceeaeab94b5df5a4fe9d94d61e4f6b0bbea96378/thefuck/utils.py#L25-L45
// immediately preceding definition is still potentially the value of the defFrom = defTo.(EssaNodeRefinement).getInput().getDefinition()
// module export.
SsaImpl::Ssa::uncertainWriteDefinitionInput(defTo, defFrom)
// Note: legacy ESSA refinement-step (e.g. for `foo.bar = X`) is
// not modelled in the new SSA beyond the cases handled above.
} }
/** /**
@@ -111,32 +104,30 @@ module ImportResolution {
// Definitions made inside `m` itself // Definitions made inside `m` itself
// //
// for code such as `foo = ...; foo.bar = ...` there will be TWO // for code such as `foo = ...; foo.bar = ...` there will be TWO
// SsaImpl::EssaDefinition/SsaImpl::EssaVariable. One for `foo = ...` (SsaImpl::AssignmentDefinition) and one // EssaDefinition/EssaVariable. One for `foo = ...` (AssignmentDefinition) and one
// for `foo.bar = ...`. The one for `foo.bar = ...` (EssaNodeRefinement). The // for `foo.bar = ...`. The one for `foo.bar = ...` (EssaNodeRefinement). The
// EssaNodeRefinement is the one that will reach the end of the module (normal // EssaNodeRefinement is the one that will reach the end of the module (normal
// exit). // exit).
// //
// However, we cannot just use the EssaNodeRefinement as the `val`, because the // However, we cannot just use the EssaNodeRefinement as the `val`, because the
// normal data-flow depends on use-use flow, and use-use flow targets CFG nodes not // normal data-flow depends on use-use flow, and use-use flow targets CFG nodes not
// EssaNodes. So we need to go back from the SsaImpl::EssaDefinition/SsaImpl::EssaVariable that // EssaNodes. So we need to go back from the EssaDefinition/EssaVariable that
// reaches the end of the module, to the first definition of the variable, and then // reaches the end of the module, to the first definition of the variable, and then
// track forwards using use-use flow to find a suitable CFG node that has flow into // track forwards using use-use flow to find a suitable CFG node that has flow into
// it from use-use flow. // it from use-use flow.
exists(SsaImpl::EssaVariable lastUseVar, SsaImpl::EssaVariable firstDef | exists(EssaVariable lastUseVar, EssaVariable firstDef |
lastUseVar.getName() = name and lastUseVar.getName() = name and
// we ignore special variable $ introduced by our analysis (not used for anything) // we ignore special variable $ introduced by our analysis (not used for anything)
// we ignore special variable * introduced by `from <pkg> import *` -- TODO: understand why we even have this? // we ignore special variable * introduced by `from <pkg> import *` -- TODO: understand why we even have this?
not name in ["$", "*"] and not name in ["$", "*"] and
exists(Cfg::ControlFlowNode exit | lastUseVar.getAUse() = m.getANormalExit() and
exit.isNormalExit() and exit.getScope() = m and lastUseVar.getAUse() = exit
) and
allowedEssaImportStep*(firstDef, lastUseVar) and allowedEssaImportStep*(firstDef, lastUseVar) and
not allowedEssaImportStep(_, firstDef) not allowedEssaImportStep(_, firstDef)
| |
not LocalFlow::defToFirstUse(firstDef, _) and not LocalFlow::defToFirstUse(firstDef, _) and
val.asCfgNode() = firstDef.getDefinition().(SsaImpl::EssaNodeDefinition).getDefiningNode() val.asCfgNode() = firstDef.getDefinition().(EssaNodeDefinition).getDefiningNode()
or or
exists(Cfg::ControlFlowNode mid, Cfg::ControlFlowNode end | exists(ControlFlowNode mid, ControlFlowNode end |
LocalFlow::defToFirstUse(firstDef, mid) and LocalFlow::defToFirstUse(firstDef, mid) and
LocalFlow::useToNextUse*(mid, end) and LocalFlow::useToNextUse*(mid, end) and
not LocalFlow::useToNextUse(end, _) and not LocalFlow::useToNextUse(end, _) and
@@ -164,9 +155,9 @@ module ImportResolution {
* handles simple cases where we can statically tell that this is the case. * handles simple cases where we can statically tell that this is the case.
*/ */
private predicate all_mentions_name(Module m, string name) { private predicate all_mentions_name(Module m, string name) {
exists(Cfg::DefinitionNode def, Cfg::SequenceNode n | exists(DefinitionNode def, SequenceNode n |
def.getValue() = n and def.getValue() = n and
def.(Cfg::NameNode).getId() = "__all__" and def.(NameNode).getId() = "__all__" and
def.getScope() = m and def.getScope() = m and
any(StringLiteral s | s.getText() = name) = n.getAnElement().getNode() any(StringLiteral s | s.getText() = name) = n.getAnElement().getNode()
) )
@@ -179,20 +170,18 @@ module ImportResolution {
*/ */
private predicate no_or_complicated_all(Module m) { private predicate no_or_complicated_all(Module m) {
// No mention of `__all__` in the module // No mention of `__all__` in the module
not exists(Cfg::DefinitionNode def | not exists(DefinitionNode def | def.getScope() = m and def.(NameNode).getId() = "__all__")
def.getScope() = m and def.(Cfg::NameNode).getId() = "__all__"
)
or or
// `__all__` is set to a non-sequence value // `__all__` is set to a non-sequence value
exists(Cfg::DefinitionNode def | exists(DefinitionNode def |
def.(Cfg::NameNode).getId() = "__all__" and def.(NameNode).getId() = "__all__" and
def.getScope() = m and def.getScope() = m and
not def.getValue() instanceof Cfg::SequenceNode not def.getValue() instanceof SequenceNode
) )
or or
// `__all__` is used in some way that doesn't involve storing a value in it. This usually means // `__all__` is used in some way that doesn't involve storing a value in it. This usually means
// it is being mutated through `append` or `extend`, which we don't handle. // it is being mutated through `append` or `extend`, which we don't handle.
exists(Cfg::NameNode n | n.getId() = "__all__" and n.getScope() = m and n.isLoad()) exists(NameNode n | n.getId() = "__all__" and n.getScope() = m and n.isLoad())
} }
private predicate potential_module_export(Module m, string name) { private predicate potential_module_export(Module m, string name) {
@@ -200,7 +189,7 @@ module ImportResolution {
or or
no_or_complicated_all(m) and no_or_complicated_all(m) and
( (
exists(Cfg::NameNode n | n.getId() = name and n.getScope() = m and name.charAt(0) != "_") exists(NameNode n | n.getId() = name and n.getScope() = m and name.charAt(0) != "_")
or or
exists(Alias a | a.getAsname().(Name).getId() = name and a.getValue().getScope() = m) exists(Alias a | a.getAsname().(Name).getId() = name and a.getValue().getScope() = m)
) )
@@ -230,12 +219,12 @@ module ImportResolution {
/** Gets a module that may have been added to `sys.modules`. */ /** Gets a module that may have been added to `sys.modules`. */
private Module sys_modules_module_with_name(string name) { private Module sys_modules_module_with_name(string name) {
exists(Cfg::ControlFlowNode n, DataFlow::Node mod | exists(ControlFlowNode n, DataFlow::Node mod |
exists(Cfg::SubscriptNode sub | exists(SubscriptNode sub |
sub.getObject() = sys_modules_reference().asCfgNode() and sub.getObject() = sys_modules_reference().asCfgNode() and
sub.getIndex() = n and sub.getIndex() = n and
n.getNode().(StringLiteral).getText() = name and n.getNode().(StringLiteral).getText() = name and
sub.(Cfg::DefinitionNode).getValue() = mod.asCfgNode() and sub.(DefinitionNode).getValue() = mod.asCfgNode() and
mod = getModuleReference(result) mod = getModuleReference(result)
) )
) )
@@ -347,11 +336,11 @@ module ImportResolution {
// name as a submodule, we always consider that this attribute _could_ be a // name as a submodule, we always consider that this attribute _could_ be a
// reference to the submodule, even if we don't know that the submodule has been // reference to the submodule, even if we don't know that the submodule has been
// imported yet. // imported yet.
exists(string submodule, Module package, SsaImpl::EssaVariable var | exists(string submodule, Module package, EssaVariable var |
submodule = var.getName() and submodule = var.getName() and
initModuleSubmoduleDefn(var.getSourceVariable().getVariable(), package) and initModuleSubmoduleDefn(var.getSourceVariable(), package) and
m = getModuleFromName(package.getPackageName() + "." + submodule) and m = getModuleFromName(package.getPackageName() + "." + submodule) and
result.asCfgNode() = var.getDefinition().(SsaImpl::EssaNodeDefinition).getDefiningNode() result.asCfgNode() = var.getDefinition().(EssaNodeDefinition).getDefiningNode()
) )
} }

View File

@@ -3,7 +3,6 @@ overlay[local]
module; module;
private import python private import python
private import semmle.python.controlflow.internal.Cfg as Cfg
private import semmle.python.dataflow.new.internal.Builtins private import semmle.python.dataflow.new.internal.Builtins
private import semmle.python.dataflow.new.internal.ImportResolution private import semmle.python.dataflow.new.internal.ImportResolution
private import semmle.python.dataflow.new.DataFlow private import semmle.python.dataflow.new.DataFlow
@@ -16,7 +15,7 @@ module ImportStar {
*/ */
overlay[local] overlay[local]
cached cached
predicate namePossiblyDefinedInImportStar(Cfg::NameNode n, string name, Scope s) { predicate namePossiblyDefinedInImportStar(NameNode n, string name, Scope s) {
n.isLoad() and n.isLoad() and
name = n.getId() and name = n.getId() and
s = n.getScope().getEnclosingScope*() and s = n.getScope().getEnclosingScope*() and
@@ -53,7 +52,7 @@ module ImportStar {
/** Holds if a global variable called `name` is assigned a value in the module `m`. */ /** Holds if a global variable called `name` is assigned a value in the module `m`. */
cached cached
predicate globalNameDefinedInModule(string name, Module m) { predicate globalNameDefinedInModule(string name, Module m) {
exists(Cfg::NameNode n | exists(NameNode n |
not exists(LocalVariable v | n.defines(v)) and not exists(LocalVariable v | n.defines(v)) and
n.isStore() and n.isStore() and
name = n.getId() and name = n.getId() and
@@ -67,7 +66,7 @@ module ImportStar {
*/ */
overlay[global] overlay[global]
cached cached
predicate importStarResolvesTo(Cfg::NameNode n, Module m) { predicate importStarResolvesTo(NameNode n, Module m) {
m = getStarImported+(n.getEnclosingModule()) and m = getStarImported+(n.getEnclosingModule()) and
globalNameDefinedInModule(n.getId(), m) and globalNameDefinedInModule(n.getId(), m) and
not isDefinedLocally(n.getNode()) not isDefinedLocally(n.getNode())
@@ -100,7 +99,7 @@ module ImportStar {
*/ */
overlay[local] overlay[local]
cached cached
Cfg::ControlFlowNode potentialImportStarBase(Scope s) { ControlFlowNode potentialImportStarBase(Scope s) {
result = any(Cfg::ImportStarNode n | n.getScope() = s).getModule() result = any(ImportStarNode n | n.getScope() = s).getModule()
} }
} }

View File

@@ -170,8 +170,6 @@ overlay[local]
module; module;
private import python private import python
private import semmle.python.controlflow.internal.Cfg as Cfg
private import semmle.python.dataflow.new.internal.SsaImpl as SsaImpl
private import DataFlowPublic private import DataFlowPublic
/** /**
@@ -180,7 +178,7 @@ private import DataFlowPublic
* This class abstracts away the differing representations of comprehensions and * This class abstracts away the differing representations of comprehensions and
* for statements. * for statements.
*/ */
class ForTarget extends Cfg::ControlFlowNode { class ForTarget extends ControlFlowNode {
Expr source; Expr source;
ForTarget() { ForTarget() {
@@ -200,7 +198,7 @@ class ForTarget extends Cfg::ControlFlowNode {
} }
/** The LHS of an assignment, it also records the assigned value. */ /** The LHS of an assignment, it also records the assigned value. */
class AssignmentTarget extends Cfg::ControlFlowNode { class AssignmentTarget extends ControlFlowNode {
Expr value; Expr value;
AssignmentTarget() { AssignmentTarget() {
@@ -211,7 +209,7 @@ class AssignmentTarget extends Cfg::ControlFlowNode {
} }
/** A direct (or top-level) target of an unpacking assignment. */ /** A direct (or top-level) target of an unpacking assignment. */
class UnpackingAssignmentDirectTarget extends Cfg::ControlFlowNode instanceof Cfg::SequenceNode { class UnpackingAssignmentDirectTarget extends ControlFlowNode instanceof SequenceNode {
Expr value; Expr value;
UnpackingAssignmentDirectTarget() { UnpackingAssignmentDirectTarget() {
@@ -224,7 +222,7 @@ class UnpackingAssignmentDirectTarget extends Cfg::ControlFlowNode instanceof Cf
} }
/** A (possibly recursive) target of an unpacking assignment. */ /** A (possibly recursive) target of an unpacking assignment. */
class UnpackingAssignmentTarget extends Cfg::ControlFlowNode { class UnpackingAssignmentTarget extends ControlFlowNode {
UnpackingAssignmentTarget() { UnpackingAssignmentTarget() {
this instanceof UnpackingAssignmentDirectTarget this instanceof UnpackingAssignmentDirectTarget
or or
@@ -233,11 +231,10 @@ class UnpackingAssignmentTarget extends Cfg::ControlFlowNode {
} }
/** A (possibly recursive) target of an unpacking assignment which is also a sequence. */ /** A (possibly recursive) target of an unpacking assignment which is also a sequence. */
class UnpackingAssignmentSequenceTarget extends UnpackingAssignmentTarget instanceof Cfg::SequenceNode class UnpackingAssignmentSequenceTarget extends UnpackingAssignmentTarget instanceof SequenceNode {
{ ControlFlowNode getElement(int i) { result = super.getElement(i) }
Cfg::ControlFlowNode getElement(int i) { result = super.getElement(i) }
Cfg::ControlFlowNode getAnElement() { result = this.getElement(_) } ControlFlowNode getAnElement() { result = this.getElement(_) }
} }
/** /**
@@ -258,7 +255,7 @@ predicate iterableUnpackingAssignmentFlowStep(Node nodeFrom, Node nodeTo) {
predicate iterableUnpackingForReadStep(CfgNode nodeFrom, Content c, Node nodeTo) { predicate iterableUnpackingForReadStep(CfgNode nodeFrom, Content c, Node nodeTo) {
exists(ForTarget target | exists(ForTarget target |
nodeFrom.getNode().getNode() = target.getSource() and nodeFrom.getNode().getNode() = target.getSource() and
target instanceof Cfg::SequenceNode and target instanceof SequenceNode and
nodeTo = TIterableSequenceNode(target) nodeTo = TIterableSequenceNode(target)
) and ) and
( (
@@ -326,11 +323,11 @@ predicate iterableUnpackingConvertingStoreStep(Node nodeFrom, Content c, Node no
*/ */
predicate iterableUnpackingElementReadStep(Node nodeFrom, Content c, Node nodeTo) { predicate iterableUnpackingElementReadStep(Node nodeFrom, Content c, Node nodeTo) {
exists( exists(
UnpackingAssignmentSequenceTarget target, int index, Cfg::ControlFlowNode element, int starIndex UnpackingAssignmentSequenceTarget target, int index, ControlFlowNode element, int starIndex
| |
target.getElement(starIndex) instanceof Cfg::StarredNode target.getElement(starIndex) instanceof StarredNode
or or
not exists(target.getAnElement().(Cfg::StarredNode)) and not exists(target.getAnElement().(StarredNode)) and
starIndex = -1 starIndex = -1
| |
nodeFrom.(CfgNode).getNode() = target and nodeFrom.(CfgNode).getNode() = target and
@@ -345,18 +342,18 @@ predicate iterableUnpackingElementReadStep(Node nodeFrom, Content c, Node nodeTo
else c.(TupleElementContent).getIndex() >= index - 1 else c.(TupleElementContent).getIndex() >= index - 1
) and ) and
( (
if element instanceof Cfg::SequenceNode if element instanceof SequenceNode
then then
// Step 5b // Step 5b
nodeTo = TIterableSequenceNode(element) nodeTo = TIterableSequenceNode(element)
else else
if element instanceof Cfg::StarredNode if element instanceof StarredNode
then then
// Step 5c // Step 5c
nodeTo = TIterableElementNode(element) nodeTo = TIterableElementNode(element)
else else
// Step 5a // Step 5a
exists(SsaImpl::MultiAssignmentDefinition mad | element = mad.getDefiningNode() | exists(MultiAssignmentDefinition mad | element = mad.getDefiningNode() |
nodeTo.(CfgNode).getNode() = element nodeTo.(CfgNode).getNode() = element
) )
) )
@@ -369,7 +366,7 @@ predicate iterableUnpackingElementReadStep(Node nodeFrom, Content c, Node nodeTo
* content type `ListElementContent`. * content type `ListElementContent`.
*/ */
predicate iterableUnpackingStarredElementStoreStep(Node nodeFrom, Content c, Node nodeTo) { predicate iterableUnpackingStarredElementStoreStep(Node nodeFrom, Content c, Node nodeTo) {
exists(Cfg::ControlFlowNode starred, SsaImpl::MultiAssignmentDefinition mad | exists(ControlFlowNode starred, MultiAssignmentDefinition mad |
starred.getNode() instanceof Starred and starred.getNode() instanceof Starred and
starred = mad.getDefiningNode() starred = mad.getDefiningNode()
| |

View File

@@ -9,7 +9,6 @@ overlay[local]
module; module;
private import python private import python
private import semmle.python.controlflow.internal.Cfg as Cfg
import DataFlowPublic import DataFlowPublic
private import DataFlowPrivate private import DataFlowPrivate
private import semmle.python.internal.CachedStages private import semmle.python.internal.CachedStages
@@ -315,7 +314,7 @@ private module Cached {
*/ */
cached cached
predicate subscript(LocalSourceNode node, CfgNode subscript, CfgNode index) { predicate subscript(LocalSourceNode node, CfgNode subscript, CfgNode index) {
exists(CfgNode seq, Cfg::SubscriptNode subscriptNode | subscriptNode = subscript.getNode() | exists(CfgNode seq, SubscriptNode subscriptNode | subscriptNode = subscript.getNode() |
node.flowsTo(seq) and node.flowsTo(seq) and
seq.getNode() = subscriptNode.getObject() and seq.getNode() = subscriptNode.getObject() and
index.getNode() = subscriptNode.getIndex() index.getNode() = subscriptNode.getIndex()

View File

@@ -91,7 +91,9 @@ predicate matchAsFlowStep(Node nodeFrom, Node nodeTo) {
or or
// the interior pattern flows to the alias // the interior pattern flows to the alias
nodeFrom.(CfgNode).getNode().getNode() = subject.getPattern() and nodeFrom.(CfgNode).getNode().getNode() = subject.getPattern() and
nodeTo.(CfgNode).getNode().getNode() = alias exists(PatternAliasDefinition pad | pad.getDefiningNode().getNode() = alias |
nodeTo.(CfgNode).getNode() = pad.getDefiningNode()
)
) )
} }
@@ -122,9 +124,11 @@ predicate matchLiteralFlowStep(Node nodeFrom, Node nodeTo) {
* syntax (toplevel): `case var:` * syntax (toplevel): `case var:`
*/ */
predicate matchCaptureFlowStep(Node nodeFrom, Node nodeTo) { predicate matchCaptureFlowStep(Node nodeFrom, Node nodeTo) {
exists(MatchCapturePattern capture | exists(MatchCapturePattern capture, Name var | capture.getVariable() = var |
nodeFrom.(CfgNode).getNode().getNode() = capture and nodeFrom.(CfgNode).getNode().getNode() = capture and
nodeTo.(CfgNode).getNode().getNode() = capture.getVariable() exists(PatternCaptureDefinition pcd | pcd.getDefiningNode().getNode() = var |
nodeTo.(CfgNode).getNode() = pcd.getDefiningNode()
)
) )
} }

View File

@@ -1,548 +0,0 @@
/**
* Provides the Python SSA implementation built on the new (shared) CFG.
*
* Mirrors the Java SSA adapter at
* `java/ql/lib/semmle/code/java/dataflow/internal/SsaImpl.qll`:
* an `InputSig` is defined in terms of positional `(BasicBlock, int)`
* variable references, and the shared
* `codeql.ssa.Ssa::Make<Location, Cfg, Input>` module is then
* instantiated.
*
* `SourceVariable` is the AST-level `Py::Variable`. Variable references
* are looked up via the CFG facade's `NameNode.defines`/`uses`/`deletes`
* predicates, which themselves are one-line bridges to AST-level
* `Name.defines`/`uses`/`deletes`.
*
* Implicit-entry definitions are inserted for:
* - non-local / global / builtin variables that are read in the scope
* but never assigned (no enclosing CFG node defines them),
* - captured variables (variables defined in an enclosing scope that
* are read inside the scope), and
* - parameters, but only if the corresponding parameter name is *not*
* itself a CFG node. With the C#-style parameter wiring already
* installed in `AstNodeImpl.qll`, parameter names *are* CFG nodes,
* so the regular `variableWrite` path handles them — no `i = -1`
* entry is needed for ordinary parameters.
*/
overlay[local?]
module;
private import python as Py
private import semmle.python.controlflow.internal.AstNodeImpl as CfgImpl
private import semmle.python.controlflow.internal.Cfg as Cfg
private import codeql.ssa.Ssa as SsaImplCommon
private import codeql.controlflow.BasicBlock as BB
/**
* Adapts the Python `Cfg` facade to the shared SSA library's `CfgSig`.
* All members are inherited from `Cfg::ControlFlowNode` and
* `Cfg::BasicBlock`.
*/
private module CfgForSsa implements BB::CfgSig<Py::Location> {
class ControlFlowNode = CfgImpl::ControlFlowNode;
class BasicBlock = CfgImpl::BasicBlock;
class EntryBasicBlock = CfgImpl::Cfg::EntryBasicBlock;
predicate dominatingEdge = CfgImpl::Cfg::dominatingEdge/2;
}
/**
* A source variable for SSA, wrapping a Python AST `Variable`.
*
* We only track variables that are read at least once in their scope —
* tracking write-only variables would be unnecessary work — *except*
* for module-scope globals, where the "read" can be external (e.g.
* `import mymodule; mymodule.x`). Such globals are tracked
* unconditionally so that import-resolution can find their defining
* write.
*/
private newtype TSsaSourceVariable =
TPyVar(Py::Variable v) {
// Has a use somewhere — read-relevant for SSA.
exists(Cfg::NameNode n | n.uses(v))
or
// Or has a deletion (treated as a write that destroys the value).
exists(Cfg::NameNode n | n.deletes(v))
or
// Or is a module-scope global written in this module — must be
// tracked even if never read locally, because importers may read
// it as an attribute on the module object.
v.getScope() instanceof Py::Module and
exists(Cfg::NameNode n | n.defines(v))
or
// Or is a parameter — parameters must always have a
// `ParameterDefinition` for dataflow argument-routing to work,
// even if the parameter is never read in its scope. Mirrors
// legacy ESSA's `ParameterDefinition` (which fired for every
// parameter binding regardless of liveness).
exists(Py::Parameter p | p.asName() = v.getAStore())
}
/**
* A source variable for SSA, wrapping a Python AST `Variable`.
*/
class SsaSourceVariable extends TSsaSourceVariable {
/** Gets the underlying Python AST variable. */
Py::Variable getVariable() { this = TPyVar(result) }
/** Gets the (textual) name of this variable. */
string getName() { result = this.getVariable().getId() }
/** Gets a textual representation of this source variable. */
string toString() { result = this.getVariable().toString() }
/** Gets the location of this source variable. */
Py::Location getLocation() { result = this.getVariable().getScope().getLocation() }
/** Gets the scope in which this variable lives. */
Py::Scope getScope() { result = this.getVariable().getScope() }
/**
* Gets a use of this variable as it appears in the source — a `NameNode`
* that loads or deletes the variable. Mirrors legacy
* `SsaSourceVariable.getASourceUse()`.
*/
Cfg::ControlFlowNode getASourceUse() {
exists(Cfg::NameNode n | result = n |
n.uses(this.getVariable()) or n.deletes(this.getVariable())
)
}
/**
* Gets an implicit use of this variable. The new SSA does not have
* implicit-use refinements, but we keep this for API parity — every
* normal-exit of the variable's scope counts as a sink, ensuring
* variables stay live to scope exit for taint-tracking.
*/
Cfg::ControlFlowNode getAnImplicitUse() {
result.isNormalExit() and result.getScope() = this.getScope()
}
/**
* Gets a use of this variable — either an explicit source use or an
* implicit use at scope exit. Mirrors legacy `SsaSourceVariable.getAUse()`.
*/
Cfg::ControlFlowNode getAUse() {
result = this.getASourceUse() or result = this.getAnImplicitUse()
}
}
/**
* Holds if `v` is a non-local read in scope `s`, in the sense that `s`
* uses `v` but does not write it within `s`. This includes globals,
* builtins, and variables captured from an enclosing function scope.
*
* The `Py::Variable` `v` lives in some defining scope (the module for
* globals, an outer function for closures, etc.); the reading scope
* `s` is the scope where the use of `v` occurs.
*/
private predicate nonLocalReadIn(Py::Variable v, Py::Scope s) {
exists(Cfg::NameNode n |
n.uses(v) and
n.getScope() = s and
not exists(Cfg::NameNode def | def.defines(v) and def.getScope() = s)
) and
// Match legacy ESSA: only create entry defs for variables that have
// at least one defining store somewhere — otherwise the entry def
// represents "nothing reaches here", which is the default anyway and
// introduces no useful flow. (Legacy's `ModuleVariable` required a
// store; this is the closure-aware generalisation.)
exists(Cfg::NameNode store | store.defines(v))
}
/**
* Holds if `bb` is the entry basic block of a scope where `v` should
* have an implicit entry definition. This covers:
* - non-local / global / builtin variables read in `s`, and
* - captured variables (defined in an enclosing scope but read in `s`).
*
* Each reading scope gets its own entry def, so a closure variable can
* have multiple entry defs across all functions/methods that read it.
*
* Parameters are *not* included: their bound `Name` is itself a CFG
* node (per the C#-style parameter wiring), so `variableWrite` fires at
* the parameter's natural CFG index.
*/
private predicate hasEntryDefIn(SsaSourceVariable v, CfgImpl::BasicBlock bb) {
exists(Py::Scope s |
nonLocalReadIn(v.getVariable(), s) and
bb = entryBlock(s)
)
}
/**
* Gets the entry basic block of scope `s`, where implicit entry
* definitions are placed (at synthetic index `-1`).
*/
private CfgImpl::BasicBlock entryBlock(Py::Scope s) {
exists(CfgImpl::ControlFlowNode entry |
entry instanceof CfgImpl::ControlFlow::EntryNode and
entry.getEnclosingCallable().asScope() = s and
result = entry.getBasicBlock()
)
}
/**
* The SSA `InputSig` for Python. References are positional
* `(BasicBlock, int)` pairs into the new CFG.
*/
private module SsaImplInput implements SsaImplCommon::InputSig<Py::Location, CfgImpl::BasicBlock> {
class SourceVariable = SsaSourceVariable;
predicate variableWrite(CfgImpl::BasicBlock bb, int i, SourceVariable v, boolean certain) {
// Explicit binding at a CFG node — includes assignments,
// parameter Names (wired in via the C# pattern), exception-handler
// `as`-bindings, import aliases, and match-pattern captures.
exists(Cfg::NameNode n |
bb.getNode(i) = n and
n.defines(v.getVariable()) and
certain = true
)
or
// `del x` — removes the binding. Modelled as a certain write that
// makes any subsequent read invalid.
exists(Cfg::NameNode n |
bb.getNode(i) = n and
n.deletes(v.getVariable()) and
certain = true
)
or
// Implicit entry definition for non-local / captured / global /
// builtin variables read in some scope. Each reading scope's entry
// block gets one such write, allowing closures: e.g. when `x` is a
// parameter of an outer function and read inside a nested
// function, both scopes get entry defs for `x`.
hasEntryDefIn(v, bb) and
i = -1 and
certain = true
or
// `from X import *` — possibly rebinds every name in the importing
// scope. Modelled as an uncertain write at the import-star's CFG
// position for every variable that lives in (or is referenced
// from) the same scope as the import-star. Mirrors legacy ESSA's
// `ImportStarRefinement` (see `essa/SsaDefinitions.qll`'s
// `import_star_refinement` predicate). The write is uncertain so
// that prior definitions of the variable remain available — the
// shared-SSA `SsaUncertainWrite` merges the new value with the
// immediately preceding definition.
exists(Cfg::ImportStarNode imp |
imp.injects(_) and
bb.getNode(i) = imp and
certain = false and
(
v.getVariable().getScope() = imp.getScope()
or
// Variable is defined in some other scope but referenced in
// the same scope as the import-star (matches legacy clause 2:
// `other.uses(v) and def.getScope() = other.getScope()`).
exists(Cfg::NameNode other |
other.uses(v.getVariable()) and
imp.getScope() = other.getScope()
)
)
)
}
predicate variableRead(CfgImpl::BasicBlock bb, int i, SourceVariable v, boolean certain) {
// Explicit source use — a `Name` load or a `del x` of the variable.
exists(Cfg::NameNode n |
bb.getNode(i) = n and
n.uses(v.getVariable()) and
certain = true
)
or
// Synthetic use at the normal exit of the variable's defining scope.
// This keeps every variable live to scope exit so that callers (e.g.
// `module_export` in ImportResolution.qll, or taint-tracking pass-through
// through unread locals) can ask "which definition reaches end of
// scope?". Mirrors legacy ESSA's `SsaSourceVariable.getAUse()` which
// included `getScope().getANormalExit()`.
exists(Cfg::ControlFlowNode exit |
exit.isNormalExit() and
exit.getScope() = v.getVariable().getScope() and
bb.getNode(i) = exit and
certain = true
)
}
}
/**
* The shared SSA instantiation for Python.
*
* Members:
* - `Definition` — the union of explicit, uncertain, and phi definitions
* - `WriteDefinition`, `UncertainWriteDefinition`, `PhiNode`
* - the standard SSA predicates (`getAUse`, `getAnUltimateDefinition`, ...).
*/
module Ssa = SsaImplCommon::Make<Py::Location, CfgForSsa, SsaImplInput>;
final class Definition = Ssa::Definition;
final class WriteDefinition = Ssa::WriteDefinition;
final class UncertainWriteDefinition = Ssa::UncertainWriteDefinition;
final class PhiNode = Ssa::PhiNode;
// ===========================================================================
// ESSA-shaped adapter layer
//
// The dataflow library (`python/ql/lib/semmle/python/dataflow/new/`) and
// related modules (`ApiGraphs.qll`, etc.) consume the legacy ESSA API
// (`EssaVariable`, `EssaDefinition`, `AssignmentDefinition`,
// `ScopeEntryDefinition`, `ParameterDefinition`, `WithDefinition`,
// `PhiFunction`, plus the `AdjacentUses` module). To migrate them off
// the legacy CFG, we expose the same API surface on top of the
// shared SSA built above.
//
// This adapter is intentionally narrow: it covers only the predicates
// that new dataflow consumes. The richer legacy ESSA — refinement
// nodes, attribute refinements, edge refinements — stays available
// via `semmle.python.essa.Essa` for points-to / legacy code.
// ===========================================================================
/**
* Gets the CFG node at which a write definition's binding takes place.
*
* For ordinary writes (assignment, deletion, parameter) this is the
* canonical CFG node of the bound Name. For implicit entry definitions
* (synthesised at position `-1` of a scope's entry BB) this is the
* scope's entry node.
*/
private Cfg::ControlFlowNode writeDefNode(Ssa::WriteDefinition def) {
exists(CfgImpl::BasicBlock bb, int i | def.definesAt(_, bb, i) |
i >= 0 and result = bb.getNode(i)
or
i = -1 and result = bb.getNode(0)
)
}
/**
* A write definition whose binding has a corresponding CFG node — i.e.
* everything that's not a phi node. Mirrors legacy ESSA's
* `EssaNodeDefinition`.
*/
class EssaNodeDefinition extends Ssa::WriteDefinition {
/** Gets the CFG node where this definition's binding takes place. */
Cfg::ControlFlowNode getDefiningNode() { result = writeDefNode(this) }
/** Gets the variable defined here (legacy name). */
SsaSourceVariable getVariable() { result = this.getSourceVariable() }
/** Gets the enclosing scope. */
Py::Scope getScope() {
exists(Cfg::ControlFlowNode n | n = this.getDefiningNode() | result = n.getScope())
}
/**
* Holds if this definition defines source variable `v` at CFG node
* `defNode`. Flatter form of `getSourceVariable()` +
* `getDefiningNode()`, matching legacy ESSA's `definedBy`.
*/
predicate definedBy(SsaSourceVariable v, Cfg::ControlFlowNode defNode) {
v = this.getSourceVariable() and defNode = this.getDefiningNode()
}
}
/**
* An assignment definition: any binding where the value being assigned
* is statically known via `Cfg::DefinitionNode.getValue()`. Includes
* plain assignments, walrus, annotated assignments, augmented
* assignments, import aliases (`import x` / `from m import x [as y]`),
* `with ... as x`, and for-target bindings (where `getValue()` returns
* the iter expression's CFG node). Excludes parameter bindings —
* those are modelled by `ParameterDefinition`.
*/
class AssignmentDefinition extends EssaNodeDefinition {
AssignmentDefinition() {
exists(Cfg::NameNode n | n = this.getDefiningNode() |
exists(n.(Cfg::DefinitionNode).getValue()) and
not n.(Cfg::ControlFlowNode).isParameter()
)
}
/** Gets the CFG node for the value being assigned, if statically known. */
Cfg::ControlFlowNode getValue() {
result = this.getDefiningNode().(Cfg::DefinitionNode).getValue()
}
}
/**
* A parameter definition — the binding of a parameter name in a
* function's scope.
*/
class ParameterDefinition extends EssaNodeDefinition {
ParameterDefinition() { this.getDefiningNode().isParameter() }
/** Gets the AST `Parameter` (a `Py::Name` in param context). */
Py::Name getParameter() { result = this.getDefiningNode().getNode() }
}
/**
* A definition introduced by a `with ... as x:` clause.
*/
class WithDefinition extends EssaNodeDefinition {
WithDefinition() {
exists(Cfg::NameNode n, Py::With w |
n = this.getDefiningNode() and
w.getOptionalVars() = n.getNode()
)
}
}
/**
* An assignment where the LHS is a tuple/list and the RHS is unpacked:
* `a, b = (1, 2)` or `a, *rest = xs`. The SSA def lives at the inner
* `Name` CFG node, but for IterableUnpacking integration we expose
* the enclosing `StarredNode` as the `getDefiningNode()` for `*rest`
* patterns — mirroring legacy ESSA's `multi_assignment_definition`,
* which placed the def at the StarredNode CFG node.
*/
class MultiAssignmentDefinition extends EssaNodeDefinition {
MultiAssignmentDefinition() {
exists(Cfg::NameNode n | n = super.getDefiningNode() |
exists(Py::Assign a, Py::Expr lhs |
a.getATarget() = lhs and
(lhs instanceof Py::Tuple or lhs instanceof Py::List) and
lhs.getASubExpression+() = n.getNode()
)
or
// For-loop with tuple/list target: `for a, b in xs:` —
// tuple-unpacking semantics applies to the for-target.
exists(Py::For f, Py::Expr lhs |
f.getTarget() = lhs and
(lhs instanceof Py::Tuple or lhs instanceof Py::List) and
lhs.getASubExpression+() = n.getNode()
)
)
}
override Cfg::ControlFlowNode getDefiningNode() {
// Default: the underlying `Name` CFG node (where the SSA def lives).
not exists(Cfg::StarredNode s |
s.getNode().(Py::Starred).getValue() = super.getDefiningNode().getNode()
) and
result = super.getDefiningNode()
or
// Exception: for `*rest`, expose the enclosing `Starred` CFG node
// so that `IterableUnpacking::iterableUnpackingStarredElementStoreStep`
// can attach the rest-list to it.
exists(Cfg::StarredNode s |
s.getNode().(Py::Starred).getValue() = super.getDefiningNode().getNode()
|
result = s
)
}
}
/**
* An implicit entry definition for a non-local / captured / global /
* builtin variable read in a scope but not defined there.
*
* Inherits from `EssaNodeDefinition` and exposes the scope's entry node
* as its defining node (matching legacy ESSA semantics).
*/
class ScopeEntryDefinition extends EssaNodeDefinition {
ScopeEntryDefinition() {
exists(CfgImpl::BasicBlock bb |
this.definesAt(_, bb, -1) and
bb instanceof CfgImpl::Cfg::EntryBasicBlock
)
}
/** Gets the enclosing scope (the scope whose entry block this def is in). */
override Py::Scope getScope() {
exists(CfgImpl::BasicBlock bb |
this.definesAt(_, bb, -1) and
result = bb.getNode(0).(Cfg::ControlFlowNode).getScope()
)
}
}
/** A phi node (alias matching legacy naming). */
class PhiFunction extends PhiNode {
/**
* Gets an input to this phi function (a definition that flows into
* the phi from one of its predecessor blocks). Mirrors legacy
* ESSA's `PhiFunction.getAnInput()`.
*/
Ssa::Definition getAnInput() { Ssa::phiHasInputFromBlock(this, result, _) }
}
/** Base class for all ESSA definitions (legacy-shaped). */
class EssaDefinition = Ssa::Definition;
/**
* An adapter representing a single SSA-defined "variable" — wrapping
* one `Ssa::Definition`. Mirrors legacy `EssaVariable` API.
*/
class EssaVariable extends Ssa::Definition {
/** Gets the underlying SSA definition (legacy name). */
Ssa::Definition getDefinition() { result = this }
/**
* Gets a CFG node where this definition is used. Includes regular
* `Name` reads as well as the synthetic scope-exit "use" registered
* via `SsaImplInput::variableRead` — mirrors legacy ESSA's
* `EssaVariable.getAUse()` which inherited the synthetic exit-use
* from `SsaSourceVariable`.
*/
Cfg::ControlFlowNode getAUse() {
exists(CfgImpl::BasicBlock bb, int i |
Ssa::ssaDefReachesRead(this.getSourceVariable(), this, bb, i) and
bb.getNode(i) = result
)
}
/** Gets the (textual) name of the underlying variable. */
string getName() { result = this.getSourceVariable().getVariable().getId() }
/** Gets the scope in which this variable lives. */
Py::Scope getScope() { result = this.getSourceVariable().getVariable().getScope() }
/** Gets an ultimate non-phi ancestor of this definition. */
EssaVariable getAnUltimateDefinition() {
if this instanceof PhiNode
then
exists(Ssa::Definition input |
Ssa::phiHasInputFromBlock(this, input, _) and
result = input.(EssaVariable).getAnUltimateDefinition()
)
else result = this
}
}
/**
* Adjacent use-use and def-use relations exposed by the shared SSA
* library. Provides the same interface as legacy
* `semmle.python.essa.SsaCompute::AdjacentUses`.
*/
module AdjacentUses {
/** Holds if `nodeFrom` and `nodeTo` are adjacent uses of the same SSA variable. */
predicate adjacentUseUse(Cfg::NameNode nodeFrom, Cfg::NameNode nodeTo) {
exists(SsaSourceVariable v, CfgImpl::BasicBlock bb1, int i1, CfgImpl::BasicBlock bb2, int i2 |
Ssa::adjacentUseUse(bb1, i1, bb2, i2, v, _) and
nodeFrom = bb1.getNode(i1) and
nodeTo = bb2.getNode(i2)
)
}
/** Holds if `use` is a first use of definition `def`. */
predicate firstUse(Ssa::Definition def, Cfg::NameNode use) {
exists(CfgImpl::BasicBlock bb, int i |
Ssa::firstUse(def, bb, i, _) and
use = bb.getNode(i)
)
}
/**
* Holds if `use` is any reachable use of definition `def`. Combines
* `firstUse` with transitive use-use adjacency.
*/
predicate useOfDef(Ssa::Definition def, Cfg::NameNode use) {
firstUse(def, use)
or
exists(Cfg::NameNode mid | useOfDef(def, mid) and adjacentUseUse(mid, use))
}
}

View File

@@ -1,6 +1,4 @@
private import python private import python
private import semmle.python.controlflow.internal.Cfg as Cfg
private import semmle.python.dataflow.new.internal.SsaImpl as SsaImpl
private import semmle.python.dataflow.new.DataFlow private import semmle.python.dataflow.new.DataFlow
private import semmle.python.dataflow.new.internal.DataFlowPrivate as DataFlowPrivate private import semmle.python.dataflow.new.internal.DataFlowPrivate as DataFlowPrivate
private import FlowSummaryImpl as FlowSummaryImpl private import FlowSummaryImpl as FlowSummaryImpl
@@ -99,7 +97,7 @@ import Cached
* and isn't a big problem in practice. * and isn't a big problem in practice.
*/ */
predicate concatStep(DataFlow::CfgNode nodeFrom, DataFlow::CfgNode nodeTo) { predicate concatStep(DataFlow::CfgNode nodeFrom, DataFlow::CfgNode nodeTo) {
exists(Cfg::BinaryExprNode add | add = nodeTo.getNode() | exists(BinaryExprNode add | add = nodeTo.getNode() |
add.getOp() instanceof Add and add.getAnOperand() = nodeFrom.getNode() add.getOp() instanceof Add and add.getAnOperand() = nodeFrom.getNode()
) )
} }
@@ -108,7 +106,7 @@ predicate concatStep(DataFlow::CfgNode nodeFrom, DataFlow::CfgNode nodeTo) {
* Holds if taint can flow from `nodeFrom` to `nodeTo` with a step related to subscripting. * Holds if taint can flow from `nodeFrom` to `nodeTo` with a step related to subscripting.
*/ */
predicate subscriptStep(DataFlow::CfgNode nodeFrom, DataFlow::CfgNode nodeTo) { predicate subscriptStep(DataFlow::CfgNode nodeFrom, DataFlow::CfgNode nodeTo) {
nodeTo.getNode().(Cfg::SubscriptNode).getObject() = nodeFrom.getNode() nodeTo.getNode().(SubscriptNode).getObject() = nodeFrom.getNode()
} }
/** /**
@@ -124,15 +122,15 @@ predicate stringManipulation(DataFlow::CfgNode nodeFrom, DataFlow::CfgNode nodeT
( (
call = API::builtin(["str", "bytes", "unicode"]).getACall() call = API::builtin(["str", "bytes", "unicode"]).getACall()
or or
call.getFunction().asCfgNode().(Cfg::NameNode).getId() in ["str", "bytes", "unicode"] call.getFunction().asCfgNode().(NameNode).getId() in ["str", "bytes", "unicode"]
) and ) and
nodeFrom in [call.getArg(0), call.getArgByName("object")] nodeFrom in [call.getArg(0), call.getArgByName("object")]
) )
or or
// String methods. Note that this doesn't recognize `meth = "foo".upper; meth()` // String methods. Note that this doesn't recognize `meth = "foo".upper; meth()`
exists(Cfg::CallNode call, string method_name, Cfg::ControlFlowNode object | exists(CallNode call, string method_name, ControlFlowNode object |
call = nodeTo.getNode() and call = nodeTo.getNode() and
object = call.getFunction().(Cfg::AttrNode).getObject(method_name) object = call.getFunction().(AttrNode).getObject(method_name)
| |
nodeFrom.getNode() = object and nodeFrom.getNode() = object and
method_name in [ method_name in [
@@ -158,7 +156,7 @@ predicate stringManipulation(DataFlow::CfgNode nodeFrom, DataFlow::CfgNode nodeT
) )
or or
// % formatting // % formatting
exists(Cfg::BinaryExprNode fmt | fmt = nodeTo.getNode() | exists(BinaryExprNode fmt | fmt = nodeTo.getNode() |
fmt.getOp() instanceof Mod and fmt.getOp() instanceof Mod and
( (
fmt.getLeft() = nodeFrom.getNode() fmt.getLeft() = nodeFrom.getNode()
@@ -168,7 +166,7 @@ predicate stringManipulation(DataFlow::CfgNode nodeFrom, DataFlow::CfgNode nodeT
) )
or or
// string multiplication -- `"foo" * 10` // string multiplication -- `"foo" * 10`
exists(Cfg::BinaryExprNode mult | mult = nodeTo.getNode() | exists(BinaryExprNode mult | mult = nodeTo.getNode() |
mult.getOp() instanceof Mult and mult.getOp() instanceof Mult and
mult.getLeft() = nodeFrom.getNode() mult.getLeft() = nodeFrom.getNode()
) )
@@ -215,8 +213,8 @@ predicate awaitStep(DataFlow::Node nodeFrom, DataFlow::Node nodeTo) {
* the variable `f` is tainted if the result of `open("foo")` is tainted. * the variable `f` is tainted if the result of `open("foo")` is tainted.
*/ */
predicate asyncWithStep(DataFlow::Node nodeFrom, DataFlow::Node nodeTo) { predicate asyncWithStep(DataFlow::Node nodeFrom, DataFlow::Node nodeTo) {
exists(With with, Cfg::ControlFlowNode contextManager, Cfg::ControlFlowNode var | exists(With with, ControlFlowNode contextManager, ControlFlowNode var |
var = any(SsaImpl::WithDefinition wd).getDefiningNode() var = any(WithDefinition wd).getDefiningNode()
| |
nodeFrom.(DataFlow::CfgNode).getNode() = contextManager and nodeFrom.(DataFlow::CfgNode).getNode() = contextManager and
nodeTo.(DataFlow::CfgNode).getNode() = var and nodeTo.(DataFlow::CfgNode).getNode() = var and

View File

@@ -2,8 +2,6 @@ import codeql.util.Unit
import codeql.typetracking.TypeTracking as Shared import codeql.typetracking.TypeTracking as Shared
import codeql.typetracking.internal.TypeTrackingImpl as SharedImpl import codeql.typetracking.internal.TypeTrackingImpl as SharedImpl
private import python private import python
private import semmle.python.controlflow.internal.Cfg as Cfg
private import semmle.python.dataflow.new.internal.SsaImpl as SsaImpl
private import semmle.python.internal.CachedStages private import semmle.python.internal.CachedStages
private import semmle.python.dataflow.new.internal.DataFlowPublic as DataFlowPublic private import semmle.python.dataflow.new.internal.DataFlowPublic as DataFlowPublic
private import semmle.python.dataflow.new.internal.DataFlowPrivate as DataFlowPrivate private import semmle.python.dataflow.new.internal.DataFlowPrivate as DataFlowPrivate
@@ -164,7 +162,7 @@ module TypeTrackingInput implements Shared::TypeTrackingInput<Location> {
// ignore the flow steps from the synthetic sequence node to the real sequence node, // ignore the flow steps from the synthetic sequence node to the real sequence node,
// since we only support one level of content in type-trackers, and the nested // since we only support one level of content in type-trackers, and the nested
// structure requires two levels at least to be useful. // structure requires two levels at least to be useful.
not exists(Cfg::SequenceNode outer | not exists(SequenceNode outer |
outer.getAnElement() = nodeTo.asCfgNode() and outer.getAnElement() = nodeTo.asCfgNode() and
IterableUnpacking::iterableUnpackingTupleFlowStep(nodeFrom, nodeTo) IterableUnpacking::iterableUnpackingTupleFlowStep(nodeFrom, nodeTo)
) )
@@ -269,7 +267,7 @@ module TypeTrackingInput implements Shared::TypeTrackingInput<Location> {
// Since we only support one level of content in type-trackers we don't actually // Since we only support one level of content in type-trackers we don't actually
// support `(aa, ab), (ba, bb) = ...`. Therefore we exclude the read-step from `(aa, // support `(aa, ab), (ba, bb) = ...`. Therefore we exclude the read-step from `(aa,
// ab)` to `aa` (since it is not needed). // ab)` to `aa` (since it is not needed).
not exists(Cfg::SequenceNode outer | not exists(SequenceNode outer |
outer.getAnElement() = nodeFrom.asCfgNode() and outer.getAnElement() = nodeFrom.asCfgNode() and
IterableUnpacking::iterableUnpackingTupleFlowStep(_, nodeFrom) IterableUnpacking::iterableUnpackingTupleFlowStep(_, nodeFrom)
) and ) and
@@ -279,7 +277,7 @@ module TypeTrackingInput implements Shared::TypeTrackingInput<Location> {
IterableUnpacking::iterableUnpackingForReadStep(_, _, seq) and IterableUnpacking::iterableUnpackingForReadStep(_, _, seq) and
IterableUnpacking::iterableUnpackingConvertingReadStep(seq, _, elem) and IterableUnpacking::iterableUnpackingConvertingReadStep(seq, _, elem) and
IterableUnpacking::iterableUnpackingConvertingStoreStep(elem, _, nodeFrom) and IterableUnpacking::iterableUnpackingConvertingStoreStep(elem, _, nodeFrom) and
nodeFrom.asCfgNode() instanceof Cfg::SequenceNode nodeFrom.asCfgNode() instanceof SequenceNode
) )
or or
TypeTrackerSummaryFlow::basicLoadStep(nodeFrom, nodeTo, DataFlowPublic::singleton(content)) TypeTrackerSummaryFlow::basicLoadStep(nodeFrom, nodeTo, DataFlowPublic::singleton(content))
@@ -317,15 +315,13 @@ module TypeTrackingInput implements Shared::TypeTrackingInput<Location> {
// //
// nodeFrom is `expr` // nodeFrom is `expr`
// nodeTo is entry node for `f` // nodeTo is entry node for `f`
exists( exists(ScopeEntryDefinition e, SsaSourceVariable var, DefinitionNode def |
SsaImpl::ScopeEntryDefinition e, SsaImpl::SsaSourceVariable var, Cfg::DefinitionNode def
|
e.getSourceVariable() = var and e.getSourceVariable() = var and
def.getNode() = var.getVariable().getAStore() var.hasDefiningNode(def)
| |
nodeTo.(DataFlowPublic::ScopeEntryDefinitionNode).getDefinition() = e and nodeTo.(DataFlowPublic::ScopeEntryDefinitionNode).getDefinition() = e and
nodeFrom.asCfgNode() = def and nodeFrom.asCfgNode() = def and
var.getVariable().getScope().getScope*() = nodeFrom.getScope() var.getScope().getScope*() = nodeFrom.getScope()
) )
} }

View File

@@ -3,9 +3,6 @@ overlay[local]
module; module;
private import python private import python
private import semmle.python.controlflow.internal.Cfg as Cfg
private import semmle.python.controlflow.internal.AstNodeImpl as CfgImpl
private import semmle.python.dataflow.new.internal.SsaImpl as SsaImpl
private import DataFlowPublic private import DataFlowPublic
private import semmle.python.dataflow.new.internal.DataFlowPrivate private import semmle.python.dataflow.new.internal.DataFlowPrivate
private import codeql.dataflow.VariableCapture as Shared private import codeql.dataflow.VariableCapture as Shared
@@ -17,10 +14,10 @@ private import codeql.dataflow.VariableCapture as Shared
// The first is the main implementation, the second is a performance motivated restriction. // The first is the main implementation, the second is a performance motivated restriction.
// The restriction is to clear any `CapturedVariableContent` before writing a new one // The restriction is to clear any `CapturedVariableContent` before writing a new one
// to avoid long access paths (see the link for a nice explanation). // to avoid long access paths (see the link for a nice explanation).
private module CaptureInput implements Shared::InputSig<Location, CfgImpl::BasicBlock> { private module CaptureInput implements Shared::InputSig<Location, Cfg::BasicBlock> {
private import python as PY private import python as PY
additional class ExprCfgNode extends Cfg::ControlFlowNode { additional class ExprCfgNode extends ControlFlowNode {
ExprCfgNode() { isExpressionNode(this) } ExprCfgNode() { isExpressionNode(this) }
} }
@@ -28,9 +25,7 @@ private module CaptureInput implements Shared::InputSig<Location, CfgImpl::Basic
predicate isConstructor() { none() } predicate isConstructor() { none() }
} }
Callable basicBlockGetEnclosingCallable(CfgImpl::BasicBlock bb) { Callable basicBlockGetEnclosingCallable(Cfg::BasicBlock bb) { result = bb.getScope() }
result = bb.getEnclosingCallable().asScope()
}
class CapturedVariable extends LocalVariable { class CapturedVariable extends LocalVariable {
Function f; Function f;
@@ -56,23 +51,21 @@ private module CaptureInput implements Shared::InputSig<Location, CfgImpl::Basic
class CapturedParameter extends CapturedVariable { class CapturedParameter extends CapturedVariable {
CapturedParameter() { this.isParameter() } CapturedParameter() { this.isParameter() }
Cfg::ControlFlowNode getCfgNode() { result.getNode().(Parameter) = this.getAnAccess() } ControlFlowNode getCfgNode() { result.getNode().(Parameter) = this.getAnAccess() }
} }
class Expr extends ExprCfgNode { class Expr extends ExprCfgNode {
predicate hasCfgNode(CfgImpl::BasicBlock bb, int i) { this = bb.getNode(i) } predicate hasCfgNode(Cfg::BasicBlock bb, int i) { this = bb.getNode(i) }
} }
class VariableWrite extends Cfg::ControlFlowNode { class VariableWrite extends ControlFlowNode {
CapturedVariable v; CapturedVariable v;
VariableWrite() { VariableWrite() { exists(DefinitionNode d | d.getNode() = v.getAStore() | this = d.getValue()) }
exists(Cfg::DefinitionNode d | d.getNode() = v.getAStore() | this = d.getValue())
}
CapturedVariable getVariable() { result = v } CapturedVariable getVariable() { result = v }
predicate hasCfgNode(CfgImpl::BasicBlock bb, int i) { this = bb.getNode(i) } predicate hasCfgNode(Cfg::BasicBlock bb, int i) { this = bb.getNode(i) }
} }
class VariableRead extends Expr { class VariableRead extends Expr {
@@ -87,14 +80,9 @@ private module CaptureInput implements Shared::InputSig<Location, CfgImpl::Basic
// TODO: Other languages have an extra case here looking like // TODO: Other languages have an extra case here looking like
// simpleAstFlowStep(nodeFrom, nodeTo) // simpleAstFlowStep(nodeFrom, nodeTo)
// we should investigate the potential benefit of adding that. // we should investigate the potential benefit of adding that.
exists(SsaImpl::EssaVariable def | exists(SsaVariable def |
def.getAUse() = nodeTo and def.getAUse() = nodeTo and
def.getAnUltimateDefinition() def.getAnUltimateDefinition().getDefinition().(DefinitionNode).getValue() = nodeFrom
.getDefinition()
.(SsaImpl::EssaNodeDefinition)
.getDefiningNode()
.(Cfg::DefinitionNode)
.getValue() = nodeFrom
) )
} }
@@ -119,7 +107,7 @@ class CapturedVariable = CaptureInput::CapturedVariable;
class ClosureExpr = CaptureInput::ClosureExpr; class ClosureExpr = CaptureInput::ClosureExpr;
module Flow = Shared::Flow<Location, Cfg::CfgSigImpl, CaptureInput>; module Flow = Shared::Flow<Location, Cfg, CaptureInput>;
private Flow::ClosureNode asClosureNode(Node n) { private Flow::ClosureNode asClosureNode(Node n) {
result = n.(SynthCaptureNode).getSynthesizedCaptureNode() result = n.(SynthCaptureNode).getSynthesizedCaptureNode()

View File

@@ -4,7 +4,6 @@
*/ */
private import python private import python
private import semmle.python.controlflow.internal.Cfg as Cfg
private import semmle.python.Concepts private import semmle.python.Concepts
private import semmle.python.ApiGraphs private import semmle.python.ApiGraphs
private import semmle.python.dataflow.new.RemoteFlowSources private import semmle.python.dataflow.new.RemoteFlowSources
@@ -158,9 +157,9 @@ module Bottle {
DataFlow::Node value; DataFlow::Node value;
HeaderWriteSubscript() { HeaderWriteSubscript() {
exists(Cfg::SubscriptNode subscript | exists(SubscriptNode subscript |
this.asCfgNode() = subscript and this.asCfgNode() = subscript and
value.asCfgNode() = subscript.(Cfg::DefinitionNode).getValue() and value.asCfgNode() = subscript.(DefinitionNode).getValue() and
name.asCfgNode() = subscript.getIndex() and name.asCfgNode() = subscript.getIndex() and
subscript.getObject() = headers().asSource().asCfgNode() subscript.getObject() = headers().asSource().asCfgNode()
) )

View File

@@ -4,7 +4,6 @@
*/ */
private import python private import python
private import semmle.python.controlflow.internal.Cfg as Cfg
private import semmle.python.dataflow.new.DataFlow private import semmle.python.dataflow.new.DataFlow
private import semmle.python.dataflow.new.RemoteFlowSources private import semmle.python.dataflow.new.RemoteFlowSources
private import semmle.python.dataflow.new.TaintTracking private import semmle.python.dataflow.new.TaintTracking
@@ -1306,7 +1305,7 @@ module PrivateDjango {
dict.(DataFlow::MethodCallNode).calls(files, "dict") dict.(DataFlow::MethodCallNode).calls(files, "dict")
) )
| |
this.asCfgNode().(Cfg::SubscriptNode).getObject() = dict.asCfgNode() this.asCfgNode().(SubscriptNode).getObject() = dict.asCfgNode()
or or
this.(DataFlow::MethodCallNode).calls(dict, "get") this.(DataFlow::MethodCallNode).calls(dict, "get")
) )
@@ -1315,7 +1314,7 @@ module PrivateDjango {
exists(DataFlow::AttrRead files, DataFlow::MethodCallNode getlistCall | exists(DataFlow::AttrRead files, DataFlow::MethodCallNode getlistCall |
files.accesses(instance(), "FILES") and files.accesses(instance(), "FILES") and
getlistCall.calls(files, "getlist") and getlistCall.calls(files, "getlist") and
this.asCfgNode().(Cfg::SubscriptNode).getObject() = getlistCall.asCfgNode() this.asCfgNode().(SubscriptNode).getObject() = getlistCall.asCfgNode()
) )
} }
} }
@@ -2217,7 +2216,7 @@ module PrivateDjango {
DataFlow::Node value; DataFlow::Node value;
DjangoResponseCookieSubscriptWrite() { DjangoResponseCookieSubscriptWrite() {
exists(Cfg::SubscriptNode subscript, DataFlow::AttrRead cookieLookup | exists(SubscriptNode subscript, DataFlow::AttrRead cookieLookup |
// To give `this` a value, we need to choose between either LHS or RHS, // To give `this` a value, we need to choose between either LHS or RHS,
// and just go with the LHS // and just go with the LHS
this.asCfgNode() = subscript this.asCfgNode() = subscript
@@ -2229,7 +2228,7 @@ module PrivateDjango {
| |
cookieLookup.flowsTo(subscriptObj) cookieLookup.flowsTo(subscriptObj)
) and ) and
value.asCfgNode() = subscript.(Cfg::DefinitionNode).getValue() and value.asCfgNode() = subscript.(DefinitionNode).getValue() and
index.asCfgNode() = subscript.getIndex() index.asCfgNode() = subscript.getIndex()
) )
} }
@@ -2250,7 +2249,7 @@ module PrivateDjango {
DataFlow::Node value; DataFlow::Node value;
DjangoResponseHeaderSubscriptWrite() { DjangoResponseHeaderSubscriptWrite() {
exists(Cfg::SubscriptNode subscript, DataFlow::AttrRead headerLookup | exists(SubscriptNode subscript, DataFlow::AttrRead headerLookup |
// To give `this` a value, we need to choose between either LHS or RHS, // To give `this` a value, we need to choose between either LHS or RHS,
// and just go with the LHS // and just go with the LHS
this.asCfgNode() = subscript this.asCfgNode() = subscript
@@ -2262,7 +2261,7 @@ module PrivateDjango {
| |
headerLookup.flowsTo(subscriptObj) headerLookup.flowsTo(subscriptObj)
) and ) and
value.asCfgNode() = subscript.(Cfg::DefinitionNode).getValue() and value.asCfgNode() = subscript.(DefinitionNode).getValue() and
index.asCfgNode() = subscript.getIndex() index.asCfgNode() = subscript.getIndex()
) )
} }
@@ -2285,14 +2284,14 @@ module PrivateDjango {
DataFlow::Node value; DataFlow::Node value;
DjangoResponseSubscriptWrite() { DjangoResponseSubscriptWrite() {
exists(Cfg::SubscriptNode subscript | exists(SubscriptNode subscript |
// To give `this` a value, we need to choose between either LHS or RHS, // To give `this` a value, we need to choose between either LHS or RHS,
// and just go with the LHS // and just go with the LHS
this.asCfgNode() = subscript this.asCfgNode() = subscript
| |
subscript.getObject() = subscript.getObject() =
DjangoImpl::DjangoHttp::Response::HttpResponse::instance().asCfgNode() and DjangoImpl::DjangoHttp::Response::HttpResponse::instance().asCfgNode() and
value.asCfgNode() = subscript.(Cfg::DefinitionNode).getValue() and value.asCfgNode() = subscript.(DefinitionNode).getValue() and
index.asCfgNode() = subscript.getIndex() index.asCfgNode() = subscript.getIndex()
) )
} }
@@ -2427,7 +2426,7 @@ module PrivateDjango {
/** Gets a reference to the result of calling the `as_view` classmethod of this class. */ /** Gets a reference to the result of calling the `as_view` classmethod of this class. */
private DataFlow::TypeTrackingNode asViewResult(DataFlow::TypeTracker t) { private DataFlow::TypeTrackingNode asViewResult(DataFlow::TypeTracker t) {
t.start() and t.start() and
result.asCfgNode().(Cfg::CallNode).getFunction() = this.asViewRef().asCfgNode() result.asCfgNode().(CallNode).getFunction() = this.asViewRef().asCfgNode()
or or
exists(DataFlow::TypeTracker t2 | result = this.asViewResult(t2).track(t2, t)) exists(DataFlow::TypeTracker t2 | result = this.asViewResult(t2).track(t2, t))
} }

View File

@@ -4,7 +4,6 @@
*/ */
private import python private import python
private import semmle.python.controlflow.internal.Cfg as Cfg
private import semmle.python.dataflow.new.DataFlow private import semmle.python.dataflow.new.DataFlow
private import semmle.python.dataflow.new.RemoteFlowSources private import semmle.python.dataflow.new.RemoteFlowSources
private import semmle.python.dataflow.new.TaintTracking private import semmle.python.dataflow.new.TaintTracking
@@ -442,7 +441,7 @@ module FastApi {
DataFlow::Node value; DataFlow::Node value;
HeaderSubscriptWrite() { HeaderSubscriptWrite() {
exists(Cfg::SubscriptNode subscript, DataFlow::AttrRead headerLookup | exists(SubscriptNode subscript, DataFlow::AttrRead headerLookup |
// To give `this` a value, we need to choose between either LHS or RHS, // To give `this` a value, we need to choose between either LHS or RHS,
// and just go with the LHS // and just go with the LHS
this.asCfgNode() = subscript this.asCfgNode() = subscript
@@ -451,7 +450,7 @@ module FastApi {
exists(DataFlow::Node subscriptObj | subscriptObj.asCfgNode() = subscript.getObject() | exists(DataFlow::Node subscriptObj | subscriptObj.asCfgNode() = subscript.getObject() |
headerLookup.flowsTo(subscriptObj) headerLookup.flowsTo(subscriptObj)
) and ) and
value.asCfgNode() = subscript.(Cfg::DefinitionNode).getValue() and value.asCfgNode() = subscript.(DefinitionNode).getValue() and
index.asCfgNode() = subscript.getIndex() index.asCfgNode() = subscript.getIndex()
) )
} }

View File

@@ -4,7 +4,6 @@
*/ */
import python import python
private import semmle.python.controlflow.internal.Cfg as Cfg
import semmle.python.dataflow.new.RemoteFlowSources import semmle.python.dataflow.new.RemoteFlowSources
import semmle.python.dataflow.new.TaintTracking import semmle.python.dataflow.new.TaintTracking
import semmle.python.ApiGraphs import semmle.python.ApiGraphs
@@ -52,9 +51,9 @@ module Gradio {
// limit only to lists of parameters given to `inputs`. // limit only to lists of parameters given to `inputs`.
( (
( (
call.getKeywordParameter("inputs").asSink().asCfgNode() instanceof Cfg::ListNode call.getKeywordParameter("inputs").asSink().asCfgNode() instanceof ListNode
or or
call.getParameter(1).asSink().asCfgNode() instanceof Cfg::ListNode call.getParameter(1).asSink().asCfgNode() instanceof ListNode
) and ) and
( (
this = call.getKeywordParameter("inputs").getASubscript().getAValueReachingSink() this = call.getKeywordParameter("inputs").getASubscript().getAValueReachingSink()
@@ -76,8 +75,8 @@ module Gradio {
exists(GradioInput call | exists(GradioInput call |
this = call.getParameter(0, "fn").getParameter(_).asSource() and this = call.getParameter(0, "fn").getParameter(_).asSource() and
// exclude lists of parameters given to `inputs` // exclude lists of parameters given to `inputs`
not call.getKeywordParameter("inputs").asSink().asCfgNode() instanceof Cfg::ListNode and not call.getKeywordParameter("inputs").asSink().asCfgNode() instanceof ListNode and
not call.getParameter(1).asSink().asCfgNode() instanceof Cfg::ListNode not call.getParameter(1).asSink().asCfgNode() instanceof ListNode
) )
} }
@@ -106,16 +105,16 @@ module Gradio {
// handle cases where there are multiple arguments passed as a list to `inputs` // handle cases where there are multiple arguments passed as a list to `inputs`
( (
( (
node.getKeywordParameter("inputs").asSink().asCfgNode() instanceof Cfg::ListNode node.getKeywordParameter("inputs").asSink().asCfgNode() instanceof ListNode
or or
node.getParameter(1).asSink().asCfgNode() instanceof Cfg::ListNode node.getParameter(1).asSink().asCfgNode() instanceof ListNode
) and ) and
exists(int i | nodeTo = node.getParameter(0, "fn").getParameter(i).asSource() | exists(int i | nodeTo = node.getParameter(0, "fn").getParameter(i).asSource() |
nodeFrom.asCfgNode() = nodeFrom.asCfgNode() =
node.getKeywordParameter("inputs").asSink().asCfgNode().(Cfg::ListNode).getElement(i) node.getKeywordParameter("inputs").asSink().asCfgNode().(ListNode).getElement(i)
or or
nodeFrom.asCfgNode() = nodeFrom.asCfgNode() =
node.getParameter(1).asSink().asCfgNode().(Cfg::ListNode).getElement(i) node.getParameter(1).asSink().asCfgNode().(ListNode).getElement(i)
) )
) )
) )

View File

@@ -4,7 +4,6 @@
*/ */
private import python private import python
private import semmle.python.controlflow.internal.Cfg as Cfg
private import semmle.python.dataflow.new.DataFlow private import semmle.python.dataflow.new.DataFlow
private import semmle.python.dataflow.new.TaintTracking private import semmle.python.dataflow.new.TaintTracking
private import semmle.python.Concepts private import semmle.python.Concepts
@@ -47,7 +46,7 @@ module MarkupSafeModel {
/** A direct instantiation of `markupsafe.Markup`. */ /** A direct instantiation of `markupsafe.Markup`. */
private class ClassInstantiation extends InstanceSource, DataFlow::CallCfgNode { private class ClassInstantiation extends InstanceSource, DataFlow::CallCfgNode {
override Cfg::CallNode node; override CallNode node;
ClassInstantiation() { this = classRef().getACall() } ClassInstantiation() { this = classRef().getACall() }
} }
@@ -65,7 +64,7 @@ module MarkupSafeModel {
/** A string concatenation with a `markupsafe.Markup` involved. */ /** A string concatenation with a `markupsafe.Markup` involved. */
class StringConcat extends Markup::InstanceSource, DataFlow::CfgNode { class StringConcat extends Markup::InstanceSource, DataFlow::CfgNode {
override Cfg::BinaryExprNode node; override BinaryExprNode node;
StringConcat() { StringConcat() {
node.getOp() instanceof Add and node.getOp() instanceof Add and
@@ -80,7 +79,7 @@ module MarkupSafeModel {
/** A %-style string format with `markupsafe.Markup` as the format string. */ /** A %-style string format with `markupsafe.Markup` as the format string. */
class PercentStringFormat extends Markup::InstanceSource, DataFlow::CfgNode { class PercentStringFormat extends Markup::InstanceSource, DataFlow::CfgNode {
override Cfg::BinaryExprNode node; override BinaryExprNode node;
PercentStringFormat() { PercentStringFormat() {
node.getOp() instanceof Mod and node.getOp() instanceof Mod and

View File

@@ -7,7 +7,6 @@
*/ */
private import python private import python
private import semmle.python.controlflow.internal.Cfg as Cfg
private import semmle.python.Concepts private import semmle.python.Concepts
private import semmle.python.ApiGraphs private import semmle.python.ApiGraphs
private import semmle.python.frameworks.data.ModelsAsData private import semmle.python.frameworks.data.ModelsAsData
@@ -57,7 +56,7 @@ module Pycurl {
{ {
OutgoingRequestCall() { OutgoingRequestCall() {
this = setopt().getACall() and this = setopt().getACall() and
this.getArg(0).asCfgNode().(Cfg::AttrNode).getName() = "URL" this.getArg(0).asCfgNode().(AttrNode).getName() = "URL"
} }
override DataFlow::Node getAUrlPart() { override DataFlow::Node getAUrlPart() {
@@ -82,7 +81,7 @@ module Pycurl {
private class CurlSslCall extends Http::Client::Request::Range instanceof DataFlow::CallCfgNode { private class CurlSslCall extends Http::Client::Request::Range instanceof DataFlow::CallCfgNode {
CurlSslCall() { CurlSslCall() {
this = setopt().getACall() and this = setopt().getACall() and
this.getArg(0).asCfgNode().(Cfg::AttrNode).getName() = ["SSL_VERIFYPEER", "SSL_VERIFYHOST"] this.getArg(0).asCfgNode().(AttrNode).getName() = ["SSL_VERIFYPEER", "SSL_VERIFYHOST"]
} }
override DataFlow::Node getAUrlPart() { none() } override DataFlow::Node getAUrlPart() { none() }

View File

@@ -7,7 +7,6 @@
*/ */
private import python private import python
private import semmle.python.controlflow.internal.Cfg as Cfg
private import semmle.python.dataflow.new.DataFlow private import semmle.python.dataflow.new.DataFlow
private import semmle.python.dataflow.new.TaintTracking private import semmle.python.dataflow.new.TaintTracking
private import semmle.python.Concepts private import semmle.python.Concepts
@@ -94,7 +93,7 @@ module Pydantic {
// be a Pydantic model. So `model[0]` will be an overapproximation, but should not // be a Pydantic model. So `model[0]` will be an overapproximation, but should not
// really cause problems (since we don't expect real code to contain such accesses) // really cause problems (since we don't expect real code to contain such accesses)
nodeFrom = instance() and nodeFrom = instance() and
nodeTo.asCfgNode().(Cfg::SubscriptNode).getObject() = nodeFrom.asCfgNode() nodeTo.asCfgNode().(SubscriptNode).getObject() = nodeFrom.asCfgNode()
} }
/** /**

View File

@@ -6,7 +6,6 @@ overlay[local?]
module; module;
private import python private import python
private import semmle.python.controlflow.internal.Cfg as Cfg
private import semmle.python.dataflow.new.DataFlow private import semmle.python.dataflow.new.DataFlow
private import semmle.python.dataflow.new.TaintTracking private import semmle.python.dataflow.new.TaintTracking
private import semmle.python.dataflow.new.RemoteFlowSources private import semmle.python.dataflow.new.RemoteFlowSources
@@ -1247,7 +1246,7 @@ module StdlibPrivate {
/** An additional taint step for calls to `os.path.join` */ /** An additional taint step for calls to `os.path.join` */
private class OsPathJoinCallAdditionalTaintStep extends TaintTracking::AdditionalTaintStep { private class OsPathJoinCallAdditionalTaintStep extends TaintTracking::AdditionalTaintStep {
override predicate step(DataFlow::Node nodeFrom, DataFlow::Node nodeTo) { override predicate step(DataFlow::Node nodeFrom, DataFlow::Node nodeTo) {
exists(Cfg::CallNode call | exists(CallNode call |
nodeTo.asCfgNode() = call and nodeTo.asCfgNode() = call and
call = OS::OsPath::join().getACall().asCfgNode() and call = OS::OsPath::join().getACall().asCfgNode() and
call.getAnArg() = nodeFrom.asCfgNode() call.getAnArg() = nodeFrom.asCfgNode()
@@ -1318,13 +1317,13 @@ module StdlibPrivate {
// run, so if we're able to, we only mark the first element as the command // run, so if we're able to, we only mark the first element as the command
// (and not the arguments to the command). // (and not the arguments to the command).
// //
result.asCfgNode() = arg_args.asCfgNode().(Cfg::SequenceNode).getElement(0) result.asCfgNode() = arg_args.asCfgNode().(SequenceNode).getElement(0)
or or
// Either the "args" argument is not a sequence (which is valid) or we where // Either the "args" argument is not a sequence (which is valid) or we where
// just not able to figure it out. Simply mark the "args" argument as the // just not able to figure it out. Simply mark the "args" argument as the
// command. // command.
// //
not arg_args.asCfgNode() instanceof Cfg::SequenceNode and not arg_args.asCfgNode() instanceof SequenceNode and
result = arg_args result = arg_args
) )
) )
@@ -1543,7 +1542,7 @@ module StdlibPrivate {
* See https://docs.python.org/3/library/functions.html#eval * See https://docs.python.org/3/library/functions.html#eval
*/ */
private class BuiltinsEvalCall extends CodeExecution::Range, DataFlow::CallCfgNode { private class BuiltinsEvalCall extends CodeExecution::Range, DataFlow::CallCfgNode {
override Cfg::CallNode node; override CallNode node;
BuiltinsEvalCall() { this = API::builtin("eval").getACall() } BuiltinsEvalCall() { this = API::builtin("eval").getACall() }
@@ -1924,7 +1923,7 @@ module StdlibPrivate {
nodeFrom = instance().getAValueReachableFromSource() and nodeFrom = instance().getAValueReachableFromSource() and
nodeTo = [getvalueRef(), getfirstRef(), getlistRef()].getAValueReachableFromSource() nodeTo = [getvalueRef(), getfirstRef(), getlistRef()].getAValueReachableFromSource()
or or
nodeFrom.asCfgNode() = nodeTo.asCfgNode().(Cfg::CallNode).getFunction() and nodeFrom.asCfgNode() = nodeTo.asCfgNode().(CallNode).getFunction() and
( (
nodeFrom = getvalueRef().getAValueReachableFromSource() and nodeFrom = getvalueRef().getAValueReachableFromSource() and
nodeTo = getvalueResult().asSource() nodeTo = getvalueResult().asSource()
@@ -1940,7 +1939,7 @@ module StdlibPrivate {
nodeFrom in [ nodeFrom in [
instance().getAValueReachableFromSource(), fieldList().getAValueReachableFromSource() instance().getAValueReachableFromSource(), fieldList().getAValueReachableFromSource()
] and ] and
nodeTo.asCfgNode().(Cfg::SubscriptNode).getObject() = nodeFrom.asCfgNode() nodeTo.asCfgNode().(SubscriptNode).getObject() = nodeFrom.asCfgNode()
or or
// Attributes on Field // Attributes on Field
nodeFrom = field().getAValueReachableFromSource() and nodeFrom = field().getAValueReachableFromSource() and
@@ -2255,8 +2254,8 @@ module StdlibPrivate {
DataFlow::CfgNode DataFlow::CfgNode
{ {
WsgirefSimpleServerApplicationReturn() { WsgirefSimpleServerApplicationReturn() {
exists(Return ret | exists(WsgirefSimpleServerApplication requestHandler, Return ret |
ret.getScope() = any(WsgirefSimpleServerApplication requestHandler) and ret.getScope() = requestHandler and
node.getNode() = ret.getValue() node.getNode() = ret.getValue()
) )
} }
@@ -2339,9 +2338,9 @@ module StdlibPrivate {
DataFlow::Node value; DataFlow::Node value;
HeaderWriteSubscript() { HeaderWriteSubscript() {
exists(Cfg::SubscriptNode subscript | exists(SubscriptNode subscript |
this.asCfgNode() = subscript and this.asCfgNode() = subscript and
value.asCfgNode() = subscript.(Cfg::DefinitionNode).getValue() and value.asCfgNode() = subscript.(DefinitionNode).getValue() and
name.asCfgNode() = subscript.getIndex() and name.asCfgNode() = subscript.getIndex() and
subscript.getObject() = instance().asCfgNode() subscript.getObject() = instance().asCfgNode()
) )
@@ -2683,7 +2682,7 @@ module StdlibPrivate {
or or
// Data injection // Data injection
// Special handling of the `/` operator // Special handling of the `/` operator
exists(Cfg::BinaryExprNode slash, DataFlow::Node pathOperand, DataFlow::TypeTracker t2 | exists(BinaryExprNode slash, DataFlow::Node pathOperand, DataFlow::TypeTracker t2 |
slash.getOp() instanceof Div and slash.getOp() instanceof Div and
pathOperand.asCfgNode() = slash.getAnOperand() and pathOperand.asCfgNode() = slash.getAnOperand() and
pathlibPath(t2).flowsTo(pathOperand) and pathlibPath(t2).flowsTo(pathOperand) and
@@ -2808,7 +2807,7 @@ module StdlibPrivate {
pathlibPath().flowsTo(nodeTo) and pathlibPath().flowsTo(nodeTo) and
( (
// Special handling of the `/` operator // Special handling of the `/` operator
exists(Cfg::BinaryExprNode slash, DataFlow::Node pathOperand | exists(BinaryExprNode slash, DataFlow::Node pathOperand |
slash.getOp() instanceof Div and slash.getOp() instanceof Div and
pathOperand.asCfgNode() = slash.getAnOperand() and pathOperand.asCfgNode() = slash.getAnOperand() and
pathlibPath().flowsTo(pathOperand) pathlibPath().flowsTo(pathOperand)
@@ -4606,9 +4605,9 @@ module StdlibPrivate {
} }
override predicate propagatesFlow(string input, string output, boolean preservesValue) { override predicate propagatesFlow(string input, string output, boolean preservesValue) {
exists(Cfg::CallNode c, string name, Cfg::ControlFlowNode n, DataFlow::AttributeContent ac | exists(CallNode c, string name, ControlFlowNode n, DataFlow::AttributeContent ac |
c.getFunction().(Cfg::NameNode).getId() = "replace" or c.getFunction().(NameNode).getId() = "replace" or
c.getFunction().(Cfg::AttrNode).getName() = "replace" c.getFunction().(AttrNode).getName() = "replace"
| |
n = c.getArgByName(name) and n = c.getArgByName(name) and
ac.getAttribute() = name and ac.getAttribute() = name and
@@ -5172,10 +5171,10 @@ module StdlibPrivate {
* See https://docs.python.org/3.9/library/stdtypes.html#str.startswith * See https://docs.python.org/3.9/library/stdtypes.html#str.startswith
*/ */
private class StartswithCall extends Path::SafeAccessCheck::Range { private class StartswithCall extends Path::SafeAccessCheck::Range {
StartswithCall() { this.(Cfg::CallNode).getFunction().(Cfg::AttrNode).getName() = "startswith" } StartswithCall() { this.(CallNode).getFunction().(AttrNode).getName() = "startswith" }
override predicate checks(Cfg::ControlFlowNode node, boolean branch) { override predicate checks(ControlFlowNode node, boolean branch) {
node = this.(Cfg::CallNode).getFunction().(Cfg::AttrNode).getObject() and node = this.(CallNode).getFunction().(AttrNode).getObject() and
branch = true branch = true
} }
} }

View File

@@ -8,7 +8,6 @@
*/ */
private import python private import python
private import semmle.python.controlflow.internal.Cfg as Cfg
private import semmle.python.Concepts private import semmle.python.Concepts
private import semmle.python.ApiGraphs private import semmle.python.ApiGraphs
private import semmle.python.security.dataflow.UrlRedirectCustomizations private import semmle.python.security.dataflow.UrlRedirectCustomizations
@@ -92,7 +91,7 @@ private module Urllib {
* A read of the `netloc` attribute of a parsed URL as returned by `urllib.parse.urlparse`, * A read of the `netloc` attribute of a parsed URL as returned by `urllib.parse.urlparse`,
* which is being checked in a way that is relevant for URL redirection vulnerabilities. * which is being checked in a way that is relevant for URL redirection vulnerabilities.
*/ */
private predicate netlocCheck(DataFlow::GuardNode g, Cfg::ControlFlowNode node, boolean branch) { private predicate netlocCheck(DataFlow::GuardNode g, ControlFlowNode node, boolean branch) {
exists(DataFlow::CallCfgNode urlParseCall, DataFlow::AttrRead netlocRead | exists(DataFlow::CallCfgNode urlParseCall, DataFlow::AttrRead netlocRead |
urlParseCall = getUrlParseCall() and urlParseCall = getUrlParseCall() and
netlocRead = urlParseCall.getAnAttributeRead("netloc") and netlocRead = urlParseCall.getAnAttributeRead("netloc") and

View File

@@ -4,7 +4,6 @@
*/ */
private import python private import python
private import semmle.python.controlflow.internal.Cfg as Cfg
private import semmle.python.dataflow.new.DataFlow private import semmle.python.dataflow.new.DataFlow
private import semmle.python.dataflow.new.RemoteFlowSources private import semmle.python.dataflow.new.RemoteFlowSources
private import semmle.python.dataflow.new.TaintTracking private import semmle.python.dataflow.new.TaintTracking
@@ -73,9 +72,9 @@ module Tornado {
DataFlow::Node value; DataFlow::Node value;
TornadoHeaderSubscriptWrite() { TornadoHeaderSubscriptWrite() {
exists(Cfg::SubscriptNode subscript | exists(SubscriptNode subscript |
subscript.getObject() = instance().asCfgNode() and subscript.getObject() = instance().asCfgNode() and
value.asCfgNode() = subscript.(Cfg::DefinitionNode).getValue() and value.asCfgNode() = subscript.(DefinitionNode).getValue() and
index.asCfgNode() = subscript.getIndex() and index.asCfgNode() = subscript.getIndex() and
this.asCfgNode() = subscript this.asCfgNode() = subscript
) )
@@ -423,7 +422,7 @@ module Tornado {
// be able to do something more structured for providing modeling of the members // be able to do something more structured for providing modeling of the members
// of a container-object. // of a container-object.
exists(DataFlow::AttrRead files | files.accesses(instance(), "cookies") | exists(DataFlow::AttrRead files | files.accesses(instance(), "cookies") |
this.asCfgNode().(Cfg::SubscriptNode).getObject() = files.asCfgNode() this.asCfgNode().(SubscriptNode).getObject() = files.asCfgNode()
or or
this.(DataFlow::MethodCallNode).calls(files, "get") this.(DataFlow::MethodCallNode).calls(files, "get")
) )
@@ -480,20 +479,20 @@ module Tornado {
// routing // routing
// --------------------------------------------------------------------------- // ---------------------------------------------------------------------------
/** Gets a sequence that defines a number of route rules */ /** Gets a sequence that defines a number of route rules */
Cfg::SequenceNode routeSetupRuleList() { SequenceNode routeSetupRuleList() {
exists(Cfg::CallNode call | exists(CallNode call |
call = any(TornadoModule::Web::Application::ClassInstantiation c).asCfgNode() call = any(TornadoModule::Web::Application::ClassInstantiation c).asCfgNode()
| |
result in [call.getArg(0), call.getArgByName("handlers")] result in [call.getArg(0), call.getArgByName("handlers")]
) )
or or
exists(Cfg::CallNode call | exists(CallNode call |
call.getFunction() = TornadoModule::Web::Application::add_handlers().asCfgNode() call.getFunction() = TornadoModule::Web::Application::add_handlers().asCfgNode()
| |
result in [call.getArg(1), call.getArgByName("host_handlers")] result in [call.getArg(1), call.getArgByName("host_handlers")]
) )
or or
result = routeSetupRuleList().getElement(_).(Cfg::TupleNode).getElement(1) result = routeSetupRuleList().getElement(_).(TupleNode).getElement(1)
} }
/** A tornado route setup. */ /** A tornado route setup. */
@@ -516,12 +515,12 @@ module Tornado {
/** A route setup using a tuple. */ /** A route setup using a tuple. */
private class TornadoTupleRouteSetup extends TornadoRouteSetup, DataFlow::CfgNode { private class TornadoTupleRouteSetup extends TornadoRouteSetup, DataFlow::CfgNode {
override Cfg::TupleNode node; override TupleNode node;
TornadoTupleRouteSetup() { TornadoTupleRouteSetup() {
node = routeSetupRuleList().getElement(_) and node = routeSetupRuleList().getElement(_) and
count(node.getElement(_)) = 2 and count(node.getElement(_)) = 2 and
not node.getElement(1) instanceof Cfg::SequenceNode not node.getElement(1) instanceof SequenceNode
} }
override DataFlow::Node getUrlPatternArg() { result.asCfgNode() = node.getElement(0) } override DataFlow::Node getUrlPatternArg() { result.asCfgNode() = node.getElement(0) }

View File

@@ -6,7 +6,6 @@
*/ */
private import python private import python
private import semmle.python.controlflow.internal.Cfg as Cfg
private import semmle.python.dataflow.new.DataFlow private import semmle.python.dataflow.new.DataFlow
private import semmle.python.dataflow.new.TaintTracking private import semmle.python.dataflow.new.TaintTracking
private import semmle.python.ApiGraphs private import semmle.python.ApiGraphs
@@ -222,9 +221,9 @@ module Werkzeug {
DataFlow::Node value; DataFlow::Node value;
HeaderWriteSubscript() { HeaderWriteSubscript() {
exists(Cfg::SubscriptNode subscript | exists(SubscriptNode subscript |
this.asCfgNode() = subscript and this.asCfgNode() = subscript and
value.asCfgNode() = subscript.(Cfg::DefinitionNode).getValue() and value.asCfgNode() = subscript.(DefinitionNode).getValue() and
name.asCfgNode() = subscript.getIndex() and name.asCfgNode() = subscript.getIndex() and
subscript.getObject() = instance().asCfgNode() subscript.getObject() = instance().asCfgNode()
) )

View File

@@ -8,7 +8,6 @@
*/ */
private import python private import python
private import semmle.python.controlflow.internal.Cfg as Cfg
private import semmle.python.dataflow.new.DataFlow private import semmle.python.dataflow.new.DataFlow
private import semmle.python.Concepts private import semmle.python.Concepts
private import semmle.python.ApiGraphs private import semmle.python.ApiGraphs
@@ -29,7 +28,7 @@ private module Yaml {
* See https://pyyaml.org/wiki/PyYAMLDocumentation (you will have to scroll down). * See https://pyyaml.org/wiki/PyYAMLDocumentation (you will have to scroll down).
*/ */
private class YamlLoadCall extends Decoding::Range, DataFlow::CallCfgNode { private class YamlLoadCall extends Decoding::Range, DataFlow::CallCfgNode {
override Cfg::CallNode node; override CallNode node;
string func_name; string func_name;
YamlLoadCall() { YamlLoadCall() {

View File

@@ -4,7 +4,6 @@
*/ */
private import python private import python
private import semmle.python.controlflow.internal.Cfg as Cfg
private import semmle.python.dataflow.new.DataFlow private import semmle.python.dataflow.new.DataFlow
private import semmle.python.dataflow.new.TaintTracking private import semmle.python.dataflow.new.TaintTracking
private import semmle.python.Concepts private import semmle.python.Concepts
@@ -112,7 +111,7 @@ module Yarl {
} }
private predicate yarlUrlIsAbsoluteCall( private predicate yarlUrlIsAbsoluteCall(
DataFlow::GuardNode g, Cfg::ControlFlowNode node, boolean branch DataFlow::GuardNode g, ControlFlowNode node, boolean branch
) { ) {
exists(ClassInstantiation instance, DataFlow::MethodCallNode call | exists(ClassInstantiation instance, DataFlow::MethodCallNode call |
call.calls(instance, "is_absolute") and call.calls(instance, "is_absolute") and

View File

@@ -11,7 +11,6 @@ private import semmle.python.dataflow.new.internal.ImportResolution
private import semmle.python.ApiGraphs private import semmle.python.ApiGraphs
private import semmle.python.filters.Tests private import semmle.python.filters.Tests
private import semmle.python.Module private import semmle.python.Module
private import semmle.python.controlflow.internal.Cfg as Cfg
// very much inspired by the draft at https://github.com/github/codeql/pull/5632 // very much inspired by the draft at https://github.com/github/codeql/pull/5632
module NotExposed { module NotExposed {
@@ -207,7 +206,7 @@ module NotExposed {
string relevantName, Location loc string relevantName, Location loc
) { ) {
loc = mod.getLocation() and loc = mod.getLocation() and
exists(API::Node relevantClass, Cfg::ControlFlowNode value | exists(API::Node relevantClass, ControlFlowNode value |
relevantClass = newOrExistingModeling(spec).getASubclass*() and relevantClass = newOrExistingModeling(spec).getASubclass*() and
ImportResolution::module_export(mod, relevantName, def) and ImportResolution::module_export(mod, relevantName, def) and
value = relevantClass.getAValueReachableFromSource().asCfgNode() and value = relevantClass.getAValueReachableFromSource().asCfgNode() and

View File

@@ -3,7 +3,6 @@
*/ */
import python import python
private import semmle.python.controlflow.internal.Cfg as Cfg
private import semmle.python.dataflow.new.DataFlow private import semmle.python.dataflow.new.DataFlow
private import semmle.python.Concepts as Concepts private import semmle.python.Concepts as Concepts
private import semmle.python.regex private import semmle.python.regex
@@ -79,7 +78,7 @@ private module FindRegexMode {
t.start() and t.start() and
exists(API::Node flag | flag_name = canonical_name(flag) and result = flag.asSource()) exists(API::Node flag | flag_name = canonical_name(flag) and result = flag.asSource())
or or
exists(Cfg::BinaryExprNode binop, DataFlow::Node operand | exists(BinaryExprNode binop, DataFlow::Node operand |
operand.getALocalSource() = re_flag_tracker(flag_name, t.continue()) and operand.getALocalSource() = re_flag_tracker(flag_name, t.continue()) and
operand.asCfgNode() = binop.getAnOperand() and operand.asCfgNode() = binop.getAnOperand() and
(binop.getOp() instanceof BitOr or binop.getOp() instanceof Add) and (binop.getOp() instanceof BitOr or binop.getOp() instanceof Add) and

View File

@@ -3,7 +3,6 @@
import python import python
import semmle.python.dataflow.new.DataFlow import semmle.python.dataflow.new.DataFlow
private import semmle.python.ApiGraphs private import semmle.python.ApiGraphs
private import semmle.python.controlflow.internal.Cfg as Cfg
/** /**
* INTERNAL: Do not use. * INTERNAL: Do not use.
@@ -30,7 +29,7 @@ private class TracebackFunctionCall extends ExceptionInfo, DataFlow::CallCfgNode
private class CaughtException extends ExceptionInfo { private class CaughtException extends ExceptionInfo {
CaughtException() { CaughtException() {
this.asExpr() = any(ExceptStmt s).getName() and this.asExpr() = any(ExceptStmt s).getName() and
this.asCfgNode().(Cfg::NameNode).defines(_) this.asCfgNode() = any(EssaNodeDefinition def).getDefiningNode()
} }
} }

View File

@@ -11,7 +11,6 @@ private import semmle.python.dataflow.new.RemoteFlowSources
private import semmle.python.dataflow.new.BarrierGuards private import semmle.python.dataflow.new.BarrierGuards
private import semmle.python.ApiGraphs private import semmle.python.ApiGraphs
private import semmle.python.frameworks.data.internal.ApiGraphModels private import semmle.python.frameworks.data.internal.ApiGraphModels
private import semmle.python.controlflow.internal.Cfg as Cfg
/** /**
* Provides default sources, sinks and sanitizers for detecting * Provides default sources, sinks and sanitizers for detecting
@@ -96,7 +95,7 @@ module ServerSideRequestForgery {
class StringConstructionAsFullUrlControlSanitizer extends FullUrlControlSanitizer { class StringConstructionAsFullUrlControlSanitizer extends FullUrlControlSanitizer {
StringConstructionAsFullUrlControlSanitizer() { StringConstructionAsFullUrlControlSanitizer() {
// string concat // string concat
exists(Cfg::BinaryExprNode add | exists(BinaryExprNode add |
add.getOp() instanceof Add and add.getOp() instanceof Add and
add.getRight() = this.asCfgNode() and add.getRight() = this.asCfgNode() and
not add.getLeft().getNode().(StringLiteral).getText().toLowerCase() in [ not add.getLeft().getNode().(StringLiteral).getText().toLowerCase() in [
@@ -105,7 +104,7 @@ module ServerSideRequestForgery {
) )
or or
// % formatting // % formatting
exists(Cfg::BinaryExprNode fmt | exists(BinaryExprNode fmt |
fmt.getOp() instanceof Mod and fmt.getOp() instanceof Mod and
fmt.getRight() = this.asCfgNode() and fmt.getRight() = this.asCfgNode() and
// detecting %-formatting is not super easy, so we simplify it to only handle // detecting %-formatting is not super easy, so we simplify it to only handle
@@ -156,9 +155,7 @@ module ServerSideRequestForgery {
} }
} }
private predicate stringRestriction( private predicate stringRestriction(DataFlow::GuardNode g, ControlFlowNode node, boolean branch) {
DataFlow::GuardNode g, Cfg::ControlFlowNode node, boolean branch
) {
exists(DataFlow::MethodCallNode call, DataFlow::Node strNode | exists(DataFlow::MethodCallNode call, DataFlow::Node strNode |
call.asCfgNode() = g and strNode.asCfgNode() = node call.asCfgNode() = g and strNode.asCfgNode() = node
| |

View File

@@ -9,7 +9,6 @@ private import semmle.python.dataflow.new.DataFlow
private import semmle.python.Concepts private import semmle.python.Concepts
private import semmle.python.dataflow.new.BarrierGuards private import semmle.python.dataflow.new.BarrierGuards
private import semmle.python.ApiGraphs private import semmle.python.ApiGraphs
private import semmle.python.controlflow.internal.Cfg as Cfg
/** /**
* Provides default sources, sinks and sanitizers for detecting * Provides default sources, sinks and sanitizers for detecting
@@ -140,8 +139,8 @@ module TarSlip {
* where `<check_path>` is any function matching `"%path"`. * where `<check_path>` is any function matching `"%path"`.
* `info` is assumed to be a `TarInfo` instance. * `info` is assumed to be a `TarInfo` instance.
*/ */
predicate tarFileInfoSanitizer(DataFlow::GuardNode g, Cfg::ControlFlowNode tarInfo, boolean branch) { predicate tarFileInfoSanitizer(DataFlow::GuardNode g, ControlFlowNode tarInfo, boolean branch) {
exists(Cfg::CallNode call, Cfg::AttrNode attr | exists(CallNode call, AttrNode attr |
g = call and g = call and
// We must test the name of the tar info object. // We must test the name of the tar info object.
attr = call.getAnArg() and attr = call.getAnArg() and
@@ -149,9 +148,9 @@ module TarSlip {
attr.getObject() = tarInfo attr.getObject() = tarInfo
| |
// The assumption that any test that matches %path is a sanitizer might be too broad. // The assumption that any test that matches %path is a sanitizer might be too broad.
call.getAChild*().(Cfg::AttrNode).getName().matches("%path") call.getAChild*().(AttrNode).getName().matches("%path")
or or
call.getAChild*().(Cfg::NameNode).getId().matches("%path") call.getAChild*().(NameNode).getId().matches("%path")
) and ) and
branch = false branch = false
} }

View File

@@ -5,7 +5,6 @@
*/ */
private import python private import python
private import semmle.python.controlflow.internal.Cfg as Cfg
private import semmle.python.dataflow.new.DataFlow private import semmle.python.dataflow.new.DataFlow
private import semmle.python.Concepts private import semmle.python.Concepts
private import semmle.python.ApiGraphs private import semmle.python.ApiGraphs
@@ -112,7 +111,7 @@ module UrlRedirect {
// Url redirection is a problem only if the user controls the prefix of the URL. // Url redirection is a problem only if the user controls the prefix of the URL.
// TODO: This is a copy of the taint-sanitizer from the old points-to query, which doesn't // TODO: This is a copy of the taint-sanitizer from the old points-to query, which doesn't
// cover formatting. // cover formatting.
exists(Cfg::BinaryExprNode string_concat | string_concat.getOp() instanceof Add | exists(BinaryExprNode string_concat | string_concat.getOp() instanceof Add |
string_concat.getRight() = this.asCfgNode() string_concat.getRight() = this.asCfgNode()
) )
} }

View File

@@ -1,5 +1,4 @@
import python import python
private import semmle.python.controlflow.internal.Cfg as Cfg
import semmle.python.dataflow.new.DataFlow import semmle.python.dataflow.new.DataFlow
private import semmle.python.dataflow.new.internal.DataFlowPrivate private import semmle.python.dataflow.new.internal.DataFlowPrivate
import FlowTest import FlowTest
@@ -24,7 +23,7 @@ import MakeTest<MakeTestSig<MaximalFlowTest>>
module MaximalFlowsConfig implements DataFlow::ConfigSig { module MaximalFlowsConfig implements DataFlow::ConfigSig {
predicate isSource(DataFlow::Node node) { predicate isSource(DataFlow::Node node) {
exists(node.getLocation().getFile().getRelativePath()) and exists(node.getLocation().getFile().getRelativePath()) and
not node.asCfgNode() instanceof Cfg::CallNode and not node.asCfgNode() instanceof CallNode and
not node.asCfgNode().getNode() instanceof Return and not node.asCfgNode().getNode() instanceof Return and
not node instanceof DataFlow::ParameterNode and not node instanceof DataFlow::ParameterNode and
not node instanceof DataFlow::PostUpdateNode and not node instanceof DataFlow::PostUpdateNode and
@@ -35,9 +34,9 @@ module MaximalFlowsConfig implements DataFlow::ConfigSig {
predicate isSink(DataFlow::Node node) { predicate isSink(DataFlow::Node node) {
exists(node.getLocation().getFile().getRelativePath()) and exists(node.getLocation().getFile().getRelativePath()) and
not any(Cfg::CallNode c).getArg(_) = node.asCfgNode() and not any(CallNode c).getArg(_) = node.asCfgNode() and
not isArgumentNode(node, _, _) and not isArgumentNode(node, _, _) and
not node.asCfgNode().(Cfg::NameNode).getId().matches("SINK%") and not node.asCfgNode().(NameNode).getId().matches("SINK%") and
not DataFlow::localFlowStep(node, _) not DataFlow::localFlowStep(node, _)
} }
} }

View File

@@ -1,5 +1,4 @@
import python import python
private import semmle.python.controlflow.internal.Cfg as Cfg
import utils.test.dataflow.FlowTest import utils.test.dataflow.FlowTest
import utils.test.dataflow.testConfig import utils.test.dataflow.testConfig
private import semmle.python.dataflow.new.internal.PrintNode private import semmle.python.dataflow.new.internal.PrintNode
@@ -20,7 +19,7 @@ query predicate missingAnnotationOnSink(Location location, string error, string
TestConfig::isSink(sink) and TestConfig::isSink(sink) and
// note: we only care about `SINK` and not `SINK_F`, so we have to reconstruct manually. // note: we only care about `SINK` and not `SINK_F`, so we have to reconstruct manually.
exists(DataFlow::CallCfgNode call | exists(DataFlow::CallCfgNode call |
call.getFunction().asCfgNode().(Cfg::NameNode).getId() = "SINK" and call.getFunction().asCfgNode().(NameNode).getId() = "SINK" and
(sink = call.getArg(_) or sink = call.getArgByName(_)) (sink = call.getArg(_) or sink = call.getArgByName(_))
) and ) and
location = sink.getLocation() and location = sink.getLocation() and

View File

@@ -1,5 +1,4 @@
import python import python
private import semmle.python.controlflow.internal.Cfg as Cfg
import utils.test.dataflow.FlowTest import utils.test.dataflow.FlowTest
import utils.test.dataflow.testTaintConfig import utils.test.dataflow.testTaintConfig
private import semmle.python.dataflow.new.internal.PrintNode private import semmle.python.dataflow.new.internal.PrintNode
@@ -19,7 +18,7 @@ query predicate missingAnnotationOnSink(Location location, string error, string
exists(DataFlow::Node sink | exists(DataFlow::Node sink |
exists(DataFlow::CallCfgNode call | exists(DataFlow::CallCfgNode call |
// note: we only care about `SINK` and not `SINK_F`, so we have to reconstruct manually. // note: we only care about `SINK` and not `SINK_F`, so we have to reconstruct manually.
call.getFunction().asCfgNode().(Cfg::NameNode).getId() = "SINK" and call.getFunction().asCfgNode().(NameNode).getId() = "SINK" and
(sink = call.getArg(_) or sink = call.getArgByName(_)) (sink = call.getArg(_) or sink = call.getArgByName(_))
) and ) and
location = sink.getLocation() and location = sink.getLocation() and

View File

@@ -1,5 +1,4 @@
import python import python
private import semmle.python.controlflow.internal.Cfg as Cfg
import semmle.python.dataflow.new.DataFlow import semmle.python.dataflow.new.DataFlow
import utils.test.InlineExpectationsTest import utils.test.InlineExpectationsTest
private import semmle.python.dataflow.new.internal.PrintNode private import semmle.python.dataflow.new.internal.PrintNode
@@ -50,7 +49,7 @@ private string fromValue(DataFlow::Node fromNode) {
pragma[inline] pragma[inline]
private string fromFunc(DataFlow::ArgumentNode fromNode) { private string fromFunc(DataFlow::ArgumentNode fromNode) {
result = fromNode.getCall().getNode().(Cfg::CallNode).getFunction().getNode().(Name).getId() result = fromNode.getCall().getNode().(CallNode).getFunction().getNode().(Name).getId()
} }
pragma[inline] pragma[inline]

View File

@@ -1,16 +1,15 @@
import python import python
private import semmle.python.controlflow.internal.Cfg as Cfg
private import semmle.python.dataflow.new.internal.PrintNode private import semmle.python.dataflow.new.internal.PrintNode
private import semmle.python.dataflow.new.internal.DataFlowPrivate as DataFlowPrivate private import semmle.python.dataflow.new.internal.DataFlowPrivate as DataFlowPrivate
private import semmle.python.ApiGraphs private import semmle.python.ApiGraphs
import utils.test.InlineExpectationsTest import utils.test.InlineExpectationsTest
signature module UnresolvedCallExpectationsSig { signature module UnresolvedCallExpectationsSig {
predicate unresolvedCall(Cfg::CallNode call); predicate unresolvedCall(CallNode call);
} }
module DefaultUnresolvedCallExpectations implements UnresolvedCallExpectationsSig { module DefaultUnresolvedCallExpectations implements UnresolvedCallExpectationsSig {
predicate unresolvedCall(Cfg::CallNode call) { predicate unresolvedCall(CallNode call) {
not exists(DataFlowPrivate::DataFlowCall dfc | not exists(DataFlowPrivate::DataFlowCall dfc |
exists(dfc.getCallable()) and dfc.getNode() = call exists(dfc.getCallable()) and dfc.getNode() = call
) and ) and
@@ -25,7 +24,7 @@ module MakeUnresolvedCallExpectations<UnresolvedCallExpectationsSig Impl> {
predicate hasActualResult(Location location, string element, string tag, string value) { predicate hasActualResult(Location location, string element, string tag, string value) {
exists(location.getFile().getRelativePath()) and exists(location.getFile().getRelativePath()) and
exists(Cfg::CallNode call | Impl::unresolvedCall(call) and call.injects(_) | exists(CallNode call | Impl::unresolvedCall(call) |
location = call.getLocation() and location = call.getLocation() and
tag = "unresolved_call" and tag = "unresolved_call" and
value = prettyExpr(call.getNode()) and value = prettyExpr(call.getNode()) and

View File

@@ -21,12 +21,11 @@
*/ */
private import python private import python
private import semmle.python.controlflow.internal.Cfg as Cfg
import semmle.python.dataflow.new.DataFlow import semmle.python.dataflow.new.DataFlow
module TestConfig implements DataFlow::ConfigSig { module TestConfig implements DataFlow::ConfigSig {
predicate isSource(DataFlow::Node node) { predicate isSource(DataFlow::Node node) {
node.(DataFlow::CfgNode).getNode().(Cfg::NameNode).getId() = "SOURCE" node.(DataFlow::CfgNode).getNode().(NameNode).getId() = "SOURCE"
or or
node.(DataFlow::CfgNode).getNode().getNode().(StringLiteral).getS() = "source" node.(DataFlow::CfgNode).getNode().getNode().(StringLiteral).getS() = "source"
or or
@@ -38,7 +37,7 @@ module TestConfig implements DataFlow::ConfigSig {
predicate isSink(DataFlow::Node node) { predicate isSink(DataFlow::Node node) {
exists(DataFlow::CallCfgNode call | exists(DataFlow::CallCfgNode call |
call.getFunction().asCfgNode().(Cfg::NameNode).getId() in ["SINK", "SINK_F"] and call.getFunction().asCfgNode().(NameNode).getId() in ["SINK", "SINK_F"] and
(node = call.getArg(_) or node = call.getArgByName(_)) and (node = call.getArg(_) or node = call.getArgByName(_)) and
not node = call.getArgByName("not_present_at_runtime") not node = call.getArgByName("not_present_at_runtime")
) )

View File

@@ -21,13 +21,12 @@
*/ */
private import python private import python
private import semmle.python.controlflow.internal.Cfg as Cfg
import semmle.python.dataflow.new.DataFlow import semmle.python.dataflow.new.DataFlow
import semmle.python.dataflow.new.TaintTracking import semmle.python.dataflow.new.TaintTracking
module TestConfig implements DataFlow::ConfigSig { module TestConfig implements DataFlow::ConfigSig {
predicate isSource(DataFlow::Node node) { predicate isSource(DataFlow::Node node) {
node.(DataFlow::CfgNode).getNode().(Cfg::NameNode).getId() = "SOURCE" node.(DataFlow::CfgNode).getNode().(NameNode).getId() = "SOURCE"
or or
node.(DataFlow::CfgNode).getNode().getNode().(StringLiteral).getS() = "source" node.(DataFlow::CfgNode).getNode().getNode().(StringLiteral).getS() = "source"
or or
@@ -38,8 +37,8 @@ module TestConfig implements DataFlow::ConfigSig {
} }
predicate isSink(DataFlow::Node node) { predicate isSink(DataFlow::Node node) {
exists(Cfg::CallNode call | exists(CallNode call |
call.getFunction().(Cfg::NameNode).getId() in ["SINK", "SINK_F"] and call.getFunction().(NameNode).getId() in ["SINK", "SINK_F"] and
node.(DataFlow::CfgNode).getNode() = call.getAnArg() node.(DataFlow::CfgNode).getNode() = call.getAnArg()
) )
} }

View File

@@ -12,8 +12,6 @@
import python import python
private import semmle.python.ApiGraphs private import semmle.python.ApiGraphs
private import semmle.python.controlflow.internal.Cfg as Cfg
private import semmle.python.Flow as Flow
API::Node iter() { result = API::builtin("iter") } API::Node iter() { result = API::builtin("iter") }
@@ -21,17 +19,17 @@ API::Node next() { result = API::builtin("next") }
API::Node stopIteration() { result = API::builtin("StopIteration") } API::Node stopIteration() { result = API::builtin("StopIteration") }
predicate call_to_iter(Flow::CallNode call, EssaVariable sequence) { predicate call_to_iter(CallNode call, EssaVariable sequence) {
call.getNode() = iter().getACall().asCfgNode().(Cfg::CallNode).getNode() and call = iter().getACall().asCfgNode() and
call.getArg(0) = sequence.getAUse() call.getArg(0) = sequence.getAUse()
} }
predicate call_to_next(Flow::CallNode call, Flow::ControlFlowNode iter) { predicate call_to_next(CallNode call, ControlFlowNode iter) {
call.getNode() = next().getACall().asCfgNode().(Cfg::CallNode).getNode() and call = next().getACall().asCfgNode() and
call.getArg(0) = iter call.getArg(0) = iter
} }
predicate call_to_next_has_default(Flow::CallNode call) { predicate call_to_next_has_default(CallNode call) {
exists(call.getArg(1)) or exists(call.getArgByName("default")) exists(call.getArg(1)) or exists(call.getArgByName("default"))
} }
@@ -51,14 +49,14 @@ predicate iter_not_exhausted(EssaVariable iterator) {
) )
} }
predicate stop_iteration_handled(Flow::CallNode call) { predicate stop_iteration_handled(CallNode call) {
exists(Try t | exists(Try t |
t.containsInScope(call.getNode()) and t.containsInScope(call.getNode()) and
t.getAHandler().getType() = stopIteration().getAValueReachableFromSource().asExpr() t.getAHandler().getType() = stopIteration().getAValueReachableFromSource().asExpr()
) )
} }
from Flow::CallNode call from CallNode call
where where
call_to_next(call, _) and call_to_next(call, _) and
not call_to_next_has_default(call) and not call_to_next_has_default(call) and

View File

@@ -11,9 +11,8 @@
import python import python
private import semmle.python.ApiGraphs private import semmle.python.ApiGraphs
private import semmle.python.controlflow.internal.Cfg as Cfg
from Cfg::CallNode call from CallNode call
where where
major_version() = 2 and major_version() = 2 and
call = API::builtin("apply").getACall().asCfgNode() call = API::builtin("apply").getACall().asCfgNode()

View File

@@ -15,7 +15,6 @@
import python import python
import semmle.python.dataflow.new.DataFlow import semmle.python.dataflow.new.DataFlow
import semmle.python.dataflow.new.internal.DataFlowDispatch import semmle.python.dataflow.new.internal.DataFlowDispatch
private import semmle.python.controlflow.internal.Cfg as Cfg
import codeql.util.Option import codeql.util.Option
/** Holds if `base` is overridden by `sub` */ /** Holds if `base` is overridden by `sub` */
@@ -144,7 +143,7 @@ predicate ignore(Function f) {
/** Gets a function that `call` may resolve to. */ /** Gets a function that `call` may resolve to. */
Function resolveCall(Call call) { Function resolveCall(Call call) {
exists(DataFlowCall dfc | call = dfc.getNode().(Cfg::CallNode).getNode() | exists(DataFlowCall dfc | call = dfc.getNode().(CallNode).getNode() |
result = viableCallable(dfc).(DataFlowFunction).getScope() result = viableCallable(dfc).(DataFlowFunction).getScope()
) )
} }

View File

@@ -4,7 +4,6 @@ import python
import semmle.python.dataflow.new.internal.DataFlowDispatch import semmle.python.dataflow.new.internal.DataFlowDispatch
import semmle.python.ApiGraphs import semmle.python.ApiGraphs
private import semmle.python.dataflow.new.internal.ReExposedInstance private import semmle.python.dataflow.new.internal.ReExposedInstance
private import semmle.python.controlflow.internal.Cfg as Cfg
/** A CFG node where a file is opened. */ /** A CFG node where a file is opened. */
abstract class FileOpenSource extends DataFlow::CfgNode { } abstract class FileOpenSource extends DataFlow::CfgNode { }
@@ -82,14 +81,12 @@ abstract class FileClose extends DataFlow::CfgNode {
} }
} }
private predicate bbSuccessor(Cfg::BasicBlock src, Cfg::BasicBlock sink) { private predicate bbSuccessor(BasicBlock src, BasicBlock sink) { sink = src.getASuccessor() }
sink = src.getASuccessor()
}
private predicate bbReachableStrict(Cfg::BasicBlock src, Cfg::BasicBlock sink) = private predicate bbReachableStrict(BasicBlock src, BasicBlock sink) =
fastTC(bbSuccessor/2)(src, sink) fastTC(bbSuccessor/2)(src, sink)
private predicate bbReachableRefl(Cfg::BasicBlock src, Cfg::BasicBlock sink) { private predicate bbReachableRefl(BasicBlock src, BasicBlock sink) {
bbReachableStrict(src, sink) or src = sink bbReachableStrict(src, sink) or src = sink
} }

View File

@@ -10,7 +10,6 @@ private import semmle.python.dataflow.new.RemoteFlowSources
private import semmle.python.ApiGraphs private import semmle.python.ApiGraphs
private import semmle.python.dataflow.new.internal.DataFlowPrivate as DataFlowPrivate private import semmle.python.dataflow.new.internal.DataFlowPrivate as DataFlowPrivate
private import semmle.python.dataflow.new.internal.TaintTrackingPrivate as TaintTrackingPrivate private import semmle.python.dataflow.new.internal.TaintTrackingPrivate as TaintTrackingPrivate
private import semmle.python.controlflow.internal.Cfg as Cfg
/** /**
* An external API that is considered "safe" from a security perspective. * An external API that is considered "safe" from a security perspective.
@@ -72,7 +71,7 @@ string apiNodeToStringRepr(API::Node node) {
) )
} }
predicate resolvedCall(Cfg::CallNode call) { predicate resolvedCall(CallNode call) {
DataFlowPrivate::resolveCall(call, _, _) or DataFlowPrivate::resolveCall(call, _, _) or
DataFlowPrivate::resolveClassCall(call, _) DataFlowPrivate::resolveClassCall(call, _)
} }

View File

@@ -14,7 +14,6 @@
import python import python
import semmle.python.dataflow.new.DataFlow import semmle.python.dataflow.new.DataFlow
import semmle.python.ApiGraphs import semmle.python.ApiGraphs
private import semmle.python.controlflow.internal.Cfg as Cfg
/* /*
* Jinja 2 Docs: * Jinja 2 Docs:
@@ -37,8 +36,8 @@ private API::Node jinja2EnvironmentOrTemplate() {
from API::CallNode call from API::CallNode call
where where
call = jinja2EnvironmentOrTemplate().getACall() and call = jinja2EnvironmentOrTemplate().getACall() and
not exists(call.asCfgNode().(Cfg::CallNode).getNode().getStarargs()) and not exists(call.asCfgNode().(CallNode).getNode().getStarargs()) and
not exists(call.asCfgNode().(Cfg::CallNode).getNode().getKwargs()) and not exists(call.asCfgNode().(CallNode).getNode().getKwargs()) and
( (
not exists(call.getArgByName("autoescape")) not exists(call.getArgByName("autoescape"))
or or

View File

@@ -5,7 +5,6 @@
private import python private import python
private import semmle.python.ApiGraphs private import semmle.python.ApiGraphs
private import semmle.python.controlflow.internal.Cfg as Cfg
import TlsLibraryModel import TlsLibraryModel
class PyOpenSslContextCreation extends ContextCreation, DataFlow::CallCfgNode { class PyOpenSslContextCreation extends ContextCreation, DataFlow::CallCfgNode {
@@ -38,10 +37,10 @@ class ConnectionCall extends ConnectionCreation, DataFlow::CallCfgNode {
// This cannot be used to unrestrict, // This cannot be used to unrestrict,
// see https://www.pyopenssl.org/en/stable/api/ssl.html#OpenSSL.SSL.Context.set_options // see https://www.pyopenssl.org/en/stable/api/ssl.html#OpenSSL.SSL.Context.set_options
class SetOptionsCall extends ProtocolRestriction, DataFlow::CallCfgNode { class SetOptionsCall extends ProtocolRestriction, DataFlow::CallCfgNode {
SetOptionsCall() { node.getFunction().(Cfg::AttrNode).getName() = "set_options" } SetOptionsCall() { node.getFunction().(AttrNode).getName() = "set_options" }
override DataFlow::CfgNode getContext() { override DataFlow::CfgNode getContext() {
result.getNode() = node.getFunction().(Cfg::AttrNode).getObject() result.getNode() = node.getFunction().(AttrNode).getObject()
} }
override ProtocolVersion getRestriction() { override ProtocolVersion getRestriction() {

View File

@@ -5,7 +5,6 @@
private import python private import python
private import semmle.python.ApiGraphs private import semmle.python.ApiGraphs
private import semmle.python.controlflow.internal.Cfg as Cfg
import TlsLibraryModel import TlsLibraryModel
class SslContextCreation extends ContextCreation, DataFlow::CallCfgNode { class SslContextCreation extends ContextCreation, DataFlow::CallCfgNode {
@@ -54,7 +53,7 @@ class OptionsAugOr extends ProtocolRestriction, DataFlow::CfgNode {
ProtocolVersion restriction; ProtocolVersion restriction;
OptionsAugOr() { OptionsAugOr() {
exists(AugAssign aa, Cfg::AttrNode attr, Expr flag | exists(AugAssign aa, AttrNode attr, Expr flag |
aa.getOperation().getOp() instanceof BitOr and aa.getOperation().getOp() instanceof BitOr and
aa.getTarget() = attr.getNode() and aa.getTarget() = attr.getNode() and
attr.getName() = "options" and attr.getName() = "options" and
@@ -81,7 +80,7 @@ class OptionsAugAndNot extends ProtocolUnrestriction, DataFlow::CfgNode {
ProtocolVersion restriction; ProtocolVersion restriction;
OptionsAugAndNot() { OptionsAugAndNot() {
exists(AugAssign aa, Cfg::AttrNode attr, Expr flag, UnaryExpr notFlag | exists(AugAssign aa, AttrNode attr, Expr flag, UnaryExpr notFlag |
aa.getOperation().getOp() instanceof BitAnd and aa.getOperation().getOp() instanceof BitAnd and
aa.getTarget() = attr.getNode() and aa.getTarget() = attr.getNode() and
attr.getName() = "options" and attr.getName() = "options" and

View File

@@ -19,7 +19,6 @@ import semmle.python.filters.Tests
private import semmle.python.dataflow.new.internal.DataFlowDispatch as DataFlowDispatch private import semmle.python.dataflow.new.internal.DataFlowDispatch as DataFlowDispatch
private import semmle.python.dataflow.new.internal.Builtins::Builtins as Builtins private import semmle.python.dataflow.new.internal.Builtins::Builtins as Builtins
private import semmle.python.frameworks.data.ModelsAsData private import semmle.python.frameworks.data.ModelsAsData
private import semmle.python.controlflow.internal.Cfg as Cfg
bindingset[char, fraction] bindingset[char, fraction]
predicate fewer_characters_than(StringLiteral str, string char, float fraction) { predicate fewer_characters_than(StringLiteral str, string char, float fraction) {
@@ -49,7 +48,7 @@ predicate capitalized_word(StringLiteral str) { str.getText().regexpMatch("[A-Z]
predicate format_string(StringLiteral str) { str.getText().matches("%{%}%") } predicate format_string(StringLiteral str) { str.getText().matches("%{%}%") }
predicate maybeCredential(Cfg::ControlFlowNode f) { predicate maybeCredential(ControlFlowNode f) {
/* A string that is not too short and unlikely to be text or an identifier. */ /* A string that is not too short and unlikely to be text or an identifier. */
exists(StringLiteral str | str = f.getNode() | exists(StringLiteral str | str = f.getNode() |
/* At least 10 characters */ /* At least 10 characters */
@@ -97,7 +96,7 @@ class CredentialSink extends DataFlow::Node {
or or
exists(Keyword k | k.getArg() = name and this.asCfgNode().getNode() = k.getValue()) exists(Keyword k | k.getArg() = name and this.asCfgNode().getNode() = k.getValue())
or or
exists(Cfg::CompareNode cmp, Cfg::NameNode n | n.getId() = name | exists(CompareNode cmp, NameNode n | n.getId() = name |
cmp.operands(this.asCfgNode(), any(Eq eq), n) cmp.operands(this.asCfgNode(), any(Eq eq), n)
or or
cmp.operands(n, any(Eq eq), this.asCfgNode()) cmp.operands(n, any(Eq eq), this.asCfgNode())

View File

@@ -14,9 +14,8 @@
import python import python
private import semmle.python.ApiGraphs private import semmle.python.ApiGraphs
private import semmle.python.dataflow.new.DataFlow private import semmle.python.dataflow.new.DataFlow
private import semmle.python.controlflow.internal.Cfg as Cfg
predicate originIsLocals(Cfg::ControlFlowNode n) { predicate originIsLocals(ControlFlowNode n) {
// Only consider the `locals()` dictionary within the scope that called `locals()`. // Only consider the `locals()` dictionary within the scope that called `locals()`.
// Once the dictionary is passed to another scope (e.g. as an argument or via an // Once the dictionary is passed to another scope (e.g. as an argument or via an
// instance attribute) it is just an ordinary mapping, and modifying it is both // instance attribute) it is just an ordinary mapping, and modifying it is both
@@ -29,19 +28,19 @@ predicate originIsLocals(Cfg::ControlFlowNode n) {
) )
} }
predicate modification_of_locals(Cfg::ControlFlowNode f) { predicate modification_of_locals(ControlFlowNode f) {
originIsLocals(f.(Cfg::SubscriptNode).getObject()) and originIsLocals(f.(SubscriptNode).getObject()) and
(f.isStore() or f.isDelete()) (f.isStore() or f.isDelete())
or or
exists(string mname, Cfg::AttrNode attr | exists(string mname, AttrNode attr |
attr = f.(Cfg::CallNode).getFunction() and attr = f.(CallNode).getFunction() and
originIsLocals(attr.getObject(mname)) originIsLocals(attr.getObject(mname))
| |
mname in ["pop", "popitem", "update", "clear"] mname in ["pop", "popitem", "update", "clear"]
) )
} }
from AstNode a, Cfg::ControlFlowNode f from AstNode a, ControlFlowNode f
where where
modification_of_locals(f) and modification_of_locals(f) and
a = f.getNode() and a = f.getNode() and

View File

@@ -14,7 +14,6 @@
import python import python
private import semmle.python.ApiGraphs private import semmle.python.ApiGraphs
private import semmle.python.controlflow.internal.Cfg as Cfg
predicate func_with_side_effects(Expr e) { predicate func_with_side_effects(Expr e) {
exists(string name | name = e.(Attribute).getName() or name = e.(Name).getId() | exists(string name | name = e.(Attribute).getName() or name = e.(Name).getId() |
@@ -25,7 +24,7 @@ predicate func_with_side_effects(Expr e) {
} }
predicate call_with_side_effect(Call e) { predicate call_with_side_effect(Call e) {
exists(Cfg::ControlFlowNode eCfg | eCfg.getNode() = e | exists(ControlFlowNode eCfg | eCfg.getNode() = e |
eCfg = eCfg =
API::moduleImport("subprocess") API::moduleImport("subprocess")
.getMember(["call", "check_call", "check_output"]) .getMember(["call", "check_call", "check_output"])

View File

@@ -13,9 +13,8 @@
import python import python
private import semmle.python.ApiGraphs private import semmle.python.ApiGraphs
private import semmle.python.controlflow.internal.Cfg as Cfg
from Cfg::CallNode call, string name from CallNode call, string name
where where
name = ["exit", "quit"] and name = ["exit", "quit"] and
call = API::builtin(name).getACall().asCfgNode() call = API::builtin(name).getACall().asCfgNode()

View File

@@ -21,7 +21,6 @@ import semmle.python.ApiGraphs
import semmle.python.dataflow.new.internal.Attributes import semmle.python.dataflow.new.internal.Attributes
import semmle.python.dataflow.new.BarrierGuards import semmle.python.dataflow.new.BarrierGuards
import semmle.python.dataflow.new.RemoteFlowSources import semmle.python.dataflow.new.RemoteFlowSources
private import semmle.python.controlflow.internal.Cfg as Cfg
/** /**
* Handle those three cases of Tarfile opens: * Handle those three cases of Tarfile opens:
@@ -76,8 +75,8 @@ private module TarSlipImprovConfig implements DataFlow::ConfigSig {
call = atfo.getReturn().getMember("extractall").getACall() and call = atfo.getReturn().getMember("extractall").getACall() and
arg = call.getArgByName("members") and arg = call.getArgByName("members") and
if if
arg.asCfgNode() instanceof Cfg::NameConstantNode or arg.asCfgNode() instanceof NameConstantNode or
arg.asCfgNode() instanceof Cfg::ListNode arg.asCfgNode() instanceof ListNode
then sink = call.getObject() then sink = call.getObject()
else else
if arg.(MethodCallNode).getMethodName() = "getmembers" if arg.(MethodCallNode).getMethodName() = "getmembers"

View File

@@ -16,7 +16,6 @@ import python
import semmle.python.dataflow.new.DataFlow import semmle.python.dataflow.new.DataFlow
import semmle.python.ApiGraphs import semmle.python.ApiGraphs
import semmle.python.dataflow.new.TaintTracking import semmle.python.dataflow.new.TaintTracking
private import semmle.python.controlflow.internal.Cfg as Cfg
class PredictableResultSource extends DataFlow::Node { class PredictableResultSource extends DataFlow::Node {
PredictableResultSource() { PredictableResultSource() {
@@ -33,9 +32,7 @@ class PredictableResultSource extends DataFlow::Node {
class TokenAssignmentValueSink extends DataFlow::Node { class TokenAssignmentValueSink extends DataFlow::Node {
TokenAssignmentValueSink() { TokenAssignmentValueSink() {
exists(string name | name.toLowerCase().matches(["%token", "%code"]) | exists(string name | name.toLowerCase().matches(["%token", "%code"]) |
exists(Cfg::DefinitionNode n | n.getValue() = this.asCfgNode() | exists(DefinitionNode n | n.getValue() = this.asCfgNode() | name = n.(NameNode).getId())
name = n.(Cfg::NameNode).getId()
)
or or
exists(DataFlow::AttrWrite aw | aw.getValue() = this | name = aw.getAttributeName()) exists(DataFlow::AttrWrite aw | aw.getValue() = this | name = aw.getAttributeName())
) )

View File

@@ -11,25 +11,25 @@
import python import python
import semmle.python.ApiGraphs import semmle.python.ApiGraphs
import semmle.python.dataflow.new.TaintTracking import semmle.python.dataflow.new.TaintTracking
import semmle.python.Flow
import semmle.python.dataflow.new.RemoteFlowSources import semmle.python.dataflow.new.RemoteFlowSources
private import semmle.python.controlflow.internal.Cfg as Cfg
/** /**
* Returns true if the control flow node may be useful in the current context. * Returns true if the control flow node may be useful in the current context.
* *
* Ideally for more completeness, we should alert on every `startswith` call and every remote flow source which gets partailly checked. But, as this can lead to lots of FPs, we apply heuristics to filter some calls. This predicate provides logic for this filteration. * Ideally for more completeness, we should alert on every `startswith` call and every remote flow source which gets partailly checked. But, as this can lead to lots of FPs, we apply heuristics to filter some calls. This predicate provides logic for this filteration.
*/ */
private predicate maybeInteresting(Cfg::ControlFlowNode c) { private predicate maybeInteresting(ControlFlowNode c) {
// Check if the name of the variable which calls the function matches the heuristic. // Check if the name of the variable which calls the function matches the heuristic.
// This would typically occur at the sink. // This would typically occur at the sink.
// This should deal with cases like // This should deal with cases like
// `origin.startswith("bla")` // `origin.startswith("bla")`
heuristics(c.(Cfg::CallNode).getFunction().(Cfg::AttrNode).getObject().(Cfg::NameNode).getId()) heuristics(c.(CallNode).getFunction().(AttrNode).getObject().(NameNode).getId())
or or
// Check if the name of the variable passed as an argument to the functions matches the heuristic. This would typically occur at the sink. // Check if the name of the variable passed as an argument to the functions matches the heuristic. This would typically occur at the sink.
// This should deal with cases like // This should deal with cases like
// `bla.startswith(origin)` // `bla.startswith(origin)`
heuristics(c.(Cfg::CallNode).getArg(0).(Cfg::NameNode).getId()) heuristics(c.(CallNode).getArg(0).(NameNode).getId())
or or
// Check if the value gets written to any interesting variable. This would typically occur at the source. // Check if the value gets written to any interesting variable. This would typically occur at the source.
// This should deal with cases like // This should deal with cases like
@@ -37,10 +37,8 @@ private predicate maybeInteresting(Cfg::ControlFlowNode c) {
exists(Variable v | heuristics(v.getId()) | c.getASuccessor*().getNode() = v.getAStore()) exists(Variable v | heuristics(v.getId()) | c.getASuccessor*().getNode() = v.getAStore())
} }
private class StringStartswithCall extends Cfg::ControlFlowNode { private class StringStartswithCall extends ControlFlowNode {
StringStartswithCall() { StringStartswithCall() { this.(CallNode).getFunction().(AttrNode).getName() = "startswith" }
this.(Cfg::CallNode).getFunction().(Cfg::AttrNode).getName() = "startswith"
}
} }
bindingset[s] bindingset[s]
@@ -68,8 +66,8 @@ module CorsBypassConfig implements DataFlow::ConfigSig {
predicate isSink(DataFlow::Node node) { predicate isSink(DataFlow::Node node) {
exists(StringStartswithCall s | exists(StringStartswithCall s |
node.asCfgNode() = s.(Cfg::CallNode).getArg(0) or node.asCfgNode() = s.(CallNode).getArg(0) or
node.asCfgNode() = s.(Cfg::CallNode).getFunction().(Cfg::AttrNode).getObject() node.asCfgNode() = s.(CallNode).getFunction().(AttrNode).getObject()
) )
} }

View File

@@ -16,7 +16,6 @@ import semmle.python.Concepts
import semmle.python.dataflow.new.TaintTracking import semmle.python.dataflow.new.TaintTracking
import semmle.python.dataflow.new.internal.DataFlowPublic import semmle.python.dataflow.new.internal.DataFlowPublic
import semmle.python.dataflow.new.RemoteFlowSources import semmle.python.dataflow.new.RemoteFlowSources
private import semmle.python.controlflow.internal.Cfg as Cfg
// The Unicode compatibility normalization calls from unicodedata, unidecode, pyunormalize // The Unicode compatibility normalization calls from unicodedata, unidecode, pyunormalize
// and textnorm modules. The use of argIdx is to constraint the argument being normalized. // and textnorm modules. The use of argIdx is to constraint the argument being normalized.
@@ -53,8 +52,8 @@ class UnicodeCompatibilityNormalize extends API::CallNode {
DataFlow::Node getPathArg() { result = this.getArg(argIdx) } DataFlow::Node getPathArg() { result = this.getArg(argIdx) }
} }
predicate underAValue(DataFlow::GuardNode g, Cfg::ControlFlowNode node, boolean branch) { predicate underAValue(DataFlow::GuardNode g, ControlFlowNode node, boolean branch) {
exists(Cfg::CompareNode cn | cn = g | exists(CompareNode cn | cn = g |
exists(API::CallNode lenCall, Cmpop op, Node n | exists(API::CallNode lenCall, Cmpop op, Node n |
lenCall = n.getALocalSource() and lenCall = n.getALocalSource() and
( (

View File

@@ -9,7 +9,6 @@ import semmle.python.ApiGraphs
import semmle.python.dataflow.new.TaintTracking import semmle.python.dataflow.new.TaintTracking
import semmle.python.frameworks.Stdlib import semmle.python.frameworks.Stdlib
import semmle.python.dataflow.new.RemoteFlowSources import semmle.python.dataflow.new.RemoteFlowSources
private import semmle.python.controlflow.internal.Cfg as Cfg
/** /**
* Handle those three cases of Tarfile opens: * Handle those three cases of Tarfile opens:
@@ -112,8 +111,8 @@ module UnsafeUnpackConfig implements DataFlow::ConfigSig {
call = atfo.getReturn().getMember("extractall").getACall() and call = atfo.getReturn().getMember("extractall").getACall() and
arg = call.getArgByName("members") and arg = call.getArgByName("members") and
if if
arg.asCfgNode() instanceof Cfg::NameConstantNode or arg.asCfgNode() instanceof NameConstantNode or
arg.asCfgNode() instanceof Cfg::ListNode arg.asCfgNode() instanceof ListNode
then sink = call.getObject() then sink = call.getObject()
else else
if arg.(MethodCallNode).getMethodName() = "getmembers" if arg.(MethodCallNode).getMethodName() = "getmembers"

View File

@@ -4,7 +4,6 @@ import semmle.python.dataflow.new.DataFlow
import semmle.python.dataflow.new.TaintTracking import semmle.python.dataflow.new.TaintTracking
import semmle.python.dataflow.new.BarrierGuards import semmle.python.dataflow.new.BarrierGuards
import semmle.python.dataflow.new.RemoteFlowSources import semmle.python.dataflow.new.RemoteFlowSources
private import semmle.python.controlflow.internal.Cfg as Cfg
/** /**
* A taint-tracking configuration for tracking untrusted user input used in file read. * A taint-tracking configuration for tracking untrusted user input used in file read.
@@ -22,7 +21,7 @@ private module CsvInjectionConfig implements DataFlow::ConfigSig {
predicate observeDiffInformedIncrementalMode() { any() } predicate observeDiffInformedIncrementalMode() { any() }
} }
private predicate startsWithCheck(DataFlow::GuardNode g, Cfg::ControlFlowNode node, boolean branch) { private predicate startsWithCheck(DataFlow::GuardNode g, ControlFlowNode node, boolean branch) {
exists(DataFlow::MethodCallNode mc | exists(DataFlow::MethodCallNode mc |
g = mc.asCfgNode() and g = mc.asCfgNode() and
mc.calls(_, "startswith") and mc.calls(_, "startswith") and

View File

@@ -6,7 +6,6 @@
import python import python
import meta.MetaMetrics import meta.MetaMetrics
private import LegacyPointsTo private import LegacyPointsTo
private import semmle.python.controlflow.internal.Cfg as Cfg
newtype TTarget = newtype TTarget =
TFunction(Function func) or TFunction(Function func) or
@@ -51,7 +50,7 @@ class TargetClass extends Target, TClass {
* A call that is (possibly) relevant for analysis quality. * A call that is (possibly) relevant for analysis quality.
* See `IgnoredFile` for details on what is excluded. * See `IgnoredFile` for details on what is excluded.
*/ */
class RelevantCall extends Cfg::CallNode { class RelevantCall extends CallNode {
RelevantCall() { not this.getLocation().getFile() instanceof IgnoredFile } RelevantCall() { not this.getLocation().getFile() instanceof IgnoredFile }
} }
@@ -61,7 +60,7 @@ module PointsToBasedCallGraph {
class ResolvableCall extends RelevantCall { class ResolvableCall extends RelevantCall {
Value targetValue; Value targetValue;
ResolvableCall() { targetValue.getACall().getNode() = this.getNode() } ResolvableCall() { targetValue.getACall() = this }
/** Gets a resolved target of this call. */ /** Gets a resolved target of this call. */
Target getTarget() { Target getTarget() {

View File

@@ -8,9 +8,8 @@
import python import python
import CallGraphQuality import CallGraphQuality
private import semmle.python.controlflow.internal.Cfg as Cfg
from Cfg::CallNode call, Target target from CallNode call, Target target
where where
target.isRelevant() and target.isRelevant() and
call.(TypeTrackingBasedCallGraph::ResolvableCall).getTarget() = target call.(TypeTrackingBasedCallGraph::ResolvableCall).getTarget() = target

View File

@@ -8,9 +8,8 @@
import python import python
import CallGraphQuality import CallGraphQuality
private import semmle.python.controlflow.internal.Cfg as Cfg
from Cfg::CallNode call, Target target from CallNode call, Target target
where where
target.isRelevant() and target.isRelevant() and
call.(PointsToBasedCallGraph::ResolvableCall).getTarget() = target and call.(PointsToBasedCallGraph::ResolvableCall).getTarget() = target and

View File

@@ -8,9 +8,8 @@
import python import python
import CallGraphQuality import CallGraphQuality
private import semmle.python.controlflow.internal.Cfg as Cfg
from Cfg::CallNode call, Target target from CallNode call, Target target
where where
target.isRelevant() and target.isRelevant() and
not call.(PointsToBasedCallGraph::ResolvableCall).getTarget() = target and not call.(PointsToBasedCallGraph::ResolvableCall).getTarget() = target and

View File

@@ -8,9 +8,8 @@
import python import python
import CallGraphQuality import CallGraphQuality
private import semmle.python.controlflow.internal.Cfg as Cfg
from Cfg::CallNode call, Target target from CallNode call, Target target
where where
target.isRelevant() and target.isRelevant() and
not call.(PointsToBasedCallGraph::ResolvableCall).getTarget() = target and not call.(PointsToBasedCallGraph::ResolvableCall).getTarget() = target and

View File

@@ -6,13 +6,12 @@
import python import python
import CallGraphQuality import CallGraphQuality
private import semmle.python.controlflow.internal.Cfg as Cfg
from string tag, int c from string tag, int c
where where
tag = "SHARED" and tag = "SHARED" and
c = c =
count(Cfg::CallNode call, Target target | count(CallNode call, Target target |
target.isRelevant() and target.isRelevant() and
call.(PointsToBasedCallGraph::ResolvableCall).getTarget() = target and call.(PointsToBasedCallGraph::ResolvableCall).getTarget() = target and
call.(TypeTrackingBasedCallGraph::ResolvableCall).getTarget() = target call.(TypeTrackingBasedCallGraph::ResolvableCall).getTarget() = target
@@ -20,7 +19,7 @@ where
or or
tag = "NEW" and tag = "NEW" and
c = c =
count(Cfg::CallNode call, Target target | count(CallNode call, Target target |
target.isRelevant() and target.isRelevant() and
not call.(PointsToBasedCallGraph::ResolvableCall).getTarget() = target and not call.(PointsToBasedCallGraph::ResolvableCall).getTarget() = target and
call.(TypeTrackingBasedCallGraph::ResolvableCall).getTarget() = target call.(TypeTrackingBasedCallGraph::ResolvableCall).getTarget() = target
@@ -28,7 +27,7 @@ where
or or
tag = "MISSING" and tag = "MISSING" and
c = c =
count(Cfg::CallNode call, Target target | count(CallNode call, Target target |
target.isRelevant() and target.isRelevant() and
call.(PointsToBasedCallGraph::ResolvableCall).getTarget() = target and call.(PointsToBasedCallGraph::ResolvableCall).getTarget() = target and
not call.(TypeTrackingBasedCallGraph::ResolvableCall).getTarget() = target not call.(TypeTrackingBasedCallGraph::ResolvableCall).getTarget() = target

View File

@@ -8,9 +8,8 @@
import python import python
import CallGraphQuality import CallGraphQuality
private import semmle.python.controlflow.internal.Cfg as Cfg
from Cfg::CallNode call, Target target from CallNode call, Target target
where where
target.isRelevant() and target.isRelevant() and
call.(PointsToBasedCallGraph::ResolvableCall).getTarget() = target and call.(PointsToBasedCallGraph::ResolvableCall).getTarget() = target and

View File

@@ -6,7 +6,6 @@
private import python private import python
private import semmle.python.dataflow.new.DataFlow private import semmle.python.dataflow.new.DataFlow
private import semmle.python.dataflow.new.BarrierGuards private import semmle.python.dataflow.new.BarrierGuards
private import semmle.python.controlflow.internal.Cfg as Cfg
/** /**
* Provides default sources, sinks and sanitizers for detecting * Provides default sources, sinks and sanitizers for detecting
@@ -77,7 +76,7 @@ module ModificationOfParameterWithDefault {
boolean nonEmpty; boolean nonEmpty;
MutableDefaultValue() { MutableDefaultValue() {
nonEmpty = mutableDefaultValue(this.asCfgNode().(Cfg::NameNode).getNode()) and nonEmpty = mutableDefaultValue(this.asCfgNode().(NameNode).getNode()) and
// Ignore sources inside the standard library. These are unlikely to be true positives. // Ignore sources inside the standard library. These are unlikely to be true positives.
exists(this.getLocation().getFile().getRelativePath()) exists(this.getLocation().getFile().getRelativePath())
} }
@@ -126,10 +125,10 @@ module ModificationOfParameterWithDefault {
class Mutation extends Sink { class Mutation extends Sink {
Mutation() { Mutation() {
// assignment to a subscript (includes slices) // assignment to a subscript (includes slices)
exists(Cfg::DefinitionNode d | d.(Cfg::SubscriptNode).getObject() = this.asCfgNode()) exists(DefinitionNode d | d.(SubscriptNode).getObject() = this.asCfgNode())
or or
// deletion of a subscript // deletion of a subscript
exists(Cfg::DeletionNode d | d.(Cfg::SubscriptNode).getObject() = this.asCfgNode()) exists(DeletionNode d | d.getTarget().(SubscriptNode).getObject() = this.asCfgNode())
or or
// augmented assignment to the value // augmented assignment to the value
exists(AugAssign a | this.asCfgNode().getNode() = a.getTarget()) exists(AugAssign a | this.asCfgNode().getNode() = a.getTarget())
@@ -142,26 +141,54 @@ module ModificationOfParameterWithDefault {
} }
} }
/** // This to reimplement some of the functionality of the DataFlow::BarrierGuard
* Holds if `g` is a direct truthiness check that proves its (single private import semmle.python.essa.SsaCompute
* operand-and-result) value truthy on the `true` branch — i.e. `if g:`.
*/
private predicate truthyGuard(DataFlow::GuardNode g, Cfg::ControlFlowNode node, boolean branch) {
node = g and branch = true
}
/** Holds if `g` is a direct falsiness check — i.e. `if g:` taken to the false branch. */ /**
private predicate falseyGuard(DataFlow::GuardNode g, Cfg::ControlFlowNode node, boolean branch) { * A data-flow node that is known to be either truthy or falsey.
node = g and branch = false *
* It handles the cases `if x` and `if not x`.
*
* For example, in the following code, `this` will be the `x` that is printed,
* which we will know is truthy:
*
* ```py
* if x:
* print(x)
* ```
*/
private class MustBe extends DataFlow::Node {
boolean truthy;
MustBe() {
exists(DataFlow::GuardNode guard, NameNode guarded, boolean branch |
// case: if x
guard = guarded and
branch = truthy
or
// case: if not x
guard.(UnaryExprNode).getNode().getOp() instanceof Not and
guarded = guard.(UnaryExprNode).getOperand() and
branch = truthy.booleanNot()
|
// guard controls this
guard.controlsBlock(this.asCfgNode().getBasicBlock(), branch) and
// there is a definition tying the guarded value to this
exists(EssaDefinition def |
AdjacentUses::useOfDef(def, this.asCfgNode()) and
AdjacentUses::useOfDef(def, guarded)
)
)
}
} }
/** Simple guard detecting truthy values. */ /** Simple guard detecting truthy values. */
private class MustBeTruthy extends MustBeNonEmpty { private class MustBeTruthy extends MustBe, MustBeNonEmpty {
MustBeTruthy() { this = DataFlow::BarrierGuard<truthyGuard/3>::getABarrierNode() } MustBeTruthy() { truthy = true }
} }
/** Simple guard detecting falsey values. */ /** Simple guard detecting falsey values. */
private class MustBeFalsey extends MustBeEmpty { private class MustBeFalsey extends MustBe, MustBeEmpty {
MustBeFalsey() { this = DataFlow::BarrierGuard<falseyGuard/3>::getABarrierNode() } MustBeFalsey() { truthy = false }
} }
} }

View File

@@ -1,5 +1,2 @@
| test.py:5:15:5:22 | ControlFlowNode for next() | Call to 'next()' in a generator. | | test.py:5:15:5:22 | ControlFlowNode for next() | Call to 'next()' in a generator. |
| test.py:10:20:10:27 | ControlFlowNode for next() | Call to 'next()' in a generator. | | test.py:10:20:10:27 | ControlFlowNode for next() | Call to 'next()' in a generator. |
| test.py:34:15:34:24 | ControlFlowNode for next() | Call to 'next()' in a generator. |
| test.py:41:19:41:28 | ControlFlowNode for next() | Call to 'next()' in a generator. |
| test.py:62:19:62:28 | ControlFlowNode for next() | Call to 'next()' in a generator. |

View File

@@ -1,3 +1,3 @@
| test.py:8:5:8:12 | Raise | Raising a $@ will result in the first element (recursively) being raised and all other elements being discarded. | test.py:7:10:7:29 | After Tuple | tuple | | test.py:8:5:8:12 | Raise | Raising a $@ will result in the first element (recursively) being raised and all other elements being discarded. | test.py:7:10:7:29 | ControlFlowNode for Tuple | tuple |
| test.py:11:5:11:32 | Raise | Raising a $@ will result in the first element (recursively) being raised and all other elements being discarded. | test.py:11:12:11:31 | After Tuple | tuple | | test.py:11:5:11:32 | Raise | Raising a $@ will result in the first element (recursively) being raised and all other elements being discarded. | test.py:11:12:11:31 | ControlFlowNode for Tuple | tuple |
| test.py:15:5:15:23 | Raise | Raising a $@ will result in the first element (recursively) being raised and all other elements being discarded. | test.py:14:10:14:19 | After Tuple | tuple | | test.py:15:5:15:23 | Raise | Raising a $@ will result in the first element (recursively) being raised and all other elements being discarded. | test.py:14:10:14:19 | ControlFlowNode for Tuple | tuple |

View File

@@ -1,2 +1,2 @@
| UseofApply.py:19:3:19:17 | After apply() | Call to the obsolete builtin function 'apply'. | | UseofApply.py:19:3:19:17 | ControlFlowNode for apply() | Call to the obsolete builtin function 'apply'. |
| expressions_test.py:2:5:2:21 | After apply() | Call to the obsolete builtin function 'apply'. | | expressions_test.py:2:5:2:21 | ControlFlowNode for apply() | Call to the obsolete builtin function 'apply'. |

View File

@@ -1 +1 @@
| expressions_test.py:6:12:6:18 | After input() | The unsafe built-in function 'input' is used in Python 2. | | expressions_test.py:6:12:6:18 | ControlFlowNode for input() | The unsafe built-in function 'input' is used in Python 2. |

View File

@@ -1,2 +1,2 @@
| test.py:10:1:10:9 | After Attribute | test.py:10:1:10:5 | myobj | foo | | test.py:10:1:10:9 | ControlFlowNode for Attribute | test.py:10:1:10:5 | ControlFlowNode for myobj | foo |
| test.py:13:1:13:21 | After getattr() | test.py:13:9:13:13 | myobj | foo | | test.py:13:1:13:21 | ControlFlowNode for getattr() | test.py:13:9:13:13 | ControlFlowNode for myobj | foo |

View File

@@ -1,5 +1,5 @@
| test.py:5:9:5:16 | __init__ | test.py:4:1:4:20 | After ClassExpr | __init__ | test.py:5:5:5:28 | FunctionExpr | | test.py:5:9:5:16 | ControlFlowNode for __init__ | test.py:4:1:4:20 | ControlFlowNode for ClassExpr | __init__ | test.py:5:5:5:28 | ControlFlowNode for FunctionExpr |
| test.py:6:9:6:16 | After Attribute | test.py:6:9:6:12 | self | foo | test.py:6:20:6:22 | foo | | test.py:6:9:6:16 | ControlFlowNode for Attribute | test.py:6:9:6:12 | ControlFlowNode for self | foo | test.py:6:20:6:22 | ControlFlowNode for foo |
| test.py:9:1:9:9 | After Attribute | test.py:0:0:0:0 | ModuleVariableNode in Module test for myobj | foo | test.py:9:13:9:17 | StringLiteral | | test.py:9:1:9:9 | ControlFlowNode for Attribute | test.py:0:0:0:0 | ModuleVariableNode in Module test for myobj | foo | test.py:9:13:9:17 | ControlFlowNode for StringLiteral |
| test.py:9:1:9:9 | After Attribute | test.py:9:1:9:5 | myobj | foo | test.py:9:13:9:17 | StringLiteral | | test.py:9:1:9:9 | ControlFlowNode for Attribute | test.py:9:1:9:5 | ControlFlowNode for myobj | foo | test.py:9:13:9:17 | ControlFlowNode for StringLiteral |
| test.py:12:1:12:25 | After setattr() | test.py:12:9:12:13 | myobj | foo | test.py:12:23:12:24 | IntegerLiteral | | test.py:12:1:12:25 | ControlFlowNode for setattr() | test.py:12:9:12:13 | ControlFlowNode for myobj | foo | test.py:12:23:12:24 | ControlFlowNode for IntegerLiteral |

View File

@@ -2,16 +2,15 @@ import python
import semmle.python.dataflow.new.DataFlow import semmle.python.dataflow.new.DataFlow
import semmle.python.dataflow.new.TaintTracking import semmle.python.dataflow.new.TaintTracking
import utils.test.InlineExpectationsTest import utils.test.InlineExpectationsTest
private import semmle.python.controlflow.internal.Cfg as Cfg
private module TestConfig implements DataFlow::ConfigSig { private module TestConfig implements DataFlow::ConfigSig {
predicate isSource(DataFlow::Node node) { predicate isSource(DataFlow::Node node) {
node.(DataFlow::CallCfgNode).getFunction().asCfgNode().(Cfg::NameNode).getId() = "source" node.(DataFlow::CallCfgNode).getFunction().asCfgNode().(NameNode).getId() = "source"
} }
predicate isSink(DataFlow::Node node) { predicate isSink(DataFlow::Node node) {
exists(DataFlow::CallCfgNode call | exists(DataFlow::CallCfgNode call |
call.getFunction().asCfgNode().(Cfg::NameNode).getId() = "sink" and call.getFunction().asCfgNode().(NameNode).getId() = "sink" and
node = call.getArg(0) node = call.getArg(0)
) )
} }

View File

@@ -1,120 +1,120 @@
| attr_clash.__init__ | __file__ | attr_clash/__init__.py:6:6:6:13 | __file__ | | attr_clash.__init__ | __file__ | attr_clash/__init__.py:6:6:6:13 | ControlFlowNode for __file__ |
| attr_clash.__init__ | clashing_attr | attr_clash/__init__.py:4:1:4:13 | clashing_attr | | attr_clash.__init__ | clashing_attr | attr_clash/__init__.py:4:1:4:13 | ControlFlowNode for clashing_attr |
| attr_clash.__init__ | enter | attr_clash/__init__.py:2:1:2:5 | enter | | attr_clash.__init__ | enter | attr_clash/__init__.py:2:1:2:5 | ControlFlowNode for enter |
| attr_clash.__init__ | exit | attr_clash/__init__.py:6:1:6:4 | exit | | attr_clash.__init__ | exit | attr_clash/__init__.py:6:1:6:4 | ControlFlowNode for exit |
| attr_clash.clashing_attr | __file__ | attr_clash/clashing_attr.py:4:6:4:13 | __file__ | | attr_clash.clashing_attr | __file__ | attr_clash/clashing_attr.py:4:6:4:13 | ControlFlowNode for __file__ |
| attr_clash.clashing_attr | enter | attr_clash/clashing_attr.py:2:1:2:5 | enter | | attr_clash.clashing_attr | enter | attr_clash/clashing_attr.py:2:1:2:5 | ControlFlowNode for enter |
| attr_clash.clashing_attr | exit | attr_clash/clashing_attr.py:4:1:4:4 | exit | | attr_clash.clashing_attr | exit | attr_clash/clashing_attr.py:4:1:4:4 | ControlFlowNode for exit |
| attr_clash.non_clashing_submodule | __file__ | attr_clash/non_clashing_submodule.py:4:6:4:13 | __file__ | | attr_clash.non_clashing_submodule | __file__ | attr_clash/non_clashing_submodule.py:4:6:4:13 | ControlFlowNode for __file__ |
| attr_clash.non_clashing_submodule | enter | attr_clash/non_clashing_submodule.py:2:1:2:5 | enter | | attr_clash.non_clashing_submodule | enter | attr_clash/non_clashing_submodule.py:2:1:2:5 | ControlFlowNode for enter |
| attr_clash.non_clashing_submodule | exit | attr_clash/non_clashing_submodule.py:4:1:4:4 | exit | | attr_clash.non_clashing_submodule | exit | attr_clash/non_clashing_submodule.py:4:1:4:4 | ControlFlowNode for exit |
| bar | __file__ | bar.py:6:6:6:13 | __file__ | | bar | __file__ | bar.py:6:6:6:13 | ControlFlowNode for __file__ |
| bar | bar_attr | bar.py:4:1:4:8 | bar_attr | | bar | bar_attr | bar.py:4:1:4:8 | ControlFlowNode for bar_attr |
| bar | enter | bar.py:2:1:2:5 | enter | | bar | enter | bar.py:2:1:2:5 | ControlFlowNode for enter |
| bar | exit | bar.py:6:1:6:4 | exit | | bar | exit | bar.py:6:1:6:4 | ControlFlowNode for exit |
| baz | __file__ | baz.py:6:6:6:13 | __file__ | | baz | __file__ | baz.py:6:6:6:13 | ControlFlowNode for __file__ |
| baz | baz_attr | baz.py:4:1:4:8 | baz_attr | | baz | baz_attr | baz.py:4:1:4:8 | ControlFlowNode for baz_attr |
| baz | enter | baz.py:2:1:2:5 | enter | | baz | enter | baz.py:2:1:2:5 | ControlFlowNode for enter |
| baz | exit | baz.py:6:1:6:4 | exit | | baz | exit | baz.py:6:1:6:4 | ControlFlowNode for exit |
| block_flow_check | SOURCE | block_flow_check.py:12:25:12:30 | SOURCE | | block_flow_check | SOURCE | block_flow_check.py:12:25:12:30 | ControlFlowNode for SOURCE |
| block_flow_check | __file__ | block_flow_check.py:14:6:14:13 | __file__ | | block_flow_check | __file__ | block_flow_check.py:14:6:14:13 | ControlFlowNode for __file__ |
| block_flow_check | check | block_flow_check.py:12:1:12:5 | check | | block_flow_check | check | block_flow_check.py:12:1:12:5 | ControlFlowNode for check |
| block_flow_check | enter | block_flow_check.py:2:1:2:5 | enter | | block_flow_check | enter | block_flow_check.py:2:1:2:5 | ControlFlowNode for enter |
| block_flow_check | exit | block_flow_check.py:14:1:14:4 | exit | | block_flow_check | exit | block_flow_check.py:14:1:14:4 | ControlFlowNode for exit |
| block_flow_check | globals | block_flow_check.py:12:33:12:39 | globals | | block_flow_check | globals | block_flow_check.py:12:33:12:39 | ControlFlowNode for globals |
| block_flow_check | object | block_flow_check.py:4:14:4:19 | object | | block_flow_check | object | block_flow_check.py:4:14:4:19 | ControlFlowNode for object |
| foo | __file__ | foo.py:14:6:14:13 | __file__ | | foo | __file__ | foo.py:14:6:14:13 | ControlFlowNode for __file__ |
| foo | __private_foo_attr | foo.py:8:1:8:18 | __private_foo_attr | | foo | __private_foo_attr | foo.py:8:1:8:18 | ControlFlowNode for __private_foo_attr |
| foo | bar_reexported | foo.py:11:8:11:10 | ImportExpr | | foo | bar_reexported | foo.py:11:8:11:10 | ControlFlowNode for ImportExpr |
| foo | bar_reexported | foo.py:12:34:12:47 | bar_reexported | | foo | bar_reexported | foo.py:12:34:12:47 | ControlFlowNode for bar_reexported |
| foo | check | foo.py:12:1:12:5 | check | | foo | check | foo.py:12:1:12:5 | ControlFlowNode for check |
| foo | enter | foo.py:2:1:2:5 | enter | | foo | enter | foo.py:2:1:2:5 | ControlFlowNode for enter |
| foo | exit | foo.py:14:1:14:4 | exit | | foo | exit | foo.py:14:1:14:4 | ControlFlowNode for exit |
| foo | foo_attr | foo.py:5:1:5:8 | foo_attr | | foo | foo_attr | foo.py:5:1:5:8 | ControlFlowNode for foo_attr |
| foo | globals | foo.py:12:71:12:77 | globals | | foo | globals | foo.py:12:71:12:77 | ControlFlowNode for globals |
| generous_export | Exception | generous_export.py:16:11:16:19 | Exception | | generous_export | Exception | generous_export.py:16:11:16:19 | ControlFlowNode for Exception |
| generous_export | SOURCE | generous_export.py:15:11:15:16 | SOURCE | | generous_export | SOURCE | generous_export.py:15:11:15:16 | ControlFlowNode for SOURCE |
| generous_export | SOURCE | generous_export.py:20:25:20:30 | SOURCE | | generous_export | SOURCE | generous_export.py:20:25:20:30 | ControlFlowNode for SOURCE |
| generous_export | __file__ | generous_export.py:22:6:22:13 | __file__ | | generous_export | __file__ | generous_export.py:22:6:22:13 | ControlFlowNode for __file__ |
| generous_export | check | generous_export.py:20:1:20:5 | check | | generous_export | check | generous_export.py:20:1:20:5 | ControlFlowNode for check |
| generous_export | enter | generous_export.py:2:1:2:5 | enter | | generous_export | enter | generous_export.py:2:1:2:5 | ControlFlowNode for enter |
| generous_export | eval | generous_export.py:10:4:10:7 | eval | | generous_export | eval | generous_export.py:10:4:10:7 | ControlFlowNode for eval |
| generous_export | exit | generous_export.py:22:1:22:4 | exit | | generous_export | exit | generous_export.py:22:1:22:4 | ControlFlowNode for exit |
| generous_export | globals | generous_export.py:20:33:20:39 | globals | | generous_export | globals | generous_export.py:20:33:20:39 | ControlFlowNode for globals |
| generous_export | object | generous_export.py:4:14:4:19 | object | | generous_export | object | generous_export.py:4:14:4:19 | ControlFlowNode for object |
| generous_export | print | generous_export.py:15:5:15:9 | print | | generous_export | print | generous_export.py:15:5:15:9 | ControlFlowNode for print |
| has_defined_all | __all__ | has_defined_all.py:7:1:7:7 | __all__ | | has_defined_all | __all__ | has_defined_all.py:7:1:7:7 | ControlFlowNode for __all__ |
| has_defined_all | __file__ | has_defined_all.py:9:6:9:13 | __file__ | | has_defined_all | __file__ | has_defined_all.py:9:6:9:13 | ControlFlowNode for __file__ |
| has_defined_all | all_defined_bar | has_defined_all.py:5:1:5:15 | all_defined_bar | | has_defined_all | all_defined_bar | has_defined_all.py:5:1:5:15 | ControlFlowNode for all_defined_bar |
| has_defined_all | all_defined_foo | has_defined_all.py:4:1:4:15 | all_defined_foo | | has_defined_all | all_defined_foo | has_defined_all.py:4:1:4:15 | ControlFlowNode for all_defined_foo |
| has_defined_all | enter | has_defined_all.py:2:1:2:5 | enter | | has_defined_all | enter | has_defined_all.py:2:1:2:5 | ControlFlowNode for enter |
| has_defined_all | exit | has_defined_all.py:9:1:9:4 | exit | | has_defined_all | exit | has_defined_all.py:9:1:9:4 | ControlFlowNode for exit |
| has_defined_all_copy | __all__ | has_defined_all_copy.py:9:1:9:7 | __all__ | | has_defined_all_copy | __all__ | has_defined_all_copy.py:9:1:9:7 | ControlFlowNode for __all__ |
| has_defined_all_copy | __file__ | has_defined_all_copy.py:11:6:11:13 | __file__ | | has_defined_all_copy | __file__ | has_defined_all_copy.py:11:6:11:13 | ControlFlowNode for __file__ |
| has_defined_all_copy | all_defined_bar_copy | has_defined_all_copy.py:7:1:7:20 | all_defined_bar_copy | | has_defined_all_copy | all_defined_bar_copy | has_defined_all_copy.py:7:1:7:20 | ControlFlowNode for all_defined_bar_copy |
| has_defined_all_copy | all_defined_foo_copy | has_defined_all_copy.py:6:1:6:20 | all_defined_foo_copy | | has_defined_all_copy | all_defined_foo_copy | has_defined_all_copy.py:6:1:6:20 | ControlFlowNode for all_defined_foo_copy |
| has_defined_all_copy | enter | has_defined_all_copy.py:4:1:4:5 | enter | | has_defined_all_copy | enter | has_defined_all_copy.py:4:1:4:5 | ControlFlowNode for enter |
| has_defined_all_copy | exit | has_defined_all_copy.py:11:1:11:4 | exit | | has_defined_all_copy | exit | has_defined_all_copy.py:11:1:11:4 | ControlFlowNode for exit |
| has_defined_all_indirection | __file__ | has_defined_all_indirection.py:6:6:6:13 | __file__ | | has_defined_all_indirection | __file__ | has_defined_all_indirection.py:6:6:6:13 | ControlFlowNode for __file__ |
| has_defined_all_indirection | all_defined_foo_copy | has_defined_all_copy.py:6:1:6:20 | all_defined_foo_copy | | has_defined_all_indirection | all_defined_foo_copy | has_defined_all_copy.py:6:1:6:20 | ControlFlowNode for all_defined_foo_copy |
| has_defined_all_indirection | enter | has_defined_all_indirection.py:2:1:2:5 | enter | | has_defined_all_indirection | enter | has_defined_all_indirection.py:2:1:2:5 | ControlFlowNode for enter |
| has_defined_all_indirection | exit | has_defined_all_indirection.py:6:1:6:4 | exit | | has_defined_all_indirection | exit | has_defined_all_indirection.py:6:1:6:4 | ControlFlowNode for exit |
| if_then_else | __file__ | if_then_else.py:16:6:16:13 | __file__ | | if_then_else | __file__ | if_then_else.py:16:6:16:13 | ControlFlowNode for __file__ |
| if_then_else | enter | if_then_else.py:2:1:2:5 | enter | | if_then_else | enter | if_then_else.py:2:1:2:5 | ControlFlowNode for enter |
| if_then_else | eval | if_then_else.py:11:8:11:11 | eval | | if_then_else | eval | if_then_else.py:11:8:11:11 | ControlFlowNode for eval |
| if_then_else | exit | if_then_else.py:16:1:16:4 | exit | | if_then_else | exit | if_then_else.py:16:1:16:4 | ControlFlowNode for exit |
| if_then_else | if_then_else_defined | if_then_else.py:7:5:7:24 | if_then_else_defined | | if_then_else | if_then_else_defined | if_then_else.py:7:5:7:24 | ControlFlowNode for if_then_else_defined |
| if_then_else | if_then_else_defined | if_then_else.py:12:9:12:28 | if_then_else_defined | | if_then_else | if_then_else_defined | if_then_else.py:12:9:12:28 | ControlFlowNode for if_then_else_defined |
| if_then_else | if_then_else_defined | if_then_else.py:14:9:14:28 | if_then_else_defined | | if_then_else | if_then_else_defined | if_then_else.py:14:9:14:28 | ControlFlowNode for if_then_else_defined |
| if_then_else_refined | SOURCE | if_then_else_refined.py:11:11:11:16 | SOURCE | | if_then_else_refined | SOURCE | if_then_else_refined.py:11:11:11:16 | ControlFlowNode for SOURCE |
| if_then_else_refined | SOURCE | if_then_else_refined.py:13:11:13:16 | SOURCE | | if_then_else_refined | SOURCE | if_then_else_refined.py:13:11:13:16 | ControlFlowNode for SOURCE |
| if_then_else_refined | __file__ | if_then_else_refined.py:19:6:19:13 | __file__ | | if_then_else_refined | __file__ | if_then_else_refined.py:19:6:19:13 | ControlFlowNode for __file__ |
| if_then_else_refined | check | if_then_else_refined.py:17:1:17:5 | check | | if_then_else_refined | check | if_then_else_refined.py:17:1:17:5 | ControlFlowNode for check |
| if_then_else_refined | enter | if_then_else_refined.py:4:1:4:5 | enter | | if_then_else_refined | enter | if_then_else_refined.py:4:1:4:5 | ControlFlowNode for enter |
| if_then_else_refined | eval | if_then_else_refined.py:10:4:10:7 | eval | | if_then_else_refined | eval | if_then_else_refined.py:10:4:10:7 | ControlFlowNode for eval |
| if_then_else_refined | exit | if_then_else_refined.py:19:1:19:4 | exit | | if_then_else_refined | exit | if_then_else_refined.py:19:1:19:4 | ControlFlowNode for exit |
| if_then_else_refined | globals | if_then_else_refined.py:17:24:17:30 | globals | | if_then_else_refined | globals | if_then_else_refined.py:17:24:17:30 | ControlFlowNode for globals |
| if_then_else_refined | src | if_then_else_refined.py:17:19:17:21 | src | | if_then_else_refined | src | if_then_else_refined.py:17:19:17:21 | ControlFlowNode for src |
| package.__init__ | __file__ | package/__init__.py:7:6:7:13 | __file__ | | package.__init__ | __file__ | package/__init__.py:7:6:7:13 | ControlFlowNode for __file__ |
| package.__init__ | attr_used_in_subpackage | package/__init__.py:4:1:4:23 | attr_used_in_subpackage | | package.__init__ | attr_used_in_subpackage | package/__init__.py:4:1:4:23 | ControlFlowNode for attr_used_in_subpackage |
| package.__init__ | enter | package/__init__.py:2:1:2:5 | enter | | package.__init__ | enter | package/__init__.py:2:1:2:5 | ControlFlowNode for enter |
| package.__init__ | exit | package/__init__.py:7:1:7:4 | exit | | package.__init__ | exit | package/__init__.py:7:1:7:4 | ControlFlowNode for exit |
| package.__init__ | package_attr | package/__init__.py:5:1:5:12 | package_attr | | package.__init__ | package_attr | package/__init__.py:5:1:5:12 | ControlFlowNode for package_attr |
| package.subpackage2.__init__ | __file__ | package/subpackage2/__init__.py:6:6:6:13 | __file__ | | package.subpackage2.__init__ | __file__ | package/subpackage2/__init__.py:6:6:6:13 | ControlFlowNode for __file__ |
| package.subpackage2.__init__ | enter | package/subpackage2/__init__.py:2:1:2:5 | enter | | package.subpackage2.__init__ | enter | package/subpackage2/__init__.py:2:1:2:5 | ControlFlowNode for enter |
| package.subpackage2.__init__ | exit | package/subpackage2/__init__.py:6:1:6:4 | exit | | package.subpackage2.__init__ | exit | package/subpackage2/__init__.py:6:1:6:4 | ControlFlowNode for exit |
| package.subpackage2.__init__ | subpackage2_attr | package/subpackage2/__init__.py:4:1:4:16 | subpackage2_attr | | package.subpackage2.__init__ | subpackage2_attr | package/subpackage2/__init__.py:4:1:4:16 | ControlFlowNode for subpackage2_attr |
| package.subpackage.__init__ | __file__ | package/subpackage/__init__.py:14:6:14:13 | __file__ | | package.subpackage.__init__ | __file__ | package/subpackage/__init__.py:14:6:14:13 | ControlFlowNode for __file__ |
| package.subpackage.__init__ | check | package/subpackage/__init__.py:12:1:12:5 | check | | package.subpackage.__init__ | check | package/subpackage/__init__.py:12:1:12:5 | ControlFlowNode for check |
| package.subpackage.__init__ | enter | package/subpackage/__init__.py:2:1:2:5 | enter | | package.subpackage.__init__ | enter | package/subpackage/__init__.py:2:1:2:5 | ControlFlowNode for enter |
| package.subpackage.__init__ | exit | package/subpackage/__init__.py:14:1:14:4 | exit | | package.subpackage.__init__ | exit | package/subpackage/__init__.py:14:1:14:4 | ControlFlowNode for exit |
| package.subpackage.__init__ | globals | package/subpackage/__init__.py:12:79:12:85 | globals | | package.subpackage.__init__ | globals | package/subpackage/__init__.py:12:79:12:85 | ControlFlowNode for globals |
| package.subpackage.__init__ | imported_attr | package/subpackage/__init__.py:7:16:7:55 | After ImportMember | | package.subpackage.__init__ | imported_attr | package/subpackage/__init__.py:7:16:7:55 | ControlFlowNode for ImportMember |
| package.subpackage.__init__ | imported_attr | package/subpackage/__init__.py:8:24:8:36 | imported_attr | | package.subpackage.__init__ | imported_attr | package/subpackage/__init__.py:8:24:8:36 | ControlFlowNode for imported_attr |
| package.subpackage.__init__ | irrelevant_attr | package/subpackage/__init__.py:11:24:11:38 | After ImportMember | | package.subpackage.__init__ | irrelevant_attr | package/subpackage/__init__.py:11:24:11:38 | ControlFlowNode for ImportMember |
| package.subpackage.__init__ | irrelevant_attr | package/subpackage/__init__.py:11:24:11:38 | irrelevant_attr | | package.subpackage.__init__ | irrelevant_attr | package/subpackage/__init__.py:11:24:11:38 | ControlFlowNode for irrelevant_attr |
| package.subpackage.__init__ | submodule | package/subpackage/__init__.py:12:35:12:43 | submodule | | package.subpackage.__init__ | submodule | package/subpackage/__init__.py:12:35:12:43 | ControlFlowNode for submodule |
| package.subpackage.__init__ | subpackage_attr | package/subpackage/__init__.py:4:1:4:15 | subpackage_attr | | package.subpackage.__init__ | subpackage_attr | package/subpackage/__init__.py:4:1:4:15 | ControlFlowNode for subpackage_attr |
| package.subpackage.submodule | __file__ | package/subpackage/submodule.py:7:6:7:13 | __file__ | | package.subpackage.submodule | __file__ | package/subpackage/submodule.py:7:6:7:13 | ControlFlowNode for __file__ |
| package.subpackage.submodule | enter | package/subpackage/submodule.py:2:1:2:5 | enter | | package.subpackage.submodule | enter | package/subpackage/submodule.py:2:1:2:5 | ControlFlowNode for enter |
| package.subpackage.submodule | exit | package/subpackage/submodule.py:7:1:7:4 | exit | | package.subpackage.submodule | exit | package/subpackage/submodule.py:7:1:7:4 | ControlFlowNode for exit |
| package.subpackage.submodule | irrelevant_attr | package/subpackage/submodule.py:5:1:5:15 | irrelevant_attr | | package.subpackage.submodule | irrelevant_attr | package/subpackage/submodule.py:5:1:5:15 | ControlFlowNode for irrelevant_attr |
| package.subpackage.submodule | submodule_attr | package/subpackage/submodule.py:4:1:4:14 | submodule_attr | | package.subpackage.submodule | submodule_attr | package/subpackage/submodule.py:4:1:4:14 | ControlFlowNode for submodule_attr |
| refined | SOURCE | refined.py:12:25:12:30 | SOURCE | | refined | SOURCE | refined.py:12:25:12:30 | ControlFlowNode for SOURCE |
| refined | __file__ | refined.py:14:6:14:13 | __file__ | | refined | __file__ | refined.py:14:6:14:13 | ControlFlowNode for __file__ |
| refined | check | refined.py:12:1:12:5 | check | | refined | check | refined.py:12:1:12:5 | ControlFlowNode for check |
| refined | enter | refined.py:2:1:2:5 | enter | | refined | enter | refined.py:2:1:2:5 | ControlFlowNode for enter |
| refined | exit | refined.py:14:1:14:4 | exit | | refined | exit | refined.py:14:1:14:4 | ControlFlowNode for exit |
| refined | globals | refined.py:12:33:12:39 | globals | | refined | globals | refined.py:12:33:12:39 | ControlFlowNode for globals |
| refined | object | refined.py:4:14:4:19 | object | | refined | object | refined.py:4:14:4:19 | ControlFlowNode for object |
| simplistic_reexport | __file__ | simplistic_reexport.py:19:6:19:13 | __file__ | | simplistic_reexport | __file__ | simplistic_reexport.py:19:6:19:13 | ControlFlowNode for __file__ |
| simplistic_reexport | bar_attr | simplistic_reexport.py:6:17:6:24 | After ImportMember | | simplistic_reexport | bar_attr | simplistic_reexport.py:6:17:6:24 | ControlFlowNode for ImportMember |
| simplistic_reexport | bar_attr | simplistic_reexport.py:10:19:10:26 | bar_attr | | simplistic_reexport | bar_attr | simplistic_reexport.py:10:19:10:26 | ControlFlowNode for bar_attr |
| simplistic_reexport | baz_attr | baz.py:4:1:4:8 | baz_attr | | simplistic_reexport | baz_attr | baz.py:4:1:4:8 | ControlFlowNode for baz_attr |
| simplistic_reexport | baz_attr | simplistic_reexport.py:17:19:17:26 | baz_attr | | simplistic_reexport | baz_attr | simplistic_reexport.py:17:19:17:26 | ControlFlowNode for baz_attr |
| simplistic_reexport | check | simplistic_reexport.py:17:1:17:5 | check | | simplistic_reexport | check | simplistic_reexport.py:17:1:17:5 | ControlFlowNode for check |
| simplistic_reexport | enter | baz.py:2:1:2:5 | enter | | simplistic_reexport | enter | baz.py:2:1:2:5 | ControlFlowNode for enter |
| simplistic_reexport | enter | simplistic_reexport.py:4:1:4:5 | enter | | simplistic_reexport | enter | simplistic_reexport.py:4:1:4:5 | ControlFlowNode for enter |
| simplistic_reexport | exit | baz.py:6:1:6:4 | exit | | simplistic_reexport | exit | baz.py:6:1:6:4 | ControlFlowNode for exit |
| simplistic_reexport | exit | simplistic_reexport.py:19:1:19:4 | exit | | simplistic_reexport | exit | simplistic_reexport.py:19:1:19:4 | ControlFlowNode for exit |
| simplistic_reexport | globals | simplistic_reexport.py:17:44:17:50 | globals | | simplistic_reexport | globals | simplistic_reexport.py:17:44:17:50 | ControlFlowNode for globals |

View File

@@ -3,7 +3,6 @@ import semmle.python.dataflow.new.DataFlow
import semmle.python.ApiGraphs import semmle.python.ApiGraphs
import utils.test.InlineExpectationsTest import utils.test.InlineExpectationsTest
import semmle.python.dataflow.new.internal.ImportResolution import semmle.python.dataflow.new.internal.ImportResolution
private import semmle.python.controlflow.internal.Cfg as Cfg
/** A string that appears on the right hand side of an assignment. */ /** A string that appears on the right hand side of an assignment. */
private class SourceString extends DataFlow::Node { private class SourceString extends DataFlow::Node {
@@ -46,7 +45,7 @@ private class VersionGuardedNode extends DataFlow::Node {
VersionGuardedNode() { VersionGuardedNode() {
version in [2, 3] and version in [2, 3] and
exists(If parent, Cfg::CompareNode c, Cfg::ControlFlowNode litCfg | exists(If parent, CompareNode c, ControlFlowNode litCfg |
parent.getBody().contains(this.asExpr()) and parent.getBody().contains(this.asExpr()) and
litCfg.getNode() = any(IntegerLiteral lit | lit.getValue() = version) litCfg.getNode() = any(IntegerLiteral lit | lit.getValue() = version)
| |

View File

@@ -2,6 +2,5 @@ testFailures
debug_callableNotUnique debug_callableNotUnique
pointsTo_found_typeTracker_notFound pointsTo_found_typeTracker_notFound
typeTracker_found_pointsTo_notFound typeTracker_found_pointsTo_notFound
| type_annotations.py:6:5:6:14 | Attribute() | Foo.method | | type_annotations.py:6:5:6:14 | ControlFlowNode for Attribute() | Foo.method |
| type_annotations.py:16:5:16:14 | Attribute() | Foo.method | | type_annotations.py:16:5:16:14 | ControlFlowNode for Attribute() | Foo.method |
| type_annotations.py:29:5:29:14 | Attribute() | Foo.method |

Some files were not shown because too many files have changed in this diff Show More