change query id

Java/Python: Fix some potential performance problems due to transitive deltas.

.github/workflows/csv-coverage-pr-artifacts.yml

@@ -6,6 +6,8 @@ on:
       - '.github/workflows/csv-coverage-pr-comment.yml'
       - '*/ql/src/**/*.ql'
       - '*/ql/src/**/*.qll'
+      - '*/ql/lib/**/*.ql'
+      - '*/ql/lib/**/*.qll'
       - 'misc/scripts/library-coverage/*.py'
       # input data files
       - '*/documentation/library-coverage/cwe-sink.csv'

.gitignore

@@ -24,3 +24,6 @@
 /codeql/
 
 csharp/extractor/Semmle.Extraction.CSharp.Driver/Properties/launchSettings.json
+
+# Avoid committing cached package components
+.codeql

README.md

@@ -4,8 +4,8 @@ This open source repository contains the standard CodeQL libraries and queries t
 
 ## How do I learn CodeQL and run queries?
 
-There is [extensive documentation](https://help.semmle.com/QL/learn-ql/) on getting started with writing CodeQL.
+There is [extensive documentation](https://codeql.github.com/docs/) on getting started with writing CodeQL.
-You can use the [interactive query console](https://lgtm.com/help/lgtm/using-query-console) on LGTM.com or the [CodeQL for Visual Studio Code](https://help.semmle.com/codeql/codeql-for-vscode.html) extension to try out your queries on any open source project that's currently being analyzed.
+You can use the [interactive query console](https://lgtm.com/help/lgtm/using-query-console) on LGTM.com or the [CodeQL for Visual Studio Code](https://codeql.github.com/docs/codeql-for-visual-studio-code/) extension to try out your queries on any open source project that's currently being analyzed.
 
 ## Contributing
 

config/identical-files.json

@@ -367,6 +367,7 @@
   ],
   "Inline Test Expectations": [
     "cpp/ql/test/TestUtilities/InlineExpectationsTest.qll",
+    "csharp/ql/test/TestUtilities/InlineExpectationsTest.qll",
     "java/ql/test/TestUtilities/InlineExpectationsTest.qll",
     "python/ql/test/TestUtilities/InlineExpectationsTest.qll"
   ],
@@ -461,5 +462,12 @@
   "ReDoS Polynomial Python/JS": [
     "javascript/ql/lib/semmle/javascript/security/performance/SuperlinearBackTracking.qll",
     "python/ql/lib/semmle/python/security/performance/SuperlinearBackTracking.qll"
+  ],
+  "CodeQL Tutorial": [
+    "cpp/ql/lib/tutorial.qll",
+    "csharp/ql/lib/tutorial.qll",
+    "java/ql/lib/tutorial.qll",
+    "javascript/ql/lib/tutorial.qll",
+    "python/ql/lib/tutorial.qll"
   ]
 }

cpp/change-notes/2021-06-10-cleartext-transmission.md

@@ -0,0 +1,2 @@
+lgtm,codescanning
+* A new query (`cpp/cleartext-transmission`) has been added. This is similar to the `cpp/cleartext-storage-file`, `cpp/cleartext-storage-buffer` and `cpp/cleartext-storage-database` queries but looks for cases where sensitive information is most likely transmitted over a network.

cpp/change-notes/2021-09-27-command-line-injection.md

@@ -0,0 +1,2 @@
+lgtm,codescanning
+* The "Uncontrolled data used in OS command" (`cpp/command-line-injection`) query has been enhanced to reduce false positive results and its `@precision` increased to `high`

cpp/change-notes/2021-09-27-overflow-static.md

@@ -0,0 +1,3 @@
+lgtm,codescanning
+* Increase precision to high for the "Static buffer overflow" query
+  (`cpp/static-buffer-overflow`). This means the query is run and displayed by default on Code Scanning and LGTM.

cpp/change-notes/2021-10-01-improper-null-termination.md

@@ -0,0 +1,2 @@
+lgtm,codescanning
+* Several improvements made to the `NullTermination.qll` library and the 'Potential improper null termination' (cpp/improper-null-termination). These changes reduce the number of false positive results for this query and related query 'User-controlled data may not be null terminated' (cpp/user-controlled-null-termination-tainted).

cpp/change-notes/2021-10-07-extraction-errors.md

@@ -0,0 +1,3 @@
+codescanning
+* Problems with extraction that in most cases won't break the analysis in a significant way are now reported as warnings rather than errors.
+* The failed extractor invocations query now has severity `error`.

cpp/ql/lib/semmle/code/cpp/Declaration.qll

@@ -275,9 +275,8 @@ class Declaration extends Locatable, @declaration {
    * `getTemplateArgumentKind(0)`.
    */
   final Locatable getTemplateArgumentKind(int index) {
-    if exists(getTemplateArgumentValue(index))
+    exists(getTemplateArgumentValue(index)) and
-    then result = getTemplateArgumentType(index)
+    result = getTemplateArgumentType(index)
-    else none()
   }
 
   /** Gets the number of template arguments for this declaration. */

cpp/ql/lib/semmle/code/cpp/File.qll

@@ -38,7 +38,7 @@ class Container extends Locatable, @container {
    * DEPRECATED: Use `getLocation` instead.
    * Gets a URL representing the location of this container.
    *
-   * For more information see [Providing URLs](https://help.semmle.com/QL/learn-ql/ql/locations.html#providing-urls).
+   * For more information see [Providing URLs](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/#providing-urls).
    */
   deprecated string getURL() { none() } // overridden by subclasses
 

cpp/ql/lib/semmle/code/cpp/Location.qll

@@ -61,7 +61,7 @@ class Location extends @location {
    * The location spans column `startcolumn` of line `startline` to
    * column `endcolumn` of line `endline` in file `filepath`.
    * For more information, see
-   * [Locations](https://help.semmle.com/QL/learn-ql/ql/locations.html).
+   * [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
    */
   predicate hasLocationInfo(
     string filepath, int startline, int startcolumn, int endline, int endcolumn

cpp/ql/lib/semmle/code/cpp/Type.qll

@@ -1650,7 +1650,6 @@ class RoutineType extends Type, @routinetype {
     i = 0 and result = "" and not exists(this.getAParameterType())
     or
     (
-      exists(this.getParameterType(i)) and
       if i < max(int j | exists(this.getParameterType(j)))
       then
         // Not the last one

cpp/ql/lib/semmle/code/cpp/XML.qll

@@ -24,7 +24,7 @@ class XMLLocatable extends @xmllocatable, TXMLLocatable {
    * The location spans column `startcolumn` of line `startline` to
    * column `endcolumn` of line `endline` in file `filepath`.
    * For more information, see
-   * [Locations](https://help.semmle.com/QL/learn-ql/ql/locations.html).
+   * [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
    */
   predicate hasLocationInfo(
     string filepath, int startline, int startcolumn, int endline, int endcolumn
@@ -108,7 +108,7 @@ class XMLParent extends @xmlparent {
   }
 
   /** Gets the text value contained in this XML parent. */
-  string getTextValue() { result = allCharactersString() }
+  string getTextValue() { result = this.allCharactersString() }
 
   /** Gets a printable representation of this XML parent. */
   string toString() { result = this.getName() }
@@ -119,7 +119,7 @@ class XMLFile extends XMLParent, File {
   XMLFile() { xmlEncoding(this, _) }
 
   /** Gets a printable representation of this XML file. */
-  override string toString() { result = getName() }
+  override string toString() { result = this.getName() }
 
   /** Gets the name of this XML file. */
   override string getName() { result = File.super.getAbsolutePath() }
@@ -129,14 +129,14 @@ class XMLFile extends XMLParent, File {
    *
    * Gets the path of this XML file.
    */
-  deprecated string getPath() { result = getAbsolutePath() }
+  deprecated string getPath() { result = this.getAbsolutePath() }
 
   /**
    * DEPRECATED: Use `getParentContainer().getAbsolutePath()` instead.
    *
    * Gets the path of the folder that contains this XML file.
    */
-  deprecated string getFolder() { result = getParentContainer().getAbsolutePath() }
+  deprecated string getFolder() { result = this.getParentContainer().getAbsolutePath() }
 
   /** Gets the encoding of this XML file. */
   string getEncoding() { xmlEncoding(this, result) }
@@ -200,7 +200,7 @@ class XMLDTD extends XMLLocatable, @xmldtd {
  */
 class XMLElement extends @xmlelement, XMLParent, XMLLocatable {
   /** Holds if this XML element has the given `name`. */
-  predicate hasName(string name) { name = getName() }
+  predicate hasName(string name) { name = this.getName() }
 
   /** Gets the name of this XML element. */
   override string getName() { xmlElements(this, result, _, _, _) }
@@ -239,7 +239,7 @@ class XMLElement extends @xmlelement, XMLParent, XMLLocatable {
   string getAttributeValue(string name) { result = this.getAttribute(name).getValue() }
 
   /** Gets a printable representation of this XML element. */
-  override string toString() { result = getName() }
+  override string toString() { result = this.getName() }
 }
 
 /**

cpp/ql/lib/semmle/code/cpp/commons/Buffer.qll

@@ -10,44 +10,11 @@ import semmle.code.cpp.dataflow.DataFlow
  *   char data[1]; // v
  * };
  * ```
- * This requires that `v` is an array of size 0 or 1, and `v` is the last member of `c`.
+ * This requires that `v` is an array of size 0 or 1.
- * In addition, if the size of the structure is taken, there must be at least one instance
- * where a `c` pointer is allocated with additional space.
- * For example, holds for `c` if it occurs as
- * ```
- * malloc(sizeof(c) + 100 * sizeof(char))
- * ```
- * but not if it only ever occurs as
- * ```
- * malloc(sizeof(c))
- * ```
  */
 predicate memberMayBeVarSize(Class c, MemberVariable v) {
-  exists(int i |
+  c = v.getDeclaringType() and
-    // `v` is the last field in `c`
+  v.getUnspecifiedType().(ArrayType).getArraySize() <= 1
-    i = max(int j | c.getCanonicalMember(j) instanceof Field | j) and
-    v = c.getCanonicalMember(i) and
-    // v is an array of size at most 1
-    v.getUnspecifiedType().(ArrayType).getArraySize() <= 1 and
-    not c instanceof Union
-  ) and
-  // If the size is taken, then arithmetic is performed on the result at least once
-  (
-    // `sizeof(c)` is not taken
-    not exists(SizeofOperator so |
-      so.(SizeofTypeOperator).getTypeOperand().getUnspecifiedType() = c or
-      so.(SizeofExprOperator).getExprOperand().getUnspecifiedType() = c
-    )
-    or
-    // or `sizeof(c)` is taken
-    exists(SizeofOperator so |
-      so.(SizeofTypeOperator).getTypeOperand().getUnspecifiedType() = c or
-      so.(SizeofExprOperator).getExprOperand().getUnspecifiedType() = c
-    |
-      // and arithmetic is performed on the result
-      so.getParent*() instanceof AddExpr
-    )
-  )
 }
 
 /**
@@ -60,10 +27,6 @@ int getBufferSize(Expr bufferExpr, Element why) {
     result = bufferVar.getUnspecifiedType().(ArrayType).getSize() and
     why = bufferVar and
     not memberMayBeVarSize(_, bufferVar) and
-    not exists(Union bufferType |
-      bufferType.getAMemberVariable() = why and
-      bufferVar.getUnspecifiedType().(ArrayType).getSize() <= 1
-    ) and
     not result = 0 // zero sized arrays are likely to have special usage, for example
     or
     // behaving a bit like a 'union' overlapping other fields.
@@ -85,13 +48,6 @@ int getBufferSize(Expr bufferExpr, Element why) {
       parentPtr.getTarget().getUnspecifiedType().(PointerType).getBaseType() = parentClass and
       result = getBufferSize(parentPtr, _) + bufferVar.getType().getSize() - parentClass.getSize()
     )
-    or
-    exists(Union bufferType |
-      bufferType.getAMemberVariable() = why and
-      why = bufferVar and
-      bufferVar.getUnspecifiedType().(ArrayType).getSize() <= 1 and
-      result = bufferType.getSize()
-    )
   )
   or
   // buffer is a fixed size dynamic allocation

cpp/ql/lib/semmle/code/cpp/commons/NullTermination.qll

@@ -1,6 +1,7 @@
 import cpp
 private import semmle.code.cpp.models.interfaces.ArrayFunction
 private import semmle.code.cpp.models.implementations.Strcat
+import semmle.code.cpp.dataflow.DataFlow
 
 private predicate mayAddNullTerminatorHelper(Expr e, VariableAccess va, Expr e0) {
   exists(StackVariable v0, Expr val |
@@ -45,22 +46,28 @@ predicate mayAddNullTerminator(Expr e, VariableAccess va) {
     ae.getRValue().getAChild*() = va
   )
   or
-  // Function call: library function, varargs function, function
+  // Function calls...
-  // containing assembler code, or function where the relevant
-  // parameter is potentially added a null terminator.
   exists(Call c, Function f, int i |
     e = c and
     f = c.getTarget() and
     not functionArgumentMustBeNullTerminated(f, i) and
     c.getAnArgumentSubExpr(i) = va
   |
-    not f.hasEntryPoint() and not functionArgumentMustBeNullTerminated(f, i)
+    // library function
+    not f.hasEntryPoint()
     or
+    // function where the relevant parameter is potentially added a null terminator
     mayAddNullTerminator(_, f.getParameter(i).getAnAccess())
     or
+    // varargs function
     f.isVarargs() and i >= f.getNumberOfParameters()
     or
+    // function containing assembler code
     exists(AsmStmt s | s.getEnclosingFunction() = f)
+    or
+    // function where the relevant parameter is returned (leaking it to be potentially null terminated elsewhere)
+    DataFlow::localFlow(DataFlow::parameterNode(f.getParameter(i)),
+      DataFlow::exprNode(any(ReturnStmt rs).getExpr()))
   )
   or
   // Call without target (e.g., function pointer call)
@@ -93,6 +100,15 @@ predicate variableMustBeNullTerminated(VariableAccess va) {
       fc.getArgument(i) = va
     )
     or
+    // String argument to a formatting function (such as `printf`)
+    exists(int n, FormatLiteral fl |
+      fc.(FormattingFunctionCall).getConversionArgument(n) = va and
+      fl = fc.(FormattingFunctionCall).getFormat() and
+      fl.getConversionType(n) instanceof PointerType and // `%s`, `%ws` etc
+      not fl.getConversionType(n) instanceof VoidPointerType and // exclude: `%p`
+      not fl.hasPrecision(n) // exclude: `%.*s`
+    )
+    or
     // Call to a wrapper function that requires null termination
     // (not itself adding a null terminator)
     exists(Function wrapper, int i, Parameter p, VariableAccess use |

cpp/ql/lib/semmle/code/cpp/commons/Printf.qll

@@ -253,6 +253,21 @@ class FormattingFunctionCall extends Expr {
     // format arguments must be known
     exists(getTarget().(FormattingFunction).getFirstFormatArgumentIndex())
   }
+
+  /**
+   * Gets the argument, if any, to which the output is written. If `isStream` is
+   * `true`, the output argument is a stream (that is, this call behaves like
+   * `fprintf`). If `isStream` is `false`, the output argument is a buffer (that
+   * is, this call behaves like `sprintf`)
+   */
+  Expr getOutputArgument(boolean isStream) {
+    result =
+      this.(Call)
+          .getArgument(this.(Call)
+                .getTarget()
+                .(FormattingFunction)
+                .getOutputParameterIndex(isStream))
+  }
 }
 
 /**

cpp/ql/lib/semmle/code/cpp/controlflow/BasicBlocks.qll

@@ -194,7 +194,7 @@ class BasicBlock extends ControlFlowNodeBase {
    * The location spans column `startcolumn` of line `startline` to
    * column `endcolumn` of line `endline` in file `filepath`.
    * For more information, see
-   * [Locations](https://help.semmle.com/QL/learn-ql/ql/locations.html).
+   * [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
    *
    * Yields no result if this basic block spans multiple source files.
    */

cpp/ql/lib/semmle/code/cpp/controlflow/internal/ConstantExprs.qll

@@ -344,14 +344,13 @@ private int convertIntToType(int val, IntegralType t) {
   then if val = 0 then result = 0 else result = 1
   else
     if t.isUnsigned()
-    then if val >= 0 and val.bitShiftRight(t.getSize() * 8) = 0 then result = val else none()
+    then val >= 0 and val.bitShiftRight(t.getSize() * 8) = 0 and result = val
     else
       if val >= 0 and val.bitShiftRight(t.getSize() * 8 - 1) = 0
       then result = val
-      else
+      else (
-        if (-(val + 1)).bitShiftRight(t.getSize() * 8 - 1) = 0
+        (-(val + 1)).bitShiftRight(t.getSize() * 8 - 1) = 0 and result = val
-        then result = val
+      )
-        else none()
 }
 
 /**

cpp/ql/lib/semmle/code/cpp/dataflow/internal/DataFlowImpl.qll

@@ -2139,7 +2139,8 @@ private predicate expensiveLen2unfolding(TypedContent tc, Configuration config)
       ) and
     accessPathApproxCostLimits(apLimit, tupleLimit) and
     apLimit < tails and
-    tupleLimit < (tails - 1) * nodes
+    tupleLimit < (tails - 1) * nodes and
+    not tc.forceHighPrecision()
   )
 }
 
@@ -2973,12 +2974,15 @@ private AccessPathApprox getATail(AccessPathApprox apa, Configuration config) {
  * expected to be expensive. Holds with `unfold = true` otherwise.
  */
 private predicate evalUnfold(AccessPathApprox apa, boolean unfold, Configuration config) {
-  exists(int aps, int nodes, int apLimit, int tupleLimit |
+  if apa.getHead().forceHighPrecision()
-    aps = countPotentialAps(apa, config) and
+  then unfold = true
-    nodes = countNodesUsingAccessPath(apa, config) and
+  else
-    accessPathCostLimits(apLimit, tupleLimit) and
+    exists(int aps, int nodes, int apLimit, int tupleLimit |
-    if apLimit < aps and tupleLimit < (aps - 1) * nodes then unfold = false else unfold = true
+      aps = countPotentialAps(apa, config) and
-  )
+      nodes = countNodesUsingAccessPath(apa, config) and
+      accessPathCostLimits(apLimit, tupleLimit) and
+      if apLimit < aps and tupleLimit < (aps - 1) * nodes then unfold = false else unfold = true
+    )
 }
 
 /**
@@ -3248,7 +3252,7 @@ class PathNode extends TPathNode {
    * The location spans column `startcolumn` of line `startline` to
    * column `endcolumn` of line `endline` in file `filepath`.
    * For more information, see
-   * [Locations](https://help.semmle.com/QL/learn-ql/ql/locations.html).
+   * [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
    */
   predicate hasLocationInfo(
     string filepath, int startline, int startcolumn, int endline, int endcolumn
@@ -4033,7 +4037,7 @@ private module FlowExploration {
      * The location spans column `startcolumn` of line `startline` to
      * column `endcolumn` of line `endline` in file `filepath`.
      * For more information, see
-     * [Locations](https://help.semmle.com/QL/learn-ql/ql/locations.html).
+     * [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
      */
     predicate hasLocationInfo(
       string filepath, int startline, int startcolumn, int endline, int endcolumn

cpp/ql/lib/semmle/code/cpp/dataflow/internal/DataFlowImpl2.qll

@@ -2139,7 +2139,8 @@ private predicate expensiveLen2unfolding(TypedContent tc, Configuration config)
       ) and
     accessPathApproxCostLimits(apLimit, tupleLimit) and
     apLimit < tails and
-    tupleLimit < (tails - 1) * nodes
+    tupleLimit < (tails - 1) * nodes and
+    not tc.forceHighPrecision()
   )
 }
 
@@ -2973,12 +2974,15 @@ private AccessPathApprox getATail(AccessPathApprox apa, Configuration config) {
  * expected to be expensive. Holds with `unfold = true` otherwise.
  */
 private predicate evalUnfold(AccessPathApprox apa, boolean unfold, Configuration config) {
-  exists(int aps, int nodes, int apLimit, int tupleLimit |
+  if apa.getHead().forceHighPrecision()
-    aps = countPotentialAps(apa, config) and
+  then unfold = true
-    nodes = countNodesUsingAccessPath(apa, config) and
+  else
-    accessPathCostLimits(apLimit, tupleLimit) and
+    exists(int aps, int nodes, int apLimit, int tupleLimit |
-    if apLimit < aps and tupleLimit < (aps - 1) * nodes then unfold = false else unfold = true
+      aps = countPotentialAps(apa, config) and
-  )
+      nodes = countNodesUsingAccessPath(apa, config) and
+      accessPathCostLimits(apLimit, tupleLimit) and
+      if apLimit < aps and tupleLimit < (aps - 1) * nodes then unfold = false else unfold = true
+    )
 }
 
 /**
@@ -3248,7 +3252,7 @@ class PathNode extends TPathNode {
    * The location spans column `startcolumn` of line `startline` to
    * column `endcolumn` of line `endline` in file `filepath`.
    * For more information, see
-   * [Locations](https://help.semmle.com/QL/learn-ql/ql/locations.html).
+   * [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
    */
   predicate hasLocationInfo(
     string filepath, int startline, int startcolumn, int endline, int endcolumn
@@ -4033,7 +4037,7 @@ private module FlowExploration {
      * The location spans column `startcolumn` of line `startline` to
      * column `endcolumn` of line `endline` in file `filepath`.
      * For more information, see
-     * [Locations](https://help.semmle.com/QL/learn-ql/ql/locations.html).
+     * [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
      */
     predicate hasLocationInfo(
       string filepath, int startline, int startcolumn, int endline, int endcolumn

cpp/ql/lib/semmle/code/cpp/dataflow/internal/DataFlowImpl3.qll

@@ -2139,7 +2139,8 @@ private predicate expensiveLen2unfolding(TypedContent tc, Configuration config)
       ) and
     accessPathApproxCostLimits(apLimit, tupleLimit) and
     apLimit < tails and
-    tupleLimit < (tails - 1) * nodes
+    tupleLimit < (tails - 1) * nodes and
+    not tc.forceHighPrecision()
   )
 }
 
@@ -2973,12 +2974,15 @@ private AccessPathApprox getATail(AccessPathApprox apa, Configuration config) {
  * expected to be expensive. Holds with `unfold = true` otherwise.
  */
 private predicate evalUnfold(AccessPathApprox apa, boolean unfold, Configuration config) {
-  exists(int aps, int nodes, int apLimit, int tupleLimit |
+  if apa.getHead().forceHighPrecision()
-    aps = countPotentialAps(apa, config) and
+  then unfold = true
-    nodes = countNodesUsingAccessPath(apa, config) and
+  else
-    accessPathCostLimits(apLimit, tupleLimit) and
+    exists(int aps, int nodes, int apLimit, int tupleLimit |
-    if apLimit < aps and tupleLimit < (aps - 1) * nodes then unfold = false else unfold = true
+      aps = countPotentialAps(apa, config) and
-  )
+      nodes = countNodesUsingAccessPath(apa, config) and
+      accessPathCostLimits(apLimit, tupleLimit) and
+      if apLimit < aps and tupleLimit < (aps - 1) * nodes then unfold = false else unfold = true
+    )
 }
 
 /**
@@ -3248,7 +3252,7 @@ class PathNode extends TPathNode {
    * The location spans column `startcolumn` of line `startline` to
    * column `endcolumn` of line `endline` in file `filepath`.
    * For more information, see
-   * [Locations](https://help.semmle.com/QL/learn-ql/ql/locations.html).
+   * [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
    */
   predicate hasLocationInfo(
     string filepath, int startline, int startcolumn, int endline, int endcolumn
@@ -4033,7 +4037,7 @@ private module FlowExploration {
      * The location spans column `startcolumn` of line `startline` to
      * column `endcolumn` of line `endline` in file `filepath`.
      * For more information, see
-     * [Locations](https://help.semmle.com/QL/learn-ql/ql/locations.html).
+     * [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
      */
     predicate hasLocationInfo(
       string filepath, int startline, int startcolumn, int endline, int endcolumn

cpp/ql/lib/semmle/code/cpp/dataflow/internal/DataFlowImpl4.qll

@@ -2139,7 +2139,8 @@ private predicate expensiveLen2unfolding(TypedContent tc, Configuration config)
       ) and
     accessPathApproxCostLimits(apLimit, tupleLimit) and
     apLimit < tails and
-    tupleLimit < (tails - 1) * nodes
+    tupleLimit < (tails - 1) * nodes and
+    not tc.forceHighPrecision()
   )
 }
 
@@ -2973,12 +2974,15 @@ private AccessPathApprox getATail(AccessPathApprox apa, Configuration config) {
  * expected to be expensive. Holds with `unfold = true` otherwise.
  */
 private predicate evalUnfold(AccessPathApprox apa, boolean unfold, Configuration config) {
-  exists(int aps, int nodes, int apLimit, int tupleLimit |
+  if apa.getHead().forceHighPrecision()
-    aps = countPotentialAps(apa, config) and
+  then unfold = true
-    nodes = countNodesUsingAccessPath(apa, config) and
+  else
-    accessPathCostLimits(apLimit, tupleLimit) and
+    exists(int aps, int nodes, int apLimit, int tupleLimit |
-    if apLimit < aps and tupleLimit < (aps - 1) * nodes then unfold = false else unfold = true
+      aps = countPotentialAps(apa, config) and
-  )
+      nodes = countNodesUsingAccessPath(apa, config) and
+      accessPathCostLimits(apLimit, tupleLimit) and
+      if apLimit < aps and tupleLimit < (aps - 1) * nodes then unfold = false else unfold = true
+    )
 }
 
 /**
@@ -3248,7 +3252,7 @@ class PathNode extends TPathNode {
    * The location spans column `startcolumn` of line `startline` to
    * column `endcolumn` of line `endline` in file `filepath`.
    * For more information, see
-   * [Locations](https://help.semmle.com/QL/learn-ql/ql/locations.html).
+   * [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
    */
   predicate hasLocationInfo(
     string filepath, int startline, int startcolumn, int endline, int endcolumn
@@ -4033,7 +4037,7 @@ private module FlowExploration {
      * The location spans column `startcolumn` of line `startline` to
      * column `endcolumn` of line `endline` in file `filepath`.
      * For more information, see
-     * [Locations](https://help.semmle.com/QL/learn-ql/ql/locations.html).
+     * [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
      */
     predicate hasLocationInfo(
       string filepath, int startline, int startcolumn, int endline, int endcolumn

cpp/ql/lib/semmle/code/cpp/dataflow/internal/DataFlowImplCommon.qll

@@ -937,7 +937,7 @@ class CallContextSpecificCall extends CallContextCall, TSpecificCall {
   }
 
   override predicate relevantFor(DataFlowCallable callable) {
-    recordDataFlowCallSite(getCall(), callable)
+    recordDataFlowCallSite(this.getCall(), callable)
   }
 
   override predicate matchesCall(DataFlowCall call) { call = this.getCall() }
@@ -1236,6 +1236,13 @@ class TypedContent extends MkTypedContent {
 
   /** Gets a textual representation of this content. */
   string toString() { result = c.toString() }
+
+  /**
+   * Holds if access paths with this `TypedContent` at their head always should
+   * be tracked at high precision. This disables adaptive access path precision
+   * for such access paths.
+   */
+  predicate forceHighPrecision() { forceHighPrecision(c) }
 }
 
 /**
@@ -1250,7 +1257,7 @@ abstract class AccessPathFront extends TAccessPathFront {
 
   TypedContent getHead() { this = TFrontHead(result) }
 
-  predicate isClearedAt(Node n) { clearsContentCached(n, getHead().getContent()) }
+  predicate isClearedAt(Node n) { clearsContentCached(n, this.getHead().getContent()) }
 }
 
 class AccessPathFrontNil extends AccessPathFront, TFrontNil {

cpp/ql/lib/semmle/code/cpp/dataflow/internal/DataFlowImplLocal.qll

@@ -2139,7 +2139,8 @@ private predicate expensiveLen2unfolding(TypedContent tc, Configuration config)
       ) and
     accessPathApproxCostLimits(apLimit, tupleLimit) and
     apLimit < tails and
-    tupleLimit < (tails - 1) * nodes
+    tupleLimit < (tails - 1) * nodes and
+    not tc.forceHighPrecision()
   )
 }
 
@@ -2973,12 +2974,15 @@ private AccessPathApprox getATail(AccessPathApprox apa, Configuration config) {
  * expected to be expensive. Holds with `unfold = true` otherwise.
  */
 private predicate evalUnfold(AccessPathApprox apa, boolean unfold, Configuration config) {
-  exists(int aps, int nodes, int apLimit, int tupleLimit |
+  if apa.getHead().forceHighPrecision()
-    aps = countPotentialAps(apa, config) and
+  then unfold = true
-    nodes = countNodesUsingAccessPath(apa, config) and
+  else
-    accessPathCostLimits(apLimit, tupleLimit) and
+    exists(int aps, int nodes, int apLimit, int tupleLimit |
-    if apLimit < aps and tupleLimit < (aps - 1) * nodes then unfold = false else unfold = true
+      aps = countPotentialAps(apa, config) and
-  )
+      nodes = countNodesUsingAccessPath(apa, config) and
+      accessPathCostLimits(apLimit, tupleLimit) and
+      if apLimit < aps and tupleLimit < (aps - 1) * nodes then unfold = false else unfold = true
+    )
 }
 
 /**
@@ -3248,7 +3252,7 @@ class PathNode extends TPathNode {
    * The location spans column `startcolumn` of line `startline` to
    * column `endcolumn` of line `endline` in file `filepath`.
    * For more information, see
-   * [Locations](https://help.semmle.com/QL/learn-ql/ql/locations.html).
+   * [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
    */
   predicate hasLocationInfo(
     string filepath, int startline, int startcolumn, int endline, int endcolumn
@@ -4033,7 +4037,7 @@ private module FlowExploration {
      * The location spans column `startcolumn` of line `startline` to
      * column `endcolumn` of line `endline` in file `filepath`.
      * For more information, see
-     * [Locations](https://help.semmle.com/QL/learn-ql/ql/locations.html).
+     * [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
      */
     predicate hasLocationInfo(
       string filepath, int startline, int startcolumn, int endline, int endcolumn

cpp/ql/lib/semmle/code/cpp/dataflow/internal/DataFlowPrivate.qll

@@ -219,15 +219,13 @@ class DataFlowExpr = Expr;
 class DataFlowType = Type;
 
 /** A function call relevant for data flow. */
-class DataFlowCall extends Expr {
+class DataFlowCall extends Expr instanceof Call {
-  DataFlowCall() { this instanceof Call }
-
   /**
    * Gets the nth argument for this call.
    *
    * The range of `n` is from `0` to `getNumberOfArguments() - 1`.
    */
-  Expr getArgument(int n) { result = this.(Call).getArgument(n) }
+  Expr getArgument(int n) { result = super.getArgument(n) }
 
   /** Gets the data flow node corresponding to this call. */
   ExprNode getNode() { result.getExpr() = this }
@@ -240,6 +238,12 @@ predicate isUnreachableInCall(Node n, DataFlowCall call) { none() } // stub impl
 
 int accessPathLimit() { result = 5 }
 
+/**
+ * Holds if access paths with `c` at their head always should be tracked at high
+ * precision. This disables adaptive access path precision for such access paths.
+ */
+predicate forceHighPrecision(Content c) { none() }
+
 /** The unit type. */
 private newtype TUnit = TMkUnit()
 

cpp/ql/lib/semmle/code/cpp/dataflow/internal/DataFlowUtil.qll

@@ -101,7 +101,7 @@ class Node extends TNode {
    * The location spans column `startcolumn` of line `startline` to
    * column `endcolumn` of line `endline` in file `filepath`.
    * For more information, see
-   * [Locations](https://help.semmle.com/QL/learn-ql/ql/locations.html).
+   * [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
    */
   predicate hasLocationInfo(
     string filepath, int startline, int startcolumn, int endline, int endcolumn

cpp/ql/lib/semmle/code/cpp/dataflow/internal/tainttracking1/TaintTrackingImpl.qll

@@ -75,24 +75,26 @@ abstract class Configuration extends DataFlow::Configuration {
   predicate isSanitizer(DataFlow::Node node) { none() }
 
   final override predicate isBarrier(DataFlow::Node node) {
-    isSanitizer(node) or
+    this.isSanitizer(node) or
     defaultTaintSanitizer(node)
   }
 
   /** Holds if taint propagation into `node` is prohibited. */
   predicate isSanitizerIn(DataFlow::Node node) { none() }
 
-  final override predicate isBarrierIn(DataFlow::Node node) { isSanitizerIn(node) }
+  final override predicate isBarrierIn(DataFlow::Node node) { this.isSanitizerIn(node) }
 
   /** Holds if taint propagation out of `node` is prohibited. */
   predicate isSanitizerOut(DataFlow::Node node) { none() }
 
-  final override predicate isBarrierOut(DataFlow::Node node) { isSanitizerOut(node) }
+  final override predicate isBarrierOut(DataFlow::Node node) { this.isSanitizerOut(node) }
 
   /** Holds if taint propagation through nodes guarded by `guard` is prohibited. */
   predicate isSanitizerGuard(DataFlow::BarrierGuard guard) { none() }
 
-  final override predicate isBarrierGuard(DataFlow::BarrierGuard guard) { isSanitizerGuard(guard) }
+  final override predicate isBarrierGuard(DataFlow::BarrierGuard guard) {
+    this.isSanitizerGuard(guard)
+  }
 
   /**
    * Holds if the additional taint propagation step from `node1` to `node2`
@@ -101,7 +103,7 @@ abstract class Configuration extends DataFlow::Configuration {
   predicate isAdditionalTaintStep(DataFlow::Node node1, DataFlow::Node node2) { none() }
 
   final override predicate isAdditionalFlowStep(DataFlow::Node node1, DataFlow::Node node2) {
-    isAdditionalTaintStep(node1, node2) or
+    this.isAdditionalTaintStep(node1, node2) or
     defaultAdditionalTaintStep(node1, node2)
   }
 

cpp/ql/lib/semmle/code/cpp/dataflow/internal/tainttracking2/TaintTrackingImpl.qll

@@ -75,24 +75,26 @@ abstract class Configuration extends DataFlow::Configuration {
   predicate isSanitizer(DataFlow::Node node) { none() }
 
   final override predicate isBarrier(DataFlow::Node node) {
-    isSanitizer(node) or
+    this.isSanitizer(node) or
     defaultTaintSanitizer(node)
   }
 
   /** Holds if taint propagation into `node` is prohibited. */
   predicate isSanitizerIn(DataFlow::Node node) { none() }
 
-  final override predicate isBarrierIn(DataFlow::Node node) { isSanitizerIn(node) }
+  final override predicate isBarrierIn(DataFlow::Node node) { this.isSanitizerIn(node) }
 
   /** Holds if taint propagation out of `node` is prohibited. */
   predicate isSanitizerOut(DataFlow::Node node) { none() }
 
-  final override predicate isBarrierOut(DataFlow::Node node) { isSanitizerOut(node) }
+  final override predicate isBarrierOut(DataFlow::Node node) { this.isSanitizerOut(node) }
 
   /** Holds if taint propagation through nodes guarded by `guard` is prohibited. */
   predicate isSanitizerGuard(DataFlow::BarrierGuard guard) { none() }
 
-  final override predicate isBarrierGuard(DataFlow::BarrierGuard guard) { isSanitizerGuard(guard) }
+  final override predicate isBarrierGuard(DataFlow::BarrierGuard guard) {
+    this.isSanitizerGuard(guard)
+  }
 
   /**
    * Holds if the additional taint propagation step from `node1` to `node2`
@@ -101,7 +103,7 @@ abstract class Configuration extends DataFlow::Configuration {
   predicate isAdditionalTaintStep(DataFlow::Node node1, DataFlow::Node node2) { none() }
 
   final override predicate isAdditionalFlowStep(DataFlow::Node node1, DataFlow::Node node2) {
-    isAdditionalTaintStep(node1, node2) or
+    this.isAdditionalTaintStep(node1, node2) or
     defaultAdditionalTaintStep(node1, node2)
   }
 

cpp/ql/lib/semmle/code/cpp/ir/dataflow/DefaultTaintTracking.qll

@@ -4,7 +4,7 @@ private import semmle.code.cpp.ir.dataflow.DataFlow
 private import semmle.code.cpp.ir.dataflow.internal.DataFlowUtil
 private import semmle.code.cpp.ir.dataflow.DataFlow3
 private import semmle.code.cpp.ir.IR
-private import semmle.code.cpp.ir.dataflow.internal.DataFlowDispatch as Dispatch
+private import semmle.code.cpp.ir.dataflow.ResolveCall
 private import semmle.code.cpp.controlflow.IRGuards
 private import semmle.code.cpp.models.interfaces.Taint
 private import semmle.code.cpp.models.interfaces.DataFlow
@@ -355,20 +355,6 @@ predicate taintedIncludingGlobalVars(Expr source, Element tainted, string global
  */
 GlobalOrNamespaceVariable globalVarFromId(string id) { id = result.getQualifiedName() }
 
-/**
- * Resolve potential target function(s) for `call`.
- *
- * If `call` is a call through a function pointer (`ExprCall`) or
- * targets a virtual method, simple data flow analysis is performed
- * in order to identify target(s).
- */
-Function resolveCall(Call call) {
-  exists(CallInstruction callInstruction |
-    callInstruction.getAST() = call and
-    result = Dispatch::viableCallable(callInstruction)
-  )
-}
-
 /**
  * Provides definitions for augmenting source/sink pairs with data-flow paths
  * between them. From a `@kind path-problem` query, import this module in the
@@ -479,7 +465,7 @@ module TaintedWithPath {
      * The location spans column `startcolumn` of line `startline` to
      * column `endcolumn` of line `endline` in file `filepath`.
      * For more information, see
-     * [Locations](https://help.semmle.com/QL/learn-ql/ql/locations.html).
+     * [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
      */
     predicate hasLocationInfo(
       string filepath, int startline, int startcolumn, int endline, int endcolumn

cpp/ql/lib/semmle/code/cpp/ir/dataflow/ResolveCall.qll

@@ -0,0 +1,23 @@
+/**
+ * Provides a predicate for non-contextual virtual dispatch and function
+ * pointer resolution.
+ */
+
+import cpp
+private import semmle.code.cpp.ir.ValueNumbering
+private import internal.DataFlowDispatch
+private import semmle.code.cpp.ir.IR
+
+/**
+ * Resolve potential target function(s) for `call`.
+ *
+ * If `call` is a call through a function pointer (`ExprCall`) or its target is
+ * a virtual member function, simple data flow analysis is performed in order
+ * to identify the possible target(s).
+ */
+Function resolveCall(Call call) {
+  exists(CallInstruction callInstruction |
+    callInstruction.getAST() = call and
+    result = viableCallable(callInstruction)
+  )
+}

cpp/ql/lib/semmle/code/cpp/ir/dataflow/internal/DataFlowImpl.qll

@@ -2139,7 +2139,8 @@ private predicate expensiveLen2unfolding(TypedContent tc, Configuration config)
       ) and
     accessPathApproxCostLimits(apLimit, tupleLimit) and
     apLimit < tails and
-    tupleLimit < (tails - 1) * nodes
+    tupleLimit < (tails - 1) * nodes and
+    not tc.forceHighPrecision()
   )
 }
 
@@ -2973,12 +2974,15 @@ private AccessPathApprox getATail(AccessPathApprox apa, Configuration config) {
  * expected to be expensive. Holds with `unfold = true` otherwise.
  */
 private predicate evalUnfold(AccessPathApprox apa, boolean unfold, Configuration config) {
-  exists(int aps, int nodes, int apLimit, int tupleLimit |
+  if apa.getHead().forceHighPrecision()
-    aps = countPotentialAps(apa, config) and
+  then unfold = true
-    nodes = countNodesUsingAccessPath(apa, config) and
+  else
-    accessPathCostLimits(apLimit, tupleLimit) and
+    exists(int aps, int nodes, int apLimit, int tupleLimit |
-    if apLimit < aps and tupleLimit < (aps - 1) * nodes then unfold = false else unfold = true
+      aps = countPotentialAps(apa, config) and
-  )
+      nodes = countNodesUsingAccessPath(apa, config) and
+      accessPathCostLimits(apLimit, tupleLimit) and
+      if apLimit < aps and tupleLimit < (aps - 1) * nodes then unfold = false else unfold = true
+    )
 }
 
 /**
@@ -3248,7 +3252,7 @@ class PathNode extends TPathNode {
    * The location spans column `startcolumn` of line `startline` to
    * column `endcolumn` of line `endline` in file `filepath`.
    * For more information, see
-   * [Locations](https://help.semmle.com/QL/learn-ql/ql/locations.html).
+   * [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
    */
   predicate hasLocationInfo(
     string filepath, int startline, int startcolumn, int endline, int endcolumn
@@ -4033,7 +4037,7 @@ private module FlowExploration {
      * The location spans column `startcolumn` of line `startline` to
      * column `endcolumn` of line `endline` in file `filepath`.
      * For more information, see
-     * [Locations](https://help.semmle.com/QL/learn-ql/ql/locations.html).
+     * [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
      */
     predicate hasLocationInfo(
       string filepath, int startline, int startcolumn, int endline, int endcolumn

cpp/ql/lib/semmle/code/cpp/ir/dataflow/internal/DataFlowImpl2.qll

@@ -2139,7 +2139,8 @@ private predicate expensiveLen2unfolding(TypedContent tc, Configuration config)
       ) and
     accessPathApproxCostLimits(apLimit, tupleLimit) and
     apLimit < tails and
-    tupleLimit < (tails - 1) * nodes
+    tupleLimit < (tails - 1) * nodes and
+    not tc.forceHighPrecision()
   )
 }
 
@@ -2973,12 +2974,15 @@ private AccessPathApprox getATail(AccessPathApprox apa, Configuration config) {
  * expected to be expensive. Holds with `unfold = true` otherwise.
  */
 private predicate evalUnfold(AccessPathApprox apa, boolean unfold, Configuration config) {
-  exists(int aps, int nodes, int apLimit, int tupleLimit |
+  if apa.getHead().forceHighPrecision()
-    aps = countPotentialAps(apa, config) and
+  then unfold = true
-    nodes = countNodesUsingAccessPath(apa, config) and
+  else
-    accessPathCostLimits(apLimit, tupleLimit) and
+    exists(int aps, int nodes, int apLimit, int tupleLimit |
-    if apLimit < aps and tupleLimit < (aps - 1) * nodes then unfold = false else unfold = true
+      aps = countPotentialAps(apa, config) and
-  )
+      nodes = countNodesUsingAccessPath(apa, config) and
+      accessPathCostLimits(apLimit, tupleLimit) and
+      if apLimit < aps and tupleLimit < (aps - 1) * nodes then unfold = false else unfold = true
+    )
 }
 
 /**
@@ -3248,7 +3252,7 @@ class PathNode extends TPathNode {
    * The location spans column `startcolumn` of line `startline` to
    * column `endcolumn` of line `endline` in file `filepath`.
    * For more information, see
-   * [Locations](https://help.semmle.com/QL/learn-ql/ql/locations.html).
+   * [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
    */
   predicate hasLocationInfo(
     string filepath, int startline, int startcolumn, int endline, int endcolumn
@@ -4033,7 +4037,7 @@ private module FlowExploration {
      * The location spans column `startcolumn` of line `startline` to
      * column `endcolumn` of line `endline` in file `filepath`.
      * For more information, see
-     * [Locations](https://help.semmle.com/QL/learn-ql/ql/locations.html).
+     * [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
      */
     predicate hasLocationInfo(
       string filepath, int startline, int startcolumn, int endline, int endcolumn

cpp/ql/lib/semmle/code/cpp/ir/dataflow/internal/DataFlowImpl3.qll

@@ -2139,7 +2139,8 @@ private predicate expensiveLen2unfolding(TypedContent tc, Configuration config)
       ) and
     accessPathApproxCostLimits(apLimit, tupleLimit) and
     apLimit < tails and
-    tupleLimit < (tails - 1) * nodes
+    tupleLimit < (tails - 1) * nodes and
+    not tc.forceHighPrecision()
   )
 }
 
@@ -2973,12 +2974,15 @@ private AccessPathApprox getATail(AccessPathApprox apa, Configuration config) {
  * expected to be expensive. Holds with `unfold = true` otherwise.
  */
 private predicate evalUnfold(AccessPathApprox apa, boolean unfold, Configuration config) {
-  exists(int aps, int nodes, int apLimit, int tupleLimit |
+  if apa.getHead().forceHighPrecision()
-    aps = countPotentialAps(apa, config) and
+  then unfold = true
-    nodes = countNodesUsingAccessPath(apa, config) and
+  else
-    accessPathCostLimits(apLimit, tupleLimit) and
+    exists(int aps, int nodes, int apLimit, int tupleLimit |
-    if apLimit < aps and tupleLimit < (aps - 1) * nodes then unfold = false else unfold = true
+      aps = countPotentialAps(apa, config) and
-  )
+      nodes = countNodesUsingAccessPath(apa, config) and
+      accessPathCostLimits(apLimit, tupleLimit) and
+      if apLimit < aps and tupleLimit < (aps - 1) * nodes then unfold = false else unfold = true
+    )
 }
 
 /**
@@ -3248,7 +3252,7 @@ class PathNode extends TPathNode {
    * The location spans column `startcolumn` of line `startline` to
    * column `endcolumn` of line `endline` in file `filepath`.
    * For more information, see
-   * [Locations](https://help.semmle.com/QL/learn-ql/ql/locations.html).
+   * [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
    */
   predicate hasLocationInfo(
     string filepath, int startline, int startcolumn, int endline, int endcolumn
@@ -4033,7 +4037,7 @@ private module FlowExploration {
      * The location spans column `startcolumn` of line `startline` to
      * column `endcolumn` of line `endline` in file `filepath`.
      * For more information, see
-     * [Locations](https://help.semmle.com/QL/learn-ql/ql/locations.html).
+     * [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
      */
     predicate hasLocationInfo(
       string filepath, int startline, int startcolumn, int endline, int endcolumn

cpp/ql/lib/semmle/code/cpp/ir/dataflow/internal/DataFlowImpl4.qll

@@ -2139,7 +2139,8 @@ private predicate expensiveLen2unfolding(TypedContent tc, Configuration config)
       ) and
     accessPathApproxCostLimits(apLimit, tupleLimit) and
     apLimit < tails and
-    tupleLimit < (tails - 1) * nodes
+    tupleLimit < (tails - 1) * nodes and
+    not tc.forceHighPrecision()
   )
 }
 
@@ -2973,12 +2974,15 @@ private AccessPathApprox getATail(AccessPathApprox apa, Configuration config) {
  * expected to be expensive. Holds with `unfold = true` otherwise.
  */
 private predicate evalUnfold(AccessPathApprox apa, boolean unfold, Configuration config) {
-  exists(int aps, int nodes, int apLimit, int tupleLimit |
+  if apa.getHead().forceHighPrecision()
-    aps = countPotentialAps(apa, config) and
+  then unfold = true
-    nodes = countNodesUsingAccessPath(apa, config) and
+  else
-    accessPathCostLimits(apLimit, tupleLimit) and
+    exists(int aps, int nodes, int apLimit, int tupleLimit |
-    if apLimit < aps and tupleLimit < (aps - 1) * nodes then unfold = false else unfold = true
+      aps = countPotentialAps(apa, config) and
-  )
+      nodes = countNodesUsingAccessPath(apa, config) and
+      accessPathCostLimits(apLimit, tupleLimit) and
+      if apLimit < aps and tupleLimit < (aps - 1) * nodes then unfold = false else unfold = true
+    )
 }
 
 /**
@@ -3248,7 +3252,7 @@ class PathNode extends TPathNode {
    * The location spans column `startcolumn` of line `startline` to
    * column `endcolumn` of line `endline` in file `filepath`.
    * For more information, see
-   * [Locations](https://help.semmle.com/QL/learn-ql/ql/locations.html).
+   * [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
    */
   predicate hasLocationInfo(
     string filepath, int startline, int startcolumn, int endline, int endcolumn
@@ -4033,7 +4037,7 @@ private module FlowExploration {
      * The location spans column `startcolumn` of line `startline` to
      * column `endcolumn` of line `endline` in file `filepath`.
      * For more information, see
-     * [Locations](https://help.semmle.com/QL/learn-ql/ql/locations.html).
+     * [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
      */
     predicate hasLocationInfo(
       string filepath, int startline, int startcolumn, int endline, int endcolumn

cpp/ql/lib/semmle/code/cpp/ir/dataflow/internal/DataFlowImplCommon.qll

@@ -937,7 +937,7 @@ class CallContextSpecificCall extends CallContextCall, TSpecificCall {
   }
 
   override predicate relevantFor(DataFlowCallable callable) {
-    recordDataFlowCallSite(getCall(), callable)
+    recordDataFlowCallSite(this.getCall(), callable)
   }
 
   override predicate matchesCall(DataFlowCall call) { call = this.getCall() }
@@ -1236,6 +1236,13 @@ class TypedContent extends MkTypedContent {
 
   /** Gets a textual representation of this content. */
   string toString() { result = c.toString() }
+
+  /**
+   * Holds if access paths with this `TypedContent` at their head always should
+   * be tracked at high precision. This disables adaptive access path precision
+   * for such access paths.
+   */
+  predicate forceHighPrecision() { forceHighPrecision(c) }
 }
 
 /**
@@ -1250,7 +1257,7 @@ abstract class AccessPathFront extends TAccessPathFront {
 
   TypedContent getHead() { this = TFrontHead(result) }
 
-  predicate isClearedAt(Node n) { clearsContentCached(n, getHead().getContent()) }
+  predicate isClearedAt(Node n) { clearsContentCached(n, this.getHead().getContent()) }
 }
 
 class AccessPathFrontNil extends AccessPathFront, TFrontNil {

cpp/ql/lib/semmle/code/cpp/ir/dataflow/internal/DataFlowPrivate.qll

@@ -466,6 +466,12 @@ predicate isUnreachableInCall(Node n, DataFlowCall call) { none() } // stub impl
 
 int accessPathLimit() { result = 5 }
 
+/**
+ * Holds if access paths with `c` at their head always should be tracked at high
+ * precision. This disables adaptive access path precision for such access paths.
+ */
+predicate forceHighPrecision(Content c) { none() }
+
 /** The unit type. */
 private newtype TUnit = TMkUnit()
 

cpp/ql/lib/semmle/code/cpp/ir/dataflow/internal/DataFlowUtil.qll

@@ -120,7 +120,7 @@ class Node extends TIRDataFlowNode {
    * The location spans column `startcolumn` of line `startline` to
    * column `endcolumn` of line `endline` in file `filepath`.
    * For more information, see
-   * [Locations](https://help.semmle.com/QL/learn-ql/ql/locations.html).
+   * [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
    */
   predicate hasLocationInfo(
     string filepath, int startline, int startcolumn, int endline, int endcolumn

cpp/ql/lib/semmle/code/cpp/ir/dataflow/internal/ModelUtil.qll

@@ -38,5 +38,8 @@ Instruction callOutput(CallInstruction call, FunctionOutput output) {
     effect.getPrimaryInstruction() = call and
     output.isParameterDerefOrQualifierObject(effect.getIndex())
   )
-  // TODO: return value dereference
+  or
+  // TODO: modify this when we get return value dereferences
+  result = call and
+  output.isReturnValueDeref()
 }

cpp/ql/lib/semmle/code/cpp/ir/dataflow/internal/tainttracking1/TaintTrackingImpl.qll

@@ -75,24 +75,26 @@ abstract class Configuration extends DataFlow::Configuration {
   predicate isSanitizer(DataFlow::Node node) { none() }
 
   final override predicate isBarrier(DataFlow::Node node) {
-    isSanitizer(node) or
+    this.isSanitizer(node) or
     defaultTaintSanitizer(node)
   }
 
   /** Holds if taint propagation into `node` is prohibited. */
   predicate isSanitizerIn(DataFlow::Node node) { none() }
 
-  final override predicate isBarrierIn(DataFlow::Node node) { isSanitizerIn(node) }
+  final override predicate isBarrierIn(DataFlow::Node node) { this.isSanitizerIn(node) }
 
   /** Holds if taint propagation out of `node` is prohibited. */
   predicate isSanitizerOut(DataFlow::Node node) { none() }
 
-  final override predicate isBarrierOut(DataFlow::Node node) { isSanitizerOut(node) }
+  final override predicate isBarrierOut(DataFlow::Node node) { this.isSanitizerOut(node) }
 
   /** Holds if taint propagation through nodes guarded by `guard` is prohibited. */
   predicate isSanitizerGuard(DataFlow::BarrierGuard guard) { none() }
 
-  final override predicate isBarrierGuard(DataFlow::BarrierGuard guard) { isSanitizerGuard(guard) }
+  final override predicate isBarrierGuard(DataFlow::BarrierGuard guard) {
+    this.isSanitizerGuard(guard)
+  }
 
   /**
    * Holds if the additional taint propagation step from `node1` to `node2`
@@ -101,7 +103,7 @@ abstract class Configuration extends DataFlow::Configuration {
   predicate isAdditionalTaintStep(DataFlow::Node node1, DataFlow::Node node2) { none() }
 
   final override predicate isAdditionalFlowStep(DataFlow::Node node1, DataFlow::Node node2) {
-    isAdditionalTaintStep(node1, node2) or
+    this.isAdditionalTaintStep(node1, node2) or
     defaultAdditionalTaintStep(node1, node2)
   }
 

cpp/ql/lib/semmle/code/cpp/ir/dataflow/internal/tainttracking2/TaintTrackingImpl.qll

@@ -75,24 +75,26 @@ abstract class Configuration extends DataFlow::Configuration {
   predicate isSanitizer(DataFlow::Node node) { none() }
 
   final override predicate isBarrier(DataFlow::Node node) {
-    isSanitizer(node) or
+    this.isSanitizer(node) or
     defaultTaintSanitizer(node)
   }
 
   /** Holds if taint propagation into `node` is prohibited. */
   predicate isSanitizerIn(DataFlow::Node node) { none() }
 
-  final override predicate isBarrierIn(DataFlow::Node node) { isSanitizerIn(node) }
+  final override predicate isBarrierIn(DataFlow::Node node) { this.isSanitizerIn(node) }
 
   /** Holds if taint propagation out of `node` is prohibited. */
   predicate isSanitizerOut(DataFlow::Node node) { none() }
 
-  final override predicate isBarrierOut(DataFlow::Node node) { isSanitizerOut(node) }
+  final override predicate isBarrierOut(DataFlow::Node node) { this.isSanitizerOut(node) }
 
   /** Holds if taint propagation through nodes guarded by `guard` is prohibited. */
   predicate isSanitizerGuard(DataFlow::BarrierGuard guard) { none() }
 
-  final override predicate isBarrierGuard(DataFlow::BarrierGuard guard) { isSanitizerGuard(guard) }
+  final override predicate isBarrierGuard(DataFlow::BarrierGuard guard) {
+    this.isSanitizerGuard(guard)
+  }
 
   /**
    * Holds if the additional taint propagation step from `node1` to `node2`
@@ -101,7 +103,7 @@ abstract class Configuration extends DataFlow::Configuration {
   predicate isAdditionalTaintStep(DataFlow::Node node1, DataFlow::Node node2) { none() }
 
   final override predicate isAdditionalFlowStep(DataFlow::Node node1, DataFlow::Node node2) {
-    isAdditionalTaintStep(node1, node2) or
+    this.isAdditionalTaintStep(node1, node2) or
     defaultAdditionalTaintStep(node1, node2)
   }
 

cpp/ql/lib/semmle/code/cpp/ir/dataflow/internal/tainttracking3/TaintTrackingImpl.qll

@@ -75,24 +75,26 @@ abstract class Configuration extends DataFlow::Configuration {
   predicate isSanitizer(DataFlow::Node node) { none() }
 
   final override predicate isBarrier(DataFlow::Node node) {
-    isSanitizer(node) or
+    this.isSanitizer(node) or
     defaultTaintSanitizer(node)
   }
 
   /** Holds if taint propagation into `node` is prohibited. */
   predicate isSanitizerIn(DataFlow::Node node) { none() }
 
-  final override predicate isBarrierIn(DataFlow::Node node) { isSanitizerIn(node) }
+  final override predicate isBarrierIn(DataFlow::Node node) { this.isSanitizerIn(node) }
 
   /** Holds if taint propagation out of `node` is prohibited. */
   predicate isSanitizerOut(DataFlow::Node node) { none() }
 
-  final override predicate isBarrierOut(DataFlow::Node node) { isSanitizerOut(node) }
+  final override predicate isBarrierOut(DataFlow::Node node) { this.isSanitizerOut(node) }
 
   /** Holds if taint propagation through nodes guarded by `guard` is prohibited. */
   predicate isSanitizerGuard(DataFlow::BarrierGuard guard) { none() }
 
-  final override predicate isBarrierGuard(DataFlow::BarrierGuard guard) { isSanitizerGuard(guard) }
+  final override predicate isBarrierGuard(DataFlow::BarrierGuard guard) {
+    this.isSanitizerGuard(guard)
+  }
 
   /**
    * Holds if the additional taint propagation step from `node1` to `node2`
@@ -101,7 +103,7 @@ abstract class Configuration extends DataFlow::Configuration {
   predicate isAdditionalTaintStep(DataFlow::Node node1, DataFlow::Node node2) { none() }
 
   final override predicate isAdditionalFlowStep(DataFlow::Node node1, DataFlow::Node node2) {
-    isAdditionalTaintStep(node1, node2) or
+    this.isAdditionalTaintStep(node1, node2) or
     defaultAdditionalTaintStep(node1, node2)
   }
 

cpp/ql/lib/semmle/code/cpp/ir/implementation/aliased_ssa/IRBlock.qll

@@ -24,7 +24,7 @@ class IRBlockBase extends TIRBlock {
   final string toString() { result = getFirstInstruction(this).toString() }
 
   /** Gets the source location of the first non-`Phi` instruction in this block. */
-  final Language::Location getLocation() { result = getFirstInstruction().getLocation() }
+  final Language::Location getLocation() { result = this.getFirstInstruction().getLocation() }
 
   /**
    * INTERNAL: Do not use.
@@ -39,7 +39,7 @@ class IRBlockBase extends TIRBlock {
     ) and
     this =
       rank[result + 1](IRBlock funcBlock, int sortOverride, int sortKey1, int sortKey2 |
-        funcBlock.getEnclosingFunction() = getEnclosingFunction() and
+        funcBlock.getEnclosingFunction() = this.getEnclosingFunction() and
         funcBlock.getFirstInstruction().hasSortKeys(sortKey1, sortKey2) and
         // Ensure that the block containing `EnterFunction` always comes first.
         if funcBlock.getFirstInstruction() instanceof EnterFunctionInstruction
@@ -59,15 +59,15 @@ class IRBlockBase extends TIRBlock {
    * Get the `Phi` instructions that appear at the start of this block.
    */
   final PhiInstruction getAPhiInstruction() {
-    Construction::getPhiInstructionBlockStart(result) = getFirstInstruction()
+    Construction::getPhiInstructionBlockStart(result) = this.getFirstInstruction()
   }
 
   /**
    * Gets an instruction in this block. This includes `Phi` instructions.
    */
   final Instruction getAnInstruction() {
-    result = getInstruction(_) or
+    result = this.getInstruction(_) or
-    result = getAPhiInstruction()
+    result = this.getAPhiInstruction()
   }
 
   /**
@@ -78,7 +78,9 @@ class IRBlockBase extends TIRBlock {
   /**
    * Gets the last instruction in this block.
    */
-  final Instruction getLastInstruction() { result = getInstruction(getInstructionCount() - 1) }
+  final Instruction getLastInstruction() {
+    result = this.getInstruction(this.getInstructionCount() - 1)
+  }
 
   /**
    * Gets the number of non-`Phi` instructions in this block.
@@ -149,7 +151,7 @@ class IRBlock extends IRBlockBase {
    * Block `A` dominates block `B` if any control flow path from the entry block of the function to
    * block `B` must pass through block `A`. A block always dominates itself.
    */
-  final predicate dominates(IRBlock block) { strictlyDominates(block) or this = block }
+  final predicate dominates(IRBlock block) { this.strictlyDominates(block) or this = block }
 
   /**
    * Gets a block on the dominance frontier of this block.
@@ -159,8 +161,8 @@ class IRBlock extends IRBlockBase {
    */
   pragma[noinline]
   final IRBlock dominanceFrontier() {
-    dominates(result.getAPredecessor()) and
+    this.dominates(result.getAPredecessor()) and
-    not strictlyDominates(result)
+    not this.strictlyDominates(result)
   }
 
   /**
@@ -189,7 +191,7 @@ class IRBlock extends IRBlockBase {
    * Block `A` post-dominates block `B` if any control flow path from `B` to the exit block of the
    * function must pass through block `A`. A block always post-dominates itself.
    */
-  final predicate postDominates(IRBlock block) { strictlyPostDominates(block) or this = block }
+  final predicate postDominates(IRBlock block) { this.strictlyPostDominates(block) or this = block }
 
   /**
    * Gets a block on the post-dominance frontier of this block.
@@ -199,16 +201,16 @@ class IRBlock extends IRBlockBase {
    */
   pragma[noinline]
   final IRBlock postPominanceFrontier() {
-    postDominates(result.getASuccessor()) and
+    this.postDominates(result.getASuccessor()) and
-    not strictlyPostDominates(result)
+    not this.strictlyPostDominates(result)
   }
 
   /**
    * Holds if this block is reachable from the entry block of its function.
    */
   final predicate isReachableFromFunctionEntry() {
-    this = getEnclosingIRFunction().getEntryBlock() or
+    this = this.getEnclosingIRFunction().getEntryBlock() or
-    getAPredecessor().isReachableFromFunctionEntry()
+    this.getAPredecessor().isReachableFromFunctionEntry()
   }
 }
 

cpp/ql/lib/semmle/code/cpp/ir/implementation/aliased_ssa/Operand.qll

@@ -46,12 +46,12 @@ class Operand extends TStageOperand {
   /**
    * Gets the location of the source code for this operand.
    */
-  final Language::Location getLocation() { result = getUse().getLocation() }
+  final Language::Location getLocation() { result = this.getUse().getLocation() }
 
   /**
    * Gets the function that contains this operand.
    */
-  final IRFunction getEnclosingIRFunction() { result = getUse().getEnclosingIRFunction() }
+  final IRFunction getEnclosingIRFunction() { result = this.getUse().getEnclosingIRFunction() }
 
   /**
    * Gets the `Instruction` that consumes this operand.
@@ -74,7 +74,7 @@ class Operand extends TStageOperand {
    */
   final Instruction getDef() {
     result = this.getAnyDef() and
-    getDefinitionOverlap() instanceof MustExactlyOverlap
+    this.getDefinitionOverlap() instanceof MustExactlyOverlap
   }
 
   /**
@@ -82,7 +82,7 @@ class Operand extends TStageOperand {
    *
    * Gets the `Instruction` that consumes this operand.
    */
-  deprecated final Instruction getUseInstruction() { result = getUse() }
+  deprecated final Instruction getUseInstruction() { result = this.getUse() }
 
   /**
    * DEPRECATED: use `getAnyDef` or `getDef`. The exact replacement for this
@@ -91,7 +91,7 @@ class Operand extends TStageOperand {
    *
    * Gets the `Instruction` whose result is the value of the operand.
    */
-  deprecated final Instruction getDefinitionInstruction() { result = getAnyDef() }
+  deprecated final Instruction getDefinitionInstruction() { result = this.getAnyDef() }
 
   /**
    * Gets the overlap relationship between the operand's definition and its use.
@@ -101,7 +101,9 @@ class Operand extends TStageOperand {
   /**
    * Holds if the result of the definition instruction does not exactly overlap this use.
    */
-  final predicate isDefinitionInexact() { not getDefinitionOverlap() instanceof MustExactlyOverlap }
+  final predicate isDefinitionInexact() {
+    not this.getDefinitionOverlap() instanceof MustExactlyOverlap
+  }
 
   /**
    * Gets a prefix to use when dumping the operand in an operand list.
@@ -121,7 +123,7 @@ class Operand extends TStageOperand {
    * For example: `this:r3_5`
    */
   final string getDumpString() {
-    result = getDumpLabel() + getInexactSpecifier() + getDefinitionId()
+    result = this.getDumpLabel() + this.getInexactSpecifier() + this.getDefinitionId()
   }
 
   /**
@@ -129,9 +131,9 @@ class Operand extends TStageOperand {
    * definition is not modeled in SSA.
    */
   private string getDefinitionId() {
-    result = getAnyDef().getResultId()
+    result = this.getAnyDef().getResultId()
     or
-    not exists(getAnyDef()) and result = "m?"
+    not exists(this.getAnyDef()) and result = "m?"
   }
 
   /**
@@ -140,7 +142,7 @@ class Operand extends TStageOperand {
    * the empty string.
    */
   private string getInexactSpecifier() {
-    if isDefinitionInexact() then result = "~" else result = ""
+    if this.isDefinitionInexact() then result = "~" else result = ""
   }
 
   /**
@@ -155,7 +157,7 @@ class Operand extends TStageOperand {
    * the definition type, such as in the case of a partial read or a read from a pointer that
    * has been cast to a different type.
    */
-  Language::LanguageType getLanguageType() { result = getAnyDef().getResultLanguageType() }
+  Language::LanguageType getLanguageType() { result = this.getAnyDef().getResultLanguageType() }
 
   /**
    * Gets the language-neutral type of the value consumed by this operand. This is usually the same
@@ -164,7 +166,7 @@ class Operand extends TStageOperand {
    * from the definition type, such as in the case of a partial read or a read from a pointer that
    * has been cast to a different type.
    */
-  final IRType getIRType() { result = getLanguageType().getIRType() }
+  final IRType getIRType() { result = this.getLanguageType().getIRType() }
 
   /**
    * Gets the type of the value consumed by this operand. This is usually the same as the
@@ -173,7 +175,7 @@ class Operand extends TStageOperand {
    * the definition type, such as in the case of a partial read or a read from a pointer that
    * has been cast to a different type.
    */
-  final Language::Type getType() { getLanguageType().hasType(result, _) }
+  final Language::Type getType() { this.getLanguageType().hasType(result, _) }
 
   /**
    * Holds if the value consumed by this operand is a glvalue. If this
@@ -182,13 +184,13 @@ class Operand extends TStageOperand {
    * not hold, the value of the operand represents a value whose type is
    * given by `getType()`.
    */
-  final predicate isGLValue() { getLanguageType().hasType(_, true) }
+  final predicate isGLValue() { this.getLanguageType().hasType(_, true) }
 
   /**
    * Gets the size of the value consumed by this operand, in bytes. If the operand does not have
    * a known constant size, this predicate does not hold.
    */
-  final int getSize() { result = getLanguageType().getByteSize() }
+  final int getSize() { result = this.getLanguageType().getByteSize() }
 }
 
 /**
@@ -205,7 +207,7 @@ class MemoryOperand extends Operand {
   /**
    * Gets the kind of memory access performed by the operand.
    */
-  MemoryAccessKind getMemoryAccess() { result = getUse().getOpcode().getReadMemoryAccess() }
+  MemoryAccessKind getMemoryAccess() { result = this.getUse().getOpcode().getReadMemoryAccess() }
 
   /**
    * Holds if the memory access performed by this operand will not always read from every bit in the
@@ -215,7 +217,7 @@ class MemoryOperand extends Operand {
    * conservative estimate of the memory that might actually be accessed at runtime (for example,
    * the global side effects of a function call).
    */
-  predicate hasMayReadMemoryAccess() { getUse().getOpcode().hasMayReadMemoryAccess() }
+  predicate hasMayReadMemoryAccess() { this.getUse().getOpcode().hasMayReadMemoryAccess() }
 
   /**
    * Returns the operand that holds the memory address from which the current operand loads its
@@ -223,8 +225,8 @@ class MemoryOperand extends Operand {
    * is `r1`.
    */
   final AddressOperand getAddressOperand() {
-    getMemoryAccess().usesAddressOperand() and
+    this.getMemoryAccess().usesAddressOperand() and
-    result.getUse() = getUse()
+    result.getUse() = this.getUse()
   }
 }
 
@@ -294,7 +296,7 @@ class NonPhiMemoryOperand extends NonPhiOperand, MemoryOperand, TNonPhiMemoryOpe
     result = unique(Instruction defInstr | hasDefinition(defInstr, _))
   }
 
-  final override Overlap getDefinitionOverlap() { hasDefinition(_, result) }
+  final override Overlap getDefinitionOverlap() { this.hasDefinition(_, result) }
 
   pragma[noinline]
   private predicate hasDefinition(Instruction defInstr, Overlap overlap) {
@@ -449,13 +451,17 @@ class PhiInputOperand extends MemoryOperand, TPhiOperand {
 
   final override Overlap getDefinitionOverlap() { result = overlap }
 
-  final override int getDumpSortOrder() { result = 11 + getPredecessorBlock().getDisplayIndex() }
+  final override int getDumpSortOrder() {
-
+    result = 11 + this.getPredecessorBlock().getDisplayIndex()
-  final override string getDumpLabel() {
-    result = "from " + getPredecessorBlock().getDisplayIndex().toString() + ":"
   }
 
-  final override string getDumpId() { result = getPredecessorBlock().getDisplayIndex().toString() }
+  final override string getDumpLabel() {
+    result = "from " + this.getPredecessorBlock().getDisplayIndex().toString() + ":"
+  }
+
+  final override string getDumpId() {
+    result = this.getPredecessorBlock().getDisplayIndex().toString()
+  }
 
   /**
    * Gets the predecessor block from which this value comes.

cpp/ql/lib/semmle/code/cpp/ir/implementation/raw/IRBlock.qll

@@ -24,7 +24,7 @@ class IRBlockBase extends TIRBlock {
   final string toString() { result = getFirstInstruction(this).toString() }
 
   /** Gets the source location of the first non-`Phi` instruction in this block. */
-  final Language::Location getLocation() { result = getFirstInstruction().getLocation() }
+  final Language::Location getLocation() { result = this.getFirstInstruction().getLocation() }
 
   /**
    * INTERNAL: Do not use.
@@ -39,7 +39,7 @@ class IRBlockBase extends TIRBlock {
     ) and
     this =
       rank[result + 1](IRBlock funcBlock, int sortOverride, int sortKey1, int sortKey2 |
-        funcBlock.getEnclosingFunction() = getEnclosingFunction() and
+        funcBlock.getEnclosingFunction() = this.getEnclosingFunction() and
         funcBlock.getFirstInstruction().hasSortKeys(sortKey1, sortKey2) and
         // Ensure that the block containing `EnterFunction` always comes first.
         if funcBlock.getFirstInstruction() instanceof EnterFunctionInstruction
@@ -59,15 +59,15 @@ class IRBlockBase extends TIRBlock {
    * Get the `Phi` instructions that appear at the start of this block.
    */
   final PhiInstruction getAPhiInstruction() {
-    Construction::getPhiInstructionBlockStart(result) = getFirstInstruction()
+    Construction::getPhiInstructionBlockStart(result) = this.getFirstInstruction()
   }
 
   /**
    * Gets an instruction in this block. This includes `Phi` instructions.
    */
   final Instruction getAnInstruction() {
-    result = getInstruction(_) or
+    result = this.getInstruction(_) or
-    result = getAPhiInstruction()
+    result = this.getAPhiInstruction()
   }
 
   /**
@@ -78,7 +78,9 @@ class IRBlockBase extends TIRBlock {
   /**
    * Gets the last instruction in this block.
    */
-  final Instruction getLastInstruction() { result = getInstruction(getInstructionCount() - 1) }
+  final Instruction getLastInstruction() {
+    result = this.getInstruction(this.getInstructionCount() - 1)
+  }
 
   /**
    * Gets the number of non-`Phi` instructions in this block.
@@ -149,7 +151,7 @@ class IRBlock extends IRBlockBase {
    * Block `A` dominates block `B` if any control flow path from the entry block of the function to
    * block `B` must pass through block `A`. A block always dominates itself.
    */
-  final predicate dominates(IRBlock block) { strictlyDominates(block) or this = block }
+  final predicate dominates(IRBlock block) { this.strictlyDominates(block) or this = block }
 
   /**
    * Gets a block on the dominance frontier of this block.
@@ -159,8 +161,8 @@ class IRBlock extends IRBlockBase {
    */
   pragma[noinline]
   final IRBlock dominanceFrontier() {
-    dominates(result.getAPredecessor()) and
+    this.dominates(result.getAPredecessor()) and
-    not strictlyDominates(result)
+    not this.strictlyDominates(result)
   }
 
   /**
@@ -189,7 +191,7 @@ class IRBlock extends IRBlockBase {
    * Block `A` post-dominates block `B` if any control flow path from `B` to the exit block of the
    * function must pass through block `A`. A block always post-dominates itself.
    */
-  final predicate postDominates(IRBlock block) { strictlyPostDominates(block) or this = block }
+  final predicate postDominates(IRBlock block) { this.strictlyPostDominates(block) or this = block }
 
   /**
    * Gets a block on the post-dominance frontier of this block.
@@ -199,16 +201,16 @@ class IRBlock extends IRBlockBase {
    */
   pragma[noinline]
   final IRBlock postPominanceFrontier() {
-    postDominates(result.getASuccessor()) and
+    this.postDominates(result.getASuccessor()) and
-    not strictlyPostDominates(result)
+    not this.strictlyPostDominates(result)
   }
 
   /**
    * Holds if this block is reachable from the entry block of its function.
    */
   final predicate isReachableFromFunctionEntry() {
-    this = getEnclosingIRFunction().getEntryBlock() or
+    this = this.getEnclosingIRFunction().getEntryBlock() or
-    getAPredecessor().isReachableFromFunctionEntry()
+    this.getAPredecessor().isReachableFromFunctionEntry()
   }
 }
 

cpp/ql/lib/semmle/code/cpp/ir/implementation/raw/Operand.qll

@@ -46,12 +46,12 @@ class Operand extends TStageOperand {
   /**
    * Gets the location of the source code for this operand.
    */
-  final Language::Location getLocation() { result = getUse().getLocation() }
+  final Language::Location getLocation() { result = this.getUse().getLocation() }
 
   /**
    * Gets the function that contains this operand.
    */
-  final IRFunction getEnclosingIRFunction() { result = getUse().getEnclosingIRFunction() }
+  final IRFunction getEnclosingIRFunction() { result = this.getUse().getEnclosingIRFunction() }
 
   /**
    * Gets the `Instruction` that consumes this operand.
@@ -74,7 +74,7 @@ class Operand extends TStageOperand {
    */
   final Instruction getDef() {
     result = this.getAnyDef() and
-    getDefinitionOverlap() instanceof MustExactlyOverlap
+    this.getDefinitionOverlap() instanceof MustExactlyOverlap
   }
 
   /**
@@ -82,7 +82,7 @@ class Operand extends TStageOperand {
    *
    * Gets the `Instruction` that consumes this operand.
    */
-  deprecated final Instruction getUseInstruction() { result = getUse() }
+  deprecated final Instruction getUseInstruction() { result = this.getUse() }
 
   /**
    * DEPRECATED: use `getAnyDef` or `getDef`. The exact replacement for this
@@ -91,7 +91,7 @@ class Operand extends TStageOperand {
    *
    * Gets the `Instruction` whose result is the value of the operand.
    */
-  deprecated final Instruction getDefinitionInstruction() { result = getAnyDef() }
+  deprecated final Instruction getDefinitionInstruction() { result = this.getAnyDef() }
 
   /**
    * Gets the overlap relationship between the operand's definition and its use.
@@ -101,7 +101,9 @@ class Operand extends TStageOperand {
   /**
    * Holds if the result of the definition instruction does not exactly overlap this use.
    */
-  final predicate isDefinitionInexact() { not getDefinitionOverlap() instanceof MustExactlyOverlap }
+  final predicate isDefinitionInexact() {
+    not this.getDefinitionOverlap() instanceof MustExactlyOverlap
+  }
 
   /**
    * Gets a prefix to use when dumping the operand in an operand list.
@@ -121,7 +123,7 @@ class Operand extends TStageOperand {
    * For example: `this:r3_5`
    */
   final string getDumpString() {
-    result = getDumpLabel() + getInexactSpecifier() + getDefinitionId()
+    result = this.getDumpLabel() + this.getInexactSpecifier() + this.getDefinitionId()
   }
 
   /**
@@ -129,9 +131,9 @@ class Operand extends TStageOperand {
    * definition is not modeled in SSA.
    */
   private string getDefinitionId() {
-    result = getAnyDef().getResultId()
+    result = this.getAnyDef().getResultId()
     or
-    not exists(getAnyDef()) and result = "m?"
+    not exists(this.getAnyDef()) and result = "m?"
   }
 
   /**
@@ -140,7 +142,7 @@ class Operand extends TStageOperand {
    * the empty string.
    */
   private string getInexactSpecifier() {
-    if isDefinitionInexact() then result = "~" else result = ""
+    if this.isDefinitionInexact() then result = "~" else result = ""
   }
 
   /**
@@ -155,7 +157,7 @@ class Operand extends TStageOperand {
    * the definition type, such as in the case of a partial read or a read from a pointer that
    * has been cast to a different type.
    */
-  Language::LanguageType getLanguageType() { result = getAnyDef().getResultLanguageType() }
+  Language::LanguageType getLanguageType() { result = this.getAnyDef().getResultLanguageType() }
 
   /**
    * Gets the language-neutral type of the value consumed by this operand. This is usually the same
@@ -164,7 +166,7 @@ class Operand extends TStageOperand {
    * from the definition type, such as in the case of a partial read or a read from a pointer that
    * has been cast to a different type.
    */
-  final IRType getIRType() { result = getLanguageType().getIRType() }
+  final IRType getIRType() { result = this.getLanguageType().getIRType() }
 
   /**
    * Gets the type of the value consumed by this operand. This is usually the same as the
@@ -173,7 +175,7 @@ class Operand extends TStageOperand {
    * the definition type, such as in the case of a partial read or a read from a pointer that
    * has been cast to a different type.
    */
-  final Language::Type getType() { getLanguageType().hasType(result, _) }
+  final Language::Type getType() { this.getLanguageType().hasType(result, _) }
 
   /**
    * Holds if the value consumed by this operand is a glvalue. If this
@@ -182,13 +184,13 @@ class Operand extends TStageOperand {
    * not hold, the value of the operand represents a value whose type is
    * given by `getType()`.
    */
-  final predicate isGLValue() { getLanguageType().hasType(_, true) }
+  final predicate isGLValue() { this.getLanguageType().hasType(_, true) }
 
   /**
    * Gets the size of the value consumed by this operand, in bytes. If the operand does not have
    * a known constant size, this predicate does not hold.
    */
-  final int getSize() { result = getLanguageType().getByteSize() }
+  final int getSize() { result = this.getLanguageType().getByteSize() }
 }
 
 /**
@@ -205,7 +207,7 @@ class MemoryOperand extends Operand {
   /**
    * Gets the kind of memory access performed by the operand.
    */
-  MemoryAccessKind getMemoryAccess() { result = getUse().getOpcode().getReadMemoryAccess() }
+  MemoryAccessKind getMemoryAccess() { result = this.getUse().getOpcode().getReadMemoryAccess() }
 
   /**
    * Holds if the memory access performed by this operand will not always read from every bit in the
@@ -215,7 +217,7 @@ class MemoryOperand extends Operand {
    * conservative estimate of the memory that might actually be accessed at runtime (for example,
    * the global side effects of a function call).
    */
-  predicate hasMayReadMemoryAccess() { getUse().getOpcode().hasMayReadMemoryAccess() }
+  predicate hasMayReadMemoryAccess() { this.getUse().getOpcode().hasMayReadMemoryAccess() }
 
   /**
    * Returns the operand that holds the memory address from which the current operand loads its
@@ -223,8 +225,8 @@ class MemoryOperand extends Operand {
    * is `r1`.
    */
   final AddressOperand getAddressOperand() {
-    getMemoryAccess().usesAddressOperand() and
+    this.getMemoryAccess().usesAddressOperand() and
-    result.getUse() = getUse()
+    result.getUse() = this.getUse()
   }
 }
 
@@ -294,7 +296,7 @@ class NonPhiMemoryOperand extends NonPhiOperand, MemoryOperand, TNonPhiMemoryOpe
     result = unique(Instruction defInstr | hasDefinition(defInstr, _))
   }
 
-  final override Overlap getDefinitionOverlap() { hasDefinition(_, result) }
+  final override Overlap getDefinitionOverlap() { this.hasDefinition(_, result) }
 
   pragma[noinline]
   private predicate hasDefinition(Instruction defInstr, Overlap overlap) {
@@ -449,13 +451,17 @@ class PhiInputOperand extends MemoryOperand, TPhiOperand {
 
   final override Overlap getDefinitionOverlap() { result = overlap }
 
-  final override int getDumpSortOrder() { result = 11 + getPredecessorBlock().getDisplayIndex() }
+  final override int getDumpSortOrder() {
-
+    result = 11 + this.getPredecessorBlock().getDisplayIndex()
-  final override string getDumpLabel() {
-    result = "from " + getPredecessorBlock().getDisplayIndex().toString() + ":"
   }
 
-  final override string getDumpId() { result = getPredecessorBlock().getDisplayIndex().toString() }
+  final override string getDumpLabel() {
+    result = "from " + this.getPredecessorBlock().getDisplayIndex().toString() + ":"
+  }
+
+  final override string getDumpId() {
+    result = this.getPredecessorBlock().getDisplayIndex().toString()
+  }
 
   /**
    * Gets the predecessor block from which this value comes.

cpp/ql/lib/semmle/code/cpp/ir/implementation/unaliased_ssa/IRBlock.qll

@@ -24,7 +24,7 @@ class IRBlockBase extends TIRBlock {
   final string toString() { result = getFirstInstruction(this).toString() }
 
   /** Gets the source location of the first non-`Phi` instruction in this block. */
-  final Language::Location getLocation() { result = getFirstInstruction().getLocation() }
+  final Language::Location getLocation() { result = this.getFirstInstruction().getLocation() }
 
   /**
    * INTERNAL: Do not use.
@@ -39,7 +39,7 @@ class IRBlockBase extends TIRBlock {
     ) and
     this =
       rank[result + 1](IRBlock funcBlock, int sortOverride, int sortKey1, int sortKey2 |
-        funcBlock.getEnclosingFunction() = getEnclosingFunction() and
+        funcBlock.getEnclosingFunction() = this.getEnclosingFunction() and
         funcBlock.getFirstInstruction().hasSortKeys(sortKey1, sortKey2) and
         // Ensure that the block containing `EnterFunction` always comes first.
         if funcBlock.getFirstInstruction() instanceof EnterFunctionInstruction
@@ -59,15 +59,15 @@ class IRBlockBase extends TIRBlock {
    * Get the `Phi` instructions that appear at the start of this block.
    */
   final PhiInstruction getAPhiInstruction() {
-    Construction::getPhiInstructionBlockStart(result) = getFirstInstruction()
+    Construction::getPhiInstructionBlockStart(result) = this.getFirstInstruction()
   }
 
   /**
    * Gets an instruction in this block. This includes `Phi` instructions.
    */
   final Instruction getAnInstruction() {
-    result = getInstruction(_) or
+    result = this.getInstruction(_) or
-    result = getAPhiInstruction()
+    result = this.getAPhiInstruction()
   }
 
   /**
@@ -78,7 +78,9 @@ class IRBlockBase extends TIRBlock {
   /**
    * Gets the last instruction in this block.
    */
-  final Instruction getLastInstruction() { result = getInstruction(getInstructionCount() - 1) }
+  final Instruction getLastInstruction() {
+    result = this.getInstruction(this.getInstructionCount() - 1)
+  }
 
   /**
    * Gets the number of non-`Phi` instructions in this block.
@@ -149,7 +151,7 @@ class IRBlock extends IRBlockBase {
    * Block `A` dominates block `B` if any control flow path from the entry block of the function to
    * block `B` must pass through block `A`. A block always dominates itself.
    */
-  final predicate dominates(IRBlock block) { strictlyDominates(block) or this = block }
+  final predicate dominates(IRBlock block) { this.strictlyDominates(block) or this = block }
 
   /**
    * Gets a block on the dominance frontier of this block.
@@ -159,8 +161,8 @@ class IRBlock extends IRBlockBase {
    */
   pragma[noinline]
   final IRBlock dominanceFrontier() {
-    dominates(result.getAPredecessor()) and
+    this.dominates(result.getAPredecessor()) and
-    not strictlyDominates(result)
+    not this.strictlyDominates(result)
   }
 
   /**
@@ -189,7 +191,7 @@ class IRBlock extends IRBlockBase {
    * Block `A` post-dominates block `B` if any control flow path from `B` to the exit block of the
    * function must pass through block `A`. A block always post-dominates itself.
    */
-  final predicate postDominates(IRBlock block) { strictlyPostDominates(block) or this = block }
+  final predicate postDominates(IRBlock block) { this.strictlyPostDominates(block) or this = block }
 
   /**
    * Gets a block on the post-dominance frontier of this block.
@@ -199,16 +201,16 @@ class IRBlock extends IRBlockBase {
    */
   pragma[noinline]
   final IRBlock postPominanceFrontier() {
-    postDominates(result.getASuccessor()) and
+    this.postDominates(result.getASuccessor()) and
-    not strictlyPostDominates(result)
+    not this.strictlyPostDominates(result)
   }
 
   /**
    * Holds if this block is reachable from the entry block of its function.
    */
   final predicate isReachableFromFunctionEntry() {
-    this = getEnclosingIRFunction().getEntryBlock() or
+    this = this.getEnclosingIRFunction().getEntryBlock() or
-    getAPredecessor().isReachableFromFunctionEntry()
+    this.getAPredecessor().isReachableFromFunctionEntry()
   }
 }
 

cpp/ql/lib/semmle/code/cpp/ir/implementation/unaliased_ssa/Operand.qll

@@ -46,12 +46,12 @@ class Operand extends TStageOperand {
   /**
    * Gets the location of the source code for this operand.
    */
-  final Language::Location getLocation() { result = getUse().getLocation() }
+  final Language::Location getLocation() { result = this.getUse().getLocation() }
 
   /**
    * Gets the function that contains this operand.
    */
-  final IRFunction getEnclosingIRFunction() { result = getUse().getEnclosingIRFunction() }
+  final IRFunction getEnclosingIRFunction() { result = this.getUse().getEnclosingIRFunction() }
 
   /**
    * Gets the `Instruction` that consumes this operand.
@@ -74,7 +74,7 @@ class Operand extends TStageOperand {
    */
   final Instruction getDef() {
     result = this.getAnyDef() and
-    getDefinitionOverlap() instanceof MustExactlyOverlap
+    this.getDefinitionOverlap() instanceof MustExactlyOverlap
   }
 
   /**
@@ -82,7 +82,7 @@ class Operand extends TStageOperand {
    *
    * Gets the `Instruction` that consumes this operand.
    */
-  deprecated final Instruction getUseInstruction() { result = getUse() }
+  deprecated final Instruction getUseInstruction() { result = this.getUse() }
 
   /**
    * DEPRECATED: use `getAnyDef` or `getDef`. The exact replacement for this
@@ -91,7 +91,7 @@ class Operand extends TStageOperand {
    *
    * Gets the `Instruction` whose result is the value of the operand.
    */
-  deprecated final Instruction getDefinitionInstruction() { result = getAnyDef() }
+  deprecated final Instruction getDefinitionInstruction() { result = this.getAnyDef() }
 
   /**
    * Gets the overlap relationship between the operand's definition and its use.
@@ -101,7 +101,9 @@ class Operand extends TStageOperand {
   /**
    * Holds if the result of the definition instruction does not exactly overlap this use.
    */
-  final predicate isDefinitionInexact() { not getDefinitionOverlap() instanceof MustExactlyOverlap }
+  final predicate isDefinitionInexact() {
+    not this.getDefinitionOverlap() instanceof MustExactlyOverlap
+  }
 
   /**
    * Gets a prefix to use when dumping the operand in an operand list.
@@ -121,7 +123,7 @@ class Operand extends TStageOperand {
    * For example: `this:r3_5`
    */
   final string getDumpString() {
-    result = getDumpLabel() + getInexactSpecifier() + getDefinitionId()
+    result = this.getDumpLabel() + this.getInexactSpecifier() + this.getDefinitionId()
   }
 
   /**
@@ -129,9 +131,9 @@ class Operand extends TStageOperand {
    * definition is not modeled in SSA.
    */
   private string getDefinitionId() {
-    result = getAnyDef().getResultId()
+    result = this.getAnyDef().getResultId()
     or
-    not exists(getAnyDef()) and result = "m?"
+    not exists(this.getAnyDef()) and result = "m?"
   }
 
   /**
@@ -140,7 +142,7 @@ class Operand extends TStageOperand {
    * the empty string.
    */
   private string getInexactSpecifier() {
-    if isDefinitionInexact() then result = "~" else result = ""
+    if this.isDefinitionInexact() then result = "~" else result = ""
   }
 
   /**
@@ -155,7 +157,7 @@ class Operand extends TStageOperand {
    * the definition type, such as in the case of a partial read or a read from a pointer that
    * has been cast to a different type.
    */
-  Language::LanguageType getLanguageType() { result = getAnyDef().getResultLanguageType() }
+  Language::LanguageType getLanguageType() { result = this.getAnyDef().getResultLanguageType() }
 
   /**
    * Gets the language-neutral type of the value consumed by this operand. This is usually the same
@@ -164,7 +166,7 @@ class Operand extends TStageOperand {
    * from the definition type, such as in the case of a partial read or a read from a pointer that
    * has been cast to a different type.
    */
-  final IRType getIRType() { result = getLanguageType().getIRType() }
+  final IRType getIRType() { result = this.getLanguageType().getIRType() }
 
   /**
    * Gets the type of the value consumed by this operand. This is usually the same as the
@@ -173,7 +175,7 @@ class Operand extends TStageOperand {
    * the definition type, such as in the case of a partial read or a read from a pointer that
    * has been cast to a different type.
    */
-  final Language::Type getType() { getLanguageType().hasType(result, _) }
+  final Language::Type getType() { this.getLanguageType().hasType(result, _) }
 
   /**
    * Holds if the value consumed by this operand is a glvalue. If this
@@ -182,13 +184,13 @@ class Operand extends TStageOperand {
    * not hold, the value of the operand represents a value whose type is
    * given by `getType()`.
    */
-  final predicate isGLValue() { getLanguageType().hasType(_, true) }
+  final predicate isGLValue() { this.getLanguageType().hasType(_, true) }
 
   /**
    * Gets the size of the value consumed by this operand, in bytes. If the operand does not have
    * a known constant size, this predicate does not hold.
    */
-  final int getSize() { result = getLanguageType().getByteSize() }
+  final int getSize() { result = this.getLanguageType().getByteSize() }
 }
 
 /**
@@ -205,7 +207,7 @@ class MemoryOperand extends Operand {
   /**
    * Gets the kind of memory access performed by the operand.
    */
-  MemoryAccessKind getMemoryAccess() { result = getUse().getOpcode().getReadMemoryAccess() }
+  MemoryAccessKind getMemoryAccess() { result = this.getUse().getOpcode().getReadMemoryAccess() }
 
   /**
    * Holds if the memory access performed by this operand will not always read from every bit in the
@@ -215,7 +217,7 @@ class MemoryOperand extends Operand {
    * conservative estimate of the memory that might actually be accessed at runtime (for example,
    * the global side effects of a function call).
    */
-  predicate hasMayReadMemoryAccess() { getUse().getOpcode().hasMayReadMemoryAccess() }
+  predicate hasMayReadMemoryAccess() { this.getUse().getOpcode().hasMayReadMemoryAccess() }
 
   /**
    * Returns the operand that holds the memory address from which the current operand loads its
@@ -223,8 +225,8 @@ class MemoryOperand extends Operand {
    * is `r1`.
    */
   final AddressOperand getAddressOperand() {
-    getMemoryAccess().usesAddressOperand() and
+    this.getMemoryAccess().usesAddressOperand() and
-    result.getUse() = getUse()
+    result.getUse() = this.getUse()
   }
 }
 
@@ -294,7 +296,7 @@ class NonPhiMemoryOperand extends NonPhiOperand, MemoryOperand, TNonPhiMemoryOpe
     result = unique(Instruction defInstr | hasDefinition(defInstr, _))
   }
 
-  final override Overlap getDefinitionOverlap() { hasDefinition(_, result) }
+  final override Overlap getDefinitionOverlap() { this.hasDefinition(_, result) }
 
   pragma[noinline]
   private predicate hasDefinition(Instruction defInstr, Overlap overlap) {
@@ -449,13 +451,17 @@ class PhiInputOperand extends MemoryOperand, TPhiOperand {
 
   final override Overlap getDefinitionOverlap() { result = overlap }
 
-  final override int getDumpSortOrder() { result = 11 + getPredecessorBlock().getDisplayIndex() }
+  final override int getDumpSortOrder() {
-
+    result = 11 + this.getPredecessorBlock().getDisplayIndex()
-  final override string getDumpLabel() {
-    result = "from " + getPredecessorBlock().getDisplayIndex().toString() + ":"
   }
 
-  final override string getDumpId() { result = getPredecessorBlock().getDisplayIndex().toString() }
+  final override string getDumpLabel() {
+    result = "from " + this.getPredecessorBlock().getDisplayIndex().toString() + ":"
+  }
+
+  final override string getDumpId() {
+    result = this.getPredecessorBlock().getDisplayIndex().toString()
+  }
 
   /**
    * Gets the predecessor block from which this value comes.

cpp/ql/lib/semmle/code/cpp/ir/internal/IntegerInterval.qll

@@ -18,10 +18,11 @@ Overlap getOverlap(IntValue defStart, IntValue defEnd, IntValue useStart, IntVal
   else
     if isLE(defStart, useStart) and isGE(defEnd, useEnd)
     then result instanceof MustTotallyOverlap
-    else
+    else (
-      if isLE(defEnd, useStart) or isGE(defStart, useEnd)
+      not isLE(defEnd, useStart) and
-      then none()
+      not isGE(defStart, useEnd) and
-      else result instanceof MayPartiallyOverlap
+      result instanceof MayPartiallyOverlap
+    )
 }
 
 /**

cpp/ql/lib/semmle/code/cpp/models/Models.qll

@@ -36,3 +36,4 @@ private import implementations.Select
 private import implementations.MySql
 private import implementations.SqLite3
 private import implementations.PostgreSql
+private import implementations.System

cpp/ql/lib/semmle/code/cpp/models/implementations/Allocation.qll

@@ -56,7 +56,7 @@ private class MallocAllocationFunction extends AllocationFunction {
       ]) and
     sizeArg = 1
     or
-    hasGlobalName(["HeapAlloc"]) and // HeapAlloc(heap, flags, size)
+    hasGlobalName("HeapAlloc") and // HeapAlloc(heap, flags, size)
     sizeArg = 2
     or
     hasGlobalName([

cpp/ql/lib/semmle/code/cpp/models/implementations/Recv.qll

@@ -85,4 +85,6 @@ private class Recv extends AliasFunction, ArrayFunction, SideEffectFunction,
     ) and
     description = "Buffer read by " + this.getName()
   }
+
+  override predicate hasSocketInput(FunctionInput input) { input.isParameter(0) }
 }

cpp/ql/lib/semmle/code/cpp/models/implementations/Send.qll

@@ -60,4 +60,6 @@ private class Send extends AliasFunction, ArrayFunction, SideEffectFunction, Rem
   override predicate hasRemoteFlowSink(FunctionInput input, string description) {
     input.isParameterDeref(1) and description = "Buffer sent by " + this.getName()
   }
+
+  override predicate hasSocketInput(FunctionInput input) { input.isParameter(0) }
 }

cpp/ql/lib/semmle/code/cpp/models/implementations/System.qll

@@ -0,0 +1,43 @@
+import cpp
+import semmle.code.cpp.models.interfaces.SideEffect
+import semmle.code.cpp.models.interfaces.Alias
+import semmle.code.cpp.models.interfaces.CommandExecution
+
+/**
+ * A function for running a command using a command interpreter.
+ */
+private class SystemFunction extends CommandExecutionFunction, ArrayFunction, AliasFunction,
+  SideEffectFunction {
+  SystemFunction() {
+    hasGlobalOrStdName("system") or // system(command)
+    hasGlobalName("popen") or // popen(command, mode)
+    // Windows variants
+    hasGlobalName("_popen") or // _popen(command, mode)
+    hasGlobalName("_wpopen") or // _wpopen(command, mode)
+    hasGlobalName("_wsystem") // _wsystem(command)
+  }
+
+  override predicate hasCommandArgument(FunctionInput input) { input.isParameterDeref(0) }
+
+  override predicate hasArrayWithNullTerminator(int bufParam) { bufParam = 0 or bufParam = 1 }
+
+  override predicate hasArrayInput(int bufParam) { bufParam = 0 or bufParam = 1 }
+
+  override predicate parameterNeverEscapes(int index) { index = 0 or index = 1 }
+
+  override predicate parameterEscapesOnlyViaReturn(int index) { none() }
+
+  override predicate parameterIsAlwaysReturned(int index) { none() }
+
+  override predicate hasOnlySpecificReadSideEffects() { any() }
+
+  override predicate hasOnlySpecificWriteSideEffects() {
+    hasGlobalOrStdName("system") or
+    hasGlobalName("_wsystem")
+  }
+
+  override predicate hasSpecificReadSideEffect(ParameterIndex i, boolean buffer) {
+    (i = 0 or i = 1) and
+    buffer = true
+  }
+}

cpp/ql/lib/semmle/code/cpp/models/interfaces/CommandExecution.qll

@@ -0,0 +1,23 @@
+/**
+ * Provides classes for modeling functions that execute new programs by
+ * interpreting string data as shell commands. To use this QL library, create
+ * a QL class extending `CommandExecutionFunction` with a characteristic
+ * predicate that selects the function or set of functions you are modeling.
+ * Within that class, override the `hasCommandArgument` predicate to indicate
+ * which parameters are interpreted as shell commands.
+ */
+
+import cpp
+import FunctionInputsAndOutputs
+import semmle.code.cpp.models.Models
+
+/**
+ * A function, such as `exec` or `popen` that starts a new process by
+ * interpreting a string as a shell command.
+ */
+abstract class CommandExecutionFunction extends Function {
+  /**
+   * Holds if `input` is interpreted as a shell command.
+   */
+  abstract predicate hasCommandArgument(FunctionInput input);
+}

cpp/ql/lib/semmle/code/cpp/models/interfaces/FlowSource.qll

@@ -18,6 +18,12 @@ abstract class RemoteFlowSourceFunction extends Function {
    * Holds if remote data described by `description` flows from `output` of a call to this function.
    */
   abstract predicate hasRemoteFlowSource(FunctionOutput output, string description);
+
+  /**
+   * Holds if remote data from this source comes from a socket described by
+   * `input`. There is no result if a socket is not specified.
+   */
+  predicate hasSocketInput(FunctionInput input) { none() }
 }
 
 /**
@@ -51,4 +57,10 @@ abstract class RemoteFlowSinkFunction extends Function {
    * send over a network connection.
    */
   abstract predicate hasRemoteFlowSink(FunctionInput input, string description);
+
+  /**
+   * Holds if data put into this sink is transmitted through a socket described
+   * by `input`. There is no result if a socket is not specified.
+   */
+  predicate hasSocketInput(FunctionInput input) { none() }
 }

cpp/ql/lib/semmle/code/cpp/rangeanalysis/NanAnalysis.qll

@@ -29,8 +29,7 @@ predicate nanExcludingComparison(ComparisonOperation guard, boolean polarity) {
  */
 private predicate excludesNan(RangeSsaDefinition def, VariableAccess v) {
   exists(VariableAccess inCond, ComparisonOperation guard, boolean branch, StackVariable lsv |
-    def.isGuardPhi(inCond, guard, branch) and
+    def.isGuardPhi(lsv, inCond, guard, branch) and
-    inCond.getTarget() = lsv and
     v = def.getAUse(lsv) and
     guard.getAnOperand() = inCond and
     nanExcludingComparison(guard, branch)

cpp/ql/lib/semmle/code/cpp/rangeanalysis/RangeSSA.qll

@@ -94,10 +94,11 @@ class RangeSsaDefinition extends ControlFlowNodeBase {
   predicate isPhiNode(StackVariable v) { exists(RangeSSA x | x.phi_node(v, this.(BasicBlock))) }
 
   /**
+   * DEPRECATED: Use isGuardPhi/4 instead
    * If this definition is a phi node corresponding to a guard,
    * then return the variable access and the guard.
    */
-  predicate isGuardPhi(VariableAccess va, Expr guard, boolean branch) {
+  deprecated predicate isGuardPhi(VariableAccess va, Expr guard, boolean branch) {
     guard_defn(va, guard, this, branch)
   }
 
@@ -142,9 +143,8 @@ class RangeSsaDefinition extends ControlFlowNodeBase {
       // below excludes definitions which can only reach guard phi
       // nodes by going through the corresponding guard.
       not exists(VariableAccess access |
-        v = access.getTarget() and
         pred.contains(access) and
-        this.isGuardPhi(access, _, _)
+        this.isGuardPhi(v, access, _, _)
       )
     )
   }

cpp/ql/lib/semmle/code/cpp/rangeanalysis/SimpleRangeAnalysis.qll

@@ -433,10 +433,7 @@ private predicate exprDependsOnDef(Expr e, RangeSsaDefinition srcDef, StackVaria
 private predicate phiDependsOnDef(
   RangeSsaDefinition phi, StackVariable v, RangeSsaDefinition srcDef, StackVariable srcVar
 ) {
-  exists(VariableAccess access, Expr guard |
+  exists(VariableAccess access, Expr guard | phi.isGuardPhi(v, access, guard, _) |
-    access = v.getAnAccess() and
-    phi.isGuardPhi(access, guard, _)
-  |
     exprDependsOnDef(guard.(ComparisonOperation).getAnOperand(), srcDef, srcVar) or
     exprDependsOnDef(access, srcDef, srcVar)
   )
@@ -1204,8 +1201,7 @@ private float boolConversionUpperBound(Expr expr) {
  */
 private float getPhiLowerBounds(StackVariable v, RangeSsaDefinition phi) {
   exists(VariableAccess access, Expr guard, boolean branch, float defLB, float guardLB |
-    access = v.getAnAccess() and
+    phi.isGuardPhi(v, access, guard, branch) and
-    phi.isGuardPhi(access, guard, branch) and
     lowerBoundFromGuard(guard, access, guardLB, branch) and
     defLB = getFullyConvertedLowerBounds(access)
   |
@@ -1230,8 +1226,7 @@ private float getPhiLowerBounds(StackVariable v, RangeSsaDefinition phi) {
 /** See comment for `getPhiLowerBounds`, above. */
 private float getPhiUpperBounds(StackVariable v, RangeSsaDefinition phi) {
   exists(VariableAccess access, Expr guard, boolean branch, float defUB, float guardUB |
-    access = v.getAnAccess() and
+    phi.isGuardPhi(v, access, guard, branch) and
-    phi.isGuardPhi(access, guard, branch) and
     upperBoundFromGuard(guard, access, guardUB, branch) and
     defUB = getFullyConvertedUpperBounds(access)
   |
@@ -1493,8 +1488,7 @@ private predicate isNEPhi(
   exists(
     ComparisonOperation cmp, boolean branch, Expr linearExpr, Expr rExpr, float p, float q, float r
   |
-    access.getTarget() = v and
+    phi.isGuardPhi(v, access, cmp, branch) and
-    phi.isGuardPhi(access, cmp, branch) and
     eqOpWithSwapAndNegate(cmp, linearExpr, rExpr, false, branch) and
     v.getUnspecifiedType() instanceof IntegralOrEnumType and // Float `!=` is too imprecise
     r = getValue(rExpr).toFloat() and
@@ -1503,8 +1497,7 @@ private predicate isNEPhi(
   )
   or
   exists(Expr op, boolean branch, Expr linearExpr, float p, float q |
-    access.getTarget() = v and
+    phi.isGuardPhi(v, access, op, branch) and
-    phi.isGuardPhi(access, op, branch) and
     eqZeroWithNegate(op, linearExpr, false, branch) and
     v.getUnspecifiedType() instanceof IntegralOrEnumType and // Float `!` is too imprecise
     linearAccess(linearExpr, access, p, q) and
@@ -1524,8 +1517,7 @@ private predicate isUnsupportedGuardPhi(Variable v, RangeSsaDefinition phi, Vari
     or
     eqZeroWithNegate(cmp, _, false, branch)
   |
-    access.getTarget() = v and
+    phi.isGuardPhi(v, access, cmp, branch) and
-    phi.isGuardPhi(access, cmp, branch) and
     not isNEPhi(v, phi, access, _)
   )
 }
@@ -1594,6 +1586,15 @@ private module SimpleRangeAnalysisCached {
     result = min([max(getTruncatedUpperBounds(expr)), getGuardedUpperBound(expr)])
   }
 
+  /** Holds if the upper bound of `expr` may have been widened. This means the the upper bound is in practice likely to be overly wide. */
+  cached
+  predicate upperBoundMayBeWidened(Expr e) {
+    isRecursiveExpr(e) and
+    // Widening is not a problem if the post-analysis in `getGuardedUpperBound` has overridden the widening.
+    // Note that the RHS of `<` may be multi-valued.
+    not getGuardedUpperBound(e) < getTruncatedUpperBounds(e)
+  }
+
   /**
    * Holds if `expr` has a provably empty range. For example:
    *

cpp/ql/lib/semmle/code/cpp/security/BufferWrite.qll

@@ -365,7 +365,7 @@ class GetsBW extends BufferWriteCall {
   /**
    * Gets the index of the parameter that is the maximum number of characters to be read.
    */
-  int getParamSize() { if exists(getArgument(1)) then result = 1 else none() }
+  int getParamSize() { exists(getArgument(1)) and result = 1 }
 
   override Type getBufferType() { result = this.getTarget().getParameter(0).getUnspecifiedType() }
 

cpp/ql/lib/semmle/code/cpp/security/CommandExecution.qll

@@ -4,42 +4,20 @@ import cpp
 import semmle.code.cpp.security.FunctionWithWrappers
 import semmle.code.cpp.models.interfaces.SideEffect
 import semmle.code.cpp.models.interfaces.Alias
+import semmle.code.cpp.models.interfaces.CommandExecution
 
 /**
  * A function for running a command using a command interpreter.
  */
-class SystemFunction extends FunctionWithWrappers, ArrayFunction, AliasFunction, SideEffectFunction {
+class SystemFunction extends FunctionWithWrappers instanceof CommandExecutionFunction {
-  SystemFunction() {
+  override predicate interestingArg(int arg) {
-    hasGlobalOrStdName("system") or // system(command)
+    exists(FunctionInput input |
-    hasGlobalName("popen") or // popen(command, mode)
+      this.(CommandExecutionFunction).hasCommandArgument(input) and
-    // Windows variants
+      (
-    hasGlobalName("_popen") or // _popen(command, mode)
+        input.isParameterDerefOrQualifierObject(arg) or
-    hasGlobalName("_wpopen") or // _wpopen(command, mode)
+        input.isParameterOrQualifierAddress(arg)
-    hasGlobalName("_wsystem") // _wsystem(command)
+      )
-  }
+    )
-
-  override predicate interestingArg(int arg) { arg = 0 }
-
-  override predicate hasArrayWithNullTerminator(int bufParam) { bufParam = 0 or bufParam = 1 }
-
-  override predicate hasArrayInput(int bufParam) { bufParam = 0 or bufParam = 1 }
-
-  override predicate parameterNeverEscapes(int index) { index = 0 or index = 1 }
-
-  override predicate parameterEscapesOnlyViaReturn(int index) { none() }
-
-  override predicate parameterIsAlwaysReturned(int index) { none() }
-
-  override predicate hasOnlySpecificReadSideEffects() { any() }
-
-  override predicate hasOnlySpecificWriteSideEffects() {
-    hasGlobalOrStdName("system") or
-    hasGlobalName("_wsystem")
-  }
-
-  override predicate hasSpecificReadSideEffect(ParameterIndex i, boolean buffer) {
-    (i = 0 or i = 1) and
-    buffer = true
   }
 }
 

cpp/ql/lib/semmle/code/cpp/security/FunctionWithWrappers.qll

@@ -17,7 +17,7 @@
 
 import cpp
 import PrintfLike
-private import TaintTracking
+private import semmle.code.cpp.ir.dataflow.ResolveCall
 
 bindingset[index]
 private string toCause(Function func, int index) {

cpp/ql/lib/semmlecode.cpp.dbscheme

@@ -245,7 +245,7 @@ svnchurn(
  * The location spans column `startcolumn` of line `startline` to
  * column `endcolumn` of line `endline` in file `file`.
  * For more information, see
- * [Locations](https://help.semmle.com/QL/learn-ql/ql/locations.html).
+ * [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
  */
 locations_default(
     /** The location of an element that is not an expression or a statement. */
@@ -262,7 +262,7 @@ locations_default(
  * The location spans column `startcolumn` of line `startline` to
  * column `endcolumn` of line `endline` in file `file`.
  * For more information, see
- * [Locations](https://help.semmle.com/QL/learn-ql/ql/locations.html).
+ * [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
  */
 locations_stmt(
     /** The location of a statement. */
@@ -279,7 +279,7 @@ locations_stmt(
  * The location spans column `startcolumn` of line `startline` to
  * column `endcolumn` of line `endline` in file `file`.
  * For more information, see
- * [Locations](https://help.semmle.com/QL/learn-ql/ql/locations.html).
+ * [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
  */
 locations_expr(
     /** The location of an expression. */

cpp/ql/src/AlertSuppression.ql

@@ -60,20 +60,18 @@ class SuppressionComment extends Comment {
 /**
  * The scope of an alert suppression comment.
  */
-class SuppressionScope extends ElementBase {
+class SuppressionScope extends ElementBase instanceof SuppressionComment {
-  SuppressionScope() { this instanceof SuppressionComment }
-
   /**
    * Holds if this element is at the specified location.
    * The location spans column `startcolumn` of line `startline` to
    * column `endcolumn` of line `endline` in file `filepath`.
    * For more information, see
-   * [Locations](https://help.semmle.com/QL/learn-ql/ql/locations.html).
+   * [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
    */
   predicate hasLocationInfo(
     string filepath, int startline, int startcolumn, int endline, int endcolumn
   ) {
-    this.(SuppressionComment).covers(filepath, startline, startcolumn, endline, endcolumn)
+    super.covers(filepath, startline, startcolumn, endline, endcolumn)
   }
 }
 

cpp/ql/src/Critical/OverflowStatic.ql

@@ -5,7 +5,7 @@
  * @kind problem
  * @problem.severity warning
  * @security-severity 9.3
- * @precision medium
+ * @precision high
  * @id cpp/static-buffer-overflow
  * @tags reliability
  *       security
@@ -55,6 +55,8 @@ predicate overflowOffsetInLoop(BufferAccess bufaccess, string msg) {
     loop.counter().getAnAccess() = bufaccess.getArrayOffset() and
     // Ensure that we don't have an upper bound on the array index that's less than the buffer size.
     not upperBound(bufaccess.getArrayOffset().getFullyConverted()) < bufaccess.bufferSize() and
+    // The upper bounds analysis must not have been widended
+    not upperBoundMayBeWidened(bufaccess.getArrayOffset().getFullyConverted()) and
     msg =
       "Potential buffer-overflow: counter '" + loop.counter().toString() + "' <= " +
         loop.limit().toString() + " but '" + bufaccess.buffer().getName() + "' has " +
@@ -130,11 +132,13 @@ predicate outOfBounds(BufferAccess bufaccess, string msg) {
     (
       access > size
       or
-      access = size and not exists(AddressOfExpr addof | bufaccess = addof.getOperand())
+      access = size and
+      not exists(AddressOfExpr addof | bufaccess = addof.getOperand()) and
+      not exists(BuiltInOperationBuiltInOffsetOf offsetof | offsetof.getAChild() = bufaccess)
     ) and
     msg =
       "Potential buffer-overflow: '" + buf + "' has size " + size.toString() + " but '" + buf + "[" +
-        access.toString() + "]' is accessed here."
+        access.toString() + "]' may be accessed here."
   )
 }
 

cpp/ql/src/Diagnostics/ExtractionErrors.ql

@@ -1,16 +0,0 @@
-/**
- * @name Extraction errors
- * @description List all extraction errors for files in the source code directory.
- * @kind diagnostic
- * @id cpp/diagnostics/extraction-errors
- */
-
-import cpp
-import ExtractionErrors
-
-from ExtractionError error
-where
-  error instanceof ExtractionUnknownError or
-  exists(error.getFile().getRelativePath())
-select error, "Extraction failed in " + error.getFile() + " with error " + error.getErrorMessage(),
-  error.getSeverity()

cpp/ql/src/Diagnostics/ExtractionProblems.qll

@@ -1,12 +1,12 @@
 /**
- * Provides a common hierarchy of all types of errors that can occur during extraction.
+ * Provides a common hierarchy of all types of problems that can occur during extraction.
  */
 
 import cpp
 
 /*
  * A note about how the C/C++ extractor emits diagnostics:
- * When the extractor frontend encounters an error, it emits a diagnostic message,
+ * When the extractor frontend encounters a problem, it emits a diagnostic message,
  * that includes a message, location and severity.
  * However, that process is best-effort and may fail (e.g. due to lack of memory).
  * Thus, if the extractor emitted at least one diagnostic of severity discretionary
@@ -15,17 +15,17 @@ import cpp
  * In the common case, this means that a compilation during which one or more errors happened also gets
  * the catch-all diagnostic.
  * This diagnostic has the empty string as file path.
- * We filter out these useless diagnostics if there is at least one error-level diagnostic
+ * We filter out these useless diagnostics if there is at least one warning-level diagnostic
  * for the affected compilation in the database.
  * Otherwise, we show it to indicate that something went wrong and that we
  * don't know what exactly happened.
  */
 
 /**
- * An error that, if present, leads to a file being marked as non-successfully extracted.
+ * A problem with a file that, if present, leads to a file being marked as non-successfully extracted.
  */
-class ReportableError extends Diagnostic {
+class ReportableWarning extends Diagnostic {
-  ReportableError() {
+  ReportableWarning() {
     (
       this instanceof CompilerDiscretionaryError or
       this instanceof CompilerError or
@@ -36,39 +36,35 @@ class ReportableError extends Diagnostic {
   }
 }
 
-private newtype TExtractionError =
+private newtype TExtractionProblem =
-  TReportableError(ReportableError err) or
+  TReportableWarning(ReportableWarning err) or
   TCompilationFailed(Compilation c, File f) {
     f = c.getAFileCompiled() and not c.normalTermination()
   } or
   // Show the catch-all diagnostic (see note above) only if we haven't seen any other error-level diagnostic
   // for that compilation
-  TUnknownError(CompilerError err) {
+  TUnknownProblem(CompilerError err) {
-    not exists(ReportableError e | e.getCompilation() = err.getCompilation())
+    not exists(ReportableWarning e | e.getCompilation() = err.getCompilation())
   }
 
 /**
- * Superclass for the extraction error hierarchy.
+ * Superclass for the extraction problem hierarchy.
  */
-class ExtractionError extends TExtractionError {
+class ExtractionProblem extends TExtractionProblem {
-  /** Gets the string representation of the error. */
+  /** Gets the string representation of the problem. */
   string toString() { none() }
 
-  /** Gets the error message for this error. */
+  /** Gets the problem message for this problem. */
-  string getErrorMessage() { none() }
+  string getProblemMessage() { none() }
 
-  /** Gets the file this error occured in. */
+  /** Gets the file this problem occured in. */
   File getFile() { none() }
 
-  /** Gets the location this error occured in. */
+  /** Gets the location this problem occured in. */
   Location getLocation() { none() }
 
-  /** Gets the SARIF severity of this error. */
+  /** Gets the SARIF severity of this problem. */
-  int getSeverity() {
+  int getSeverity() { none() }
-    // Unfortunately, we can't distinguish between errors and fatal errors in SARIF,
-    // so all errors have severity 2.
-    result = 2
-  }
 }
 
 /**
@@ -79,7 +75,7 @@ class ExtractionError extends TExtractionError {
  *  - stack overflow
  *  - out of memory
  */
-class ExtractionUnrecoverableError extends ExtractionError, TCompilationFailed {
+class ExtractionUnrecoverableError extends ExtractionProblem, TCompilationFailed {
   Compilation c;
   File f;
 
@@ -89,49 +85,67 @@ class ExtractionUnrecoverableError extends ExtractionError, TCompilationFailed {
     result = "Unrecoverable extraction error while compiling " + f.toString()
   }
 
-  override string getErrorMessage() { result = "unrecoverable compilation failure." }
+  override string getProblemMessage() { result = "unrecoverable compilation failure." }
 
   override File getFile() { result = f }
 
   override Location getLocation() { result = f.getLocation() }
+
+  override int getSeverity() {
+    // These extractor errors break the analysis, so we mark them in SARIF as
+    // [errors](https://docs.oasis-open.org/sarif/sarif/v2.1.0/csprd01/sarif-v2.1.0-csprd01.html#_Toc10541338).
+    result = 2
+  }
 }
 
 /**
- * A recoverable extraction error.
+ * A recoverable extraction warning.
  * These are compiler errors from the frontend.
  * Upon encountering one of these, we still continue extraction, but the
  * database will be incomplete for that file.
  */
-class ExtractionRecoverableError extends ExtractionError, TReportableError {
+class ExtractionRecoverableWarning extends ExtractionProblem, TReportableWarning {
-  ReportableError err;
+  ReportableWarning err;
 
-  ExtractionRecoverableError() { this = TReportableError(err) }
+  ExtractionRecoverableWarning() { this = TReportableWarning(err) }
 
   override string toString() { result = "Recoverable extraction error: " + err }
 
-  override string getErrorMessage() { result = err.getFullMessage() }
+  override string getProblemMessage() { result = err.getFullMessage() }
 
   override File getFile() { result = err.getFile() }
 
   override Location getLocation() { result = err.getLocation() }
+
+  override int getSeverity() {
+    // Recoverable extraction problems don't tend to break the analysis, so we mark them in SARIF as
+    // [warnings](https://docs.oasis-open.org/sarif/sarif/v2.1.0/csprd01/sarif-v2.1.0-csprd01.html#_Toc10541338).
+    result = 1
+  }
 }
 
 /**
- * An unknown error happened during extraction.
+ * An unknown problem happened during extraction.
- * These are only displayed if we know that we encountered an error during extraction,
+ * These are only displayed if we know that we encountered an problem during extraction,
  * but, for some reason, failed to emit a proper diagnostic with location information
- * and error message.
+ * and problem message.
  */
-class ExtractionUnknownError extends ExtractionError, TUnknownError {
+class ExtractionUnknownProblem extends ExtractionProblem, TUnknownProblem {
   CompilerError err;
 
-  ExtractionUnknownError() { this = TUnknownError(err) }
+  ExtractionUnknownProblem() { this = TUnknownProblem(err) }
 
-  override string toString() { result = "Unknown extraction error: " + err }
+  override string toString() { result = "Unknown extraction problem: " + err }
 
-  override string getErrorMessage() { result = err.getFullMessage() }
+  override string getProblemMessage() { result = err.getFullMessage() }
 
   override File getFile() { result = err.getFile() }
 
   override Location getLocation() { result = err.getLocation() }
+
+  override int getSeverity() {
+    // Unknown extraction problems don't tend to break the analysis, so we mark them in SARIF as
+    // [warnings](https://docs.oasis-open.org/sarif/sarif/v2.1.0/csprd01/sarif-v2.1.0-csprd01.html#_Toc10541338).
+    result = 1
+  }
 }

cpp/ql/src/Diagnostics/ExtractionWarnings.ql

@@ -0,0 +1,18 @@
+/**
+ * @name Extraction warnings
+ * @description List all extraction warnings for files in the source code directory.
+ * @kind diagnostic
+ * @id cpp/diagnostics/extraction-warnings
+ */
+
+import cpp
+import ExtractionProblems
+
+from ExtractionProblem warning
+where
+  warning instanceof ExtractionRecoverableWarning and exists(warning.getFile().getRelativePath())
+  or
+  warning instanceof ExtractionUnknownProblem
+select warning,
+  "Extraction failed in " + warning.getFile() + " with warning " + warning.getProblemMessage(),
+  warning.getSeverity()

cpp/ql/src/Diagnostics/FailedExtractorInvocations.ql

@@ -13,6 +13,9 @@ string describe(Compilation c) {
   else result = "extractor invocation " + concat(int i | | c.getArgument(i), " " order by i)
 }
 
+/** Gets the SARIF severity level that indicates an error. */
+private int getErrorSeverity() { result = 2 }
+
 from Compilation c
 where not c.normalTermination()
-select "Extraction aborted for " + describe(c)
+select "Extraction aborted for " + describe(c), getErrorSeverity()

cpp/ql/src/Diagnostics/SuccessfullyExtractedFiles.ql

@@ -1,15 +1,15 @@
 /**
  * @name Successfully extracted files
- * @description Lists all files in the source code directory that were extracted without encountering an error in the file.
+ * @description Lists all files in the source code directory that were extracted without encountering a problem in the file.
  * @kind diagnostic
  * @id cpp/diagnostics/successfully-extracted-files
  */
 
 import cpp
-import ExtractionErrors
+import ExtractionProblems
 
 from File f
 where
-  not exists(ExtractionError e | e.getFile() = f) and
+  not exists(ExtractionProblem e | e.getFile() = f) and
   exists(f.getRelativePath())
-select f, ""
+select f, "File successfully extracted"

cpp/ql/src/Documentation/CommentedOutCode.qll

@@ -198,7 +198,7 @@ class CommentBlock extends Comment {
    * The location spans column `startcolumn` of line `startline` to
    * column `endcolumn` of line `endline` in file `filepath`.
    * For more information, see
-   * [Locations](https://help.semmle.com/QL/learn-ql/ql/locations.html).
+   * [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
    */
   predicate hasLocationInfo(
     string filepath, int startline, int startcolumn, int endline, int endcolumn

cpp/ql/src/Likely Bugs/Arithmetic/IntMultToLong.ql

@@ -166,6 +166,7 @@ class VarAnalyzableExpr extends AnalyzableExpr, VariableAccess {
  * Holds if `t` is not an instance of `IntegralType`,
  * or if `me` cannot be proven to not overflow
  */
+pragma[inline]
 predicate overflows(MulExpr me, Type t) {
   t instanceof IntegralType
   implies

cpp/ql/src/Likely Bugs/Arithmetic/PointlessComparison.ql

@@ -50,10 +50,7 @@ where
     // If either of the operands is constant, then don't include it.
     (
       if cmp.getLeftOperand().isConstant()
-      then
+      then not cmp.getRightOperand().isConstant() and reason = rightReason
-        if cmp.getRightOperand().isConstant()
-        then none() // Both operands are constant so don't create a message.
-        else reason = rightReason
       else
         if cmp.getRightOperand().isConstant()
         then reason = leftReason

cpp/ql/src/Likely Bugs/Memory Management/ImproperNullTermination.ql

@@ -5,6 +5,7 @@
  * @kind problem
  * @id cpp/improper-null-termination
  * @problem.severity warning
+ * @precision medium
  * @security-severity 7.8
  * @tags security
  *       external/cwe/cwe-170
@@ -53,6 +54,7 @@ class ImproperNullTerminationReachability extends StackVariableReachabilityWithR
   override predicate isBarrier(ControlFlowNode node, StackVariable v) {
     exprDefinition(v, node, _) or
     mayAddNullTerminator(node, v.getAnAccess()) or
+    node.(AddressOfExpr).getOperand() = v.getAnAccess() or // address taken
     isSinkActual(node, v) // only report first use
   }
 }

cpp/ql/src/Metrics/Internal/CallableExtents.ql

@@ -18,7 +18,7 @@ class RangeFunction extends Function {
    * The location spans column `startcolumn` of line `startline` to
    * column `endcolumn` of line `endline` in file `filepath`.
    * For more information, see
-   * [Locations](https://help.semmle.com/QL/learn-ql/ql/locations.html).
+   * [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
    */
   predicate hasLocationInfo(string path, int sl, int sc, int el, int ec) {
     super.getLocation().hasLocationInfo(path, sl, sc, _, _) and

cpp/ql/src/Security/CWE/CWE-078/ExecTainted.ql

@@ -3,10 +3,10 @@
  * @description Using user-supplied data in an OS command, without
  *              neutralizing special elements, can make code vulnerable
  *              to command injection.
- * @kind problem
+ * @kind path-problem
  * @problem.severity error
  * @security-severity 9.8
- * @precision low
+ * @precision high
  * @id cpp/command-line-injection
  * @tags security
  *       external/cwe/cwe-078
@@ -16,13 +16,204 @@
 import cpp
 import semmle.code.cpp.security.CommandExecution
 import semmle.code.cpp.security.Security
-import semmle.code.cpp.security.TaintTracking
+import semmle.code.cpp.valuenumbering.GlobalValueNumbering
+import semmle.code.cpp.ir.IR
+import semmle.code.cpp.ir.dataflow.TaintTracking
+import semmle.code.cpp.ir.dataflow.TaintTracking2
+import semmle.code.cpp.security.FlowSources
+import semmle.code.cpp.models.implementations.Strcat
 
-from Expr taintedArg, Expr taintSource, string taintCause, string callChain
+Expr sinkAsArgumentIndirection(DataFlow::Node sink) {
+  result =
+    sink.asOperand()
+        .(SideEffectOperand)
+        .getAddressOperand()
+        .getAnyDef()
+        .getUnconvertedResultExpression()
+}
+
+/**
+ * Holds if `fst` is a string that is used in a format or concatenation function resulting in `snd`,
+ * and is *not* placed at the start of the resulting string. This indicates that the author did not
+ * expect `fst` to control what program is run if the resulting string is eventually interpreted as
+ * a command line, for example as an argument to `system`.
+ */
+predicate interestingConcatenation(DataFlow::Node fst, DataFlow::Node snd) {
+  exists(FormattingFunctionCall call, int index, FormatLiteral literal |
+    sinkAsArgumentIndirection(fst) = call.getConversionArgument(index) and
+    snd.asDefiningArgument() = call.getOutputArgument(false) and
+    literal = call.getFormat() and
+    not literal.getConvSpecOffset(index) = 0 and
+    literal.getConversionChar(index) = ["s", "S"]
+  )
+  or
+  // strcat and friends
+  exists(StrcatFunction strcatFunc, CallInstruction call, ReadSideEffectInstruction rse |
+    call.getStaticCallTarget() = strcatFunc and
+    rse.getArgumentDef() = call.getArgument(strcatFunc.getParamSrc()) and
+    fst.asOperand() = rse.getSideEffectOperand() and
+    snd.asInstruction().(WriteSideEffectInstruction).getDestinationAddress() =
+      call.getArgument(strcatFunc.getParamDest())
+  )
+  or
+  exists(CallInstruction call, Operator op, ReadSideEffectInstruction rse |
+    call.getStaticCallTarget() = op and
+    op.hasQualifiedName("std", "operator+") and
+    op.getType().(UserType).hasQualifiedName("std", "basic_string") and
+    call.getArgument(1) = rse.getArgumentOperand().getAnyDef() and // left operand
+    fst.asOperand() = rse.getSideEffectOperand() and
+    call = snd.asInstruction()
+  )
+}
+
+class TaintToConcatenationConfiguration extends TaintTracking::Configuration {
+  TaintToConcatenationConfiguration() { this = "TaintToConcatenationConfiguration" }
+
+  override predicate isSource(DataFlow::Node source) { source instanceof FlowSource }
+
+  override predicate isSink(DataFlow::Node sink) { interestingConcatenation(sink, _) }
+
+  override predicate isSanitizer(DataFlow::Node node) {
+    node.asInstruction().getResultType() instanceof IntegralType
+    or
+    node.asInstruction().getResultType() instanceof FloatingPointType
+  }
+}
+
+class ExecTaintConfiguration extends TaintTracking2::Configuration {
+  ExecTaintConfiguration() { this = "ExecTaintConfiguration" }
+
+  override predicate isSource(DataFlow::Node source) {
+    exists(DataFlow::Node prevSink, TaintToConcatenationConfiguration conf |
+      conf.hasFlow(_, prevSink) and
+      interestingConcatenation(prevSink, source)
+    )
+  }
+
+  override predicate isSink(DataFlow::Node sink) {
+    shellCommand(sinkAsArgumentIndirection(sink), _)
+  }
+
+  override predicate isSanitizerOut(DataFlow::Node node) {
+    isSink(node) // Prevent duplicates along a call chain, since `shellCommand` will include wrappers
+  }
+}
+
+module StitchedPathGraph {
+  // There's a different PathNode class for each DataFlowImplN.qll, so we can't simply combine the
+  // PathGraph predicates directly. Instead, we use a newtype so there's a single type that
+  // contains both sets of PathNodes.
+  newtype TMergedPathNode =
+    TPathNode1(DataFlow::PathNode node) or
+    TPathNode2(DataFlow2::PathNode node)
+
+  // this wraps the toString and location predicates so we can use the merged node type in a
+  // selection
+  class MergedPathNode extends TMergedPathNode {
+    string toString() {
+      exists(DataFlow::PathNode n |
+        this = TPathNode1(n) and
+        result = n.toString()
+      )
+      or
+      exists(DataFlow2::PathNode n |
+        this = TPathNode2(n) and
+        result = n.toString()
+      )
+    }
+
+    DataFlow::Node getNode() {
+      exists(DataFlow::PathNode n |
+        this = TPathNode1(n) and
+        result = n.getNode()
+      )
+      or
+      exists(DataFlow2::PathNode n |
+        this = TPathNode2(n) and
+        result = n.getNode()
+      )
+    }
+
+    DataFlow::PathNode getPathNode1() { this = TPathNode1(result) }
+
+    DataFlow2::PathNode getPathNode2() { this = TPathNode2(result) }
+
+    predicate hasLocationInfo(
+      string filepath, int startline, int startcolumn, int endline, int endcolumn
+    ) {
+      exists(DataFlow::PathNode n |
+        this = TPathNode1(n) and
+        n.hasLocationInfo(filepath, startline, startcolumn, endline, endcolumn)
+      )
+      or
+      exists(DataFlow2::PathNode n |
+        this = TPathNode2(n) and
+        n.hasLocationInfo(filepath, startline, startcolumn, endline, endcolumn)
+      )
+    }
+  }
+
+  query predicate edges(MergedPathNode a, MergedPathNode b) {
+    exists(DataFlow::PathNode an, DataFlow::PathNode bn |
+      a = TPathNode1(an) and
+      b = TPathNode1(bn) and
+      DataFlow::PathGraph::edges(an, bn)
+    )
+    or
+    exists(DataFlow2::PathNode an, DataFlow2::PathNode bn |
+      a = TPathNode2(an) and
+      b = TPathNode2(bn) and
+      DataFlow2::PathGraph::edges(an, bn)
+    )
+    or
+    // This is where paths from the two configurations are connected. `interestingConcatenation`
+    // is the only thing in this module that's actually specific to the query - everything else is
+    // just using types and predicates from the DataFlow library.
+    interestingConcatenation(a.getNode(), b.getNode()) and
+    a instanceof TPathNode1 and
+    b instanceof TPathNode2
+  }
+
+  query predicate nodes(MergedPathNode mpn, string key, string val) {
+    // here we just need the union of the underlying `nodes` predicates
+    exists(DataFlow::PathNode n |
+      mpn = TPathNode1(n) and
+      DataFlow::PathGraph::nodes(n, key, val)
+    )
+    or
+    exists(DataFlow2::PathNode n |
+      mpn = TPathNode2(n) and
+      DataFlow2::PathGraph::nodes(n, key, val)
+    )
+  }
+
+  query predicate subpaths(
+    MergedPathNode arg, MergedPathNode par, MergedPathNode ret, MergedPathNode out
+  ) {
+    // just forward subpaths from the underlying libraries. This might be slightly awkward when
+    // the concatenation is deep in a call chain.
+    DataFlow::PathGraph::subpaths(arg.getPathNode1(), par.getPathNode1(), ret.getPathNode1(),
+      out.getPathNode1())
+    or
+    DataFlow2::PathGraph::subpaths(arg.getPathNode2(), par.getPathNode2(), ret.getPathNode2(),
+      out.getPathNode2())
+  }
+}
+
+import StitchedPathGraph
+
+from
+  DataFlow::PathNode sourceNode, DataFlow::PathNode concatSink, DataFlow2::PathNode concatSource,
+  DataFlow2::PathNode sinkNode, string taintCause, string callChain,
+  TaintToConcatenationConfiguration conf1, ExecTaintConfiguration conf2
 where
-  shellCommand(taintedArg, callChain) and
+  taintCause = sourceNode.getNode().(FlowSource).getSourceType() and
-  tainted(taintSource, taintedArg) and
+  conf1.hasFlowPath(sourceNode, concatSink) and
-  isUserInput(taintSource, taintCause)
+  interestingConcatenation(concatSink.getNode(), concatSource.getNode()) and // this loses call context
-select taintedArg,
+  conf2.hasFlowPath(concatSource, sinkNode) and
-  "This argument to an OS command is derived from $@ and then passed to " + callChain, taintSource,
+  shellCommand(sinkAsArgumentIndirection(sinkNode.getNode()), callChain)
-  "user input (" + taintCause + ")"
+select sinkAsArgumentIndirection(sinkNode.getNode()), TPathNode1(sourceNode).(MergedPathNode),
+  TPathNode2(sinkNode).(MergedPathNode),
+  "This argument to an OS command is derived from $@, dangerously concatenated into $@, and then passed to "
+    + callChain, sourceNode, "user input (" + taintCause + ")", concatSource,
+  concatSource.toString()

cpp/ql/src/Security/CWE/CWE-170/ImproperNullTerminationTainted.ql

@@ -5,6 +5,7 @@
  * @kind problem
  * @id cpp/user-controlled-null-termination-tainted
  * @problem.severity warning
+ * @precision medium
  * @security-severity 10.0
  * @tags security
  *       external/cwe/cwe-170

cpp/ql/src/Security/CWE/CWE-311/CleartextStorage.inc.qhelp

@@ -9,7 +9,7 @@ storage.</p>
 </overview>
 <recommendation>
 
-<p>Ensure that sensitive information is always encrypted before being stored, especially before writing to a file.
+<p>Ensure that sensitive information is always encrypted before being stored to a file or transmitted over the network.
 It may be wise to encrypt information before it is put into a buffer that may be readable in memory.</p>
 
 <p>In general, decrypt sensitive information only at the point where it is necessary for it to be used in

cpp/ql/src/Security/CWE/CWE-311/CleartextTransmission.qhelp

@@ -0,0 +1,5 @@
+<!DOCTYPE qhelp PUBLIC
+  "-//Semmle//qhelp//EN"
+  "qhelp.dtd">
+<qhelp>
+<include src="CleartextStorage.inc.qhelp" /></qhelp>

cpp/ql/src/Security/CWE/CWE-311/CleartextTransmission.ql

@@ -0,0 +1,124 @@
+/**
+ * @name Cleartext transmission of sensitive information
+ * @description Transmitting sensitive information across a network in
+ *              cleartext can expose it to an attacker.
+ * @kind path-problem
+ * @problem.severity warning
+ * @security-severity 7.5
+ * @precision medium
+ * @id cpp/cleartext-transmission
+ * @tags security
+ *       external/cwe/cwe-319
+ */
+
+import cpp
+import semmle.code.cpp.security.SensitiveExprs
+import semmle.code.cpp.dataflow.TaintTracking
+import semmle.code.cpp.models.interfaces.FlowSource
+import DataFlow::PathGraph
+
+/**
+ * A function call that sends or receives data over a network.
+ */
+abstract class NetworkSendRecv extends FunctionCall {
+  /**
+   * Gets the expression for the socket or similar object used for sending or
+   * receiving data (if any).
+   */
+  abstract Expr getSocketExpr();
+
+  /**
+   * Gets the expression for the buffer to be sent from / received into.
+   */
+  abstract Expr getDataExpr();
+}
+
+/**
+ * A function call that sends data over a network.
+ *
+ * note: functions such as `write` may be writing to a network source or a file. We could attempt to determine which, and sort results into `cpp/cleartext-transmission` and perhaps `cpp/cleartext-storage-file`. In practice it usually isn't very important which query reports a result as long as its reported exactly once.
+ */
+class NetworkSend extends NetworkSendRecv {
+  RemoteFlowSinkFunction target;
+
+  NetworkSend() { target = this.getTarget() }
+
+  override Expr getSocketExpr() {
+    exists(FunctionInput input, int arg |
+      target.hasSocketInput(input) and
+      input.isParameter(arg) and
+      result = this.getArgument(arg)
+    )
+  }
+
+  override Expr getDataExpr() {
+    exists(FunctionInput input, int arg |
+      target.hasRemoteFlowSink(input, _) and
+      input.isParameterDeref(arg) and
+      result = this.getArgument(arg)
+    )
+  }
+}
+
+/**
+ * A function call that receives data over a network.
+ */
+class NetworkRecv extends NetworkSendRecv {
+  RemoteFlowSourceFunction target;
+
+  NetworkRecv() { target = this.getTarget() }
+
+  override Expr getSocketExpr() {
+    exists(FunctionInput input, int arg |
+      target.hasSocketInput(input) and
+      input.isParameter(arg) and
+      result = this.getArgument(arg)
+    )
+  }
+
+  override Expr getDataExpr() {
+    exists(FunctionOutput output, int arg |
+      target.hasRemoteFlowSource(output, _) and
+      output.isParameterDeref(arg) and
+      result = this.getArgument(arg)
+    )
+  }
+}
+
+/**
+ * Taint flow from a sensitive expression to a network operation with data
+ * tainted by that expression.
+ */
+class SensitiveSendRecvConfiguration extends TaintTracking::Configuration {
+  SensitiveSendRecvConfiguration() { this = "SensitiveSendRecvConfiguration" }
+
+  override predicate isSource(DataFlow::Node source) { source.asExpr() instanceof SensitiveExpr }
+
+  override predicate isSink(DataFlow::Node sink) {
+    exists(NetworkSendRecv transmission |
+      sink.asExpr() = transmission.getDataExpr() and
+      // a zero socket descriptor is standard input, which is not interesting for this query.
+      not exists(Zero zero |
+        DataFlow::localFlow(DataFlow::exprNode(zero),
+          DataFlow::exprNode(transmission.getSocketExpr()))
+      )
+    )
+  }
+}
+
+from
+  SensitiveSendRecvConfiguration config, DataFlow::PathNode source, DataFlow::PathNode sink,
+  NetworkSendRecv transmission, string msg
+where
+  config.hasFlowPath(source, sink) and
+  sink.getNode().asExpr() = transmission.getDataExpr() and
+  if transmission instanceof NetworkSend
+  then
+    msg =
+      "This operation transmits '" + sink.toString() +
+        "', which may contain unencrypted sensitive data from $@"
+  else
+    msg =
+      "This operation receives into '" + sink.toString() +
+        "', which may put unencrypted sensitive data into $@"
+select transmission, source, sink, msg, source, source.getNode().asExpr().toString()

cpp/ql/src/Security/CWE/CWE-468/IncorrectPointerScalingCommon.qll

@@ -121,16 +121,14 @@ predicate exprSourceType(Expr use, Type sourceType, Location sourceLoc) {
           else
             if use instanceof CrementOperation
             then exprSourceType(use.(CrementOperation).getOperand(), sourceType, sourceLoc)
-            else
+            else (
               // Conversions are not in the AST, so ignore them.
-              if use instanceof Conversion
+              not use instanceof Conversion and
-              then none()
+              // Source expressions
-              else (
+              sourceType = use.getUnspecifiedType() and
-                // Source expressions
+              isPointerType(sourceType) and
-                sourceType = use.getUnspecifiedType() and
+              sourceLoc = use.getLocation()
-                isPointerType(sourceType) and
+            )
-                sourceLoc = use.getLocation()
-              )
 }
 
 /**

cpp/ql/src/Security/CWE/CWE-570/IncorrectAllocationErrorHandling.ql

@@ -4,10 +4,13 @@
  * @kind problem
  * @id cpp/incorrect-allocation-error-handling
  * @problem.severity warning
+ * @security-severity 7.5
  * @precision medium
  * @tags correctness
  *       security
  *       external/cwe/cwe-570
+ *       external/cwe/cwe-252
+ *       external/cwe/cwe-755
  */
 
 import cpp

cpp/ql/src/definitions.qll

@@ -24,7 +24,7 @@ class Top extends Element {
    * The location spans column `startcolumn` of line `startline` to
    * column `endcolumn` of line `endline` in file `filepath`.
    * For more information, see
-   * [Locations](https://help.semmle.com/QL/learn-ql/ql/locations.html).
+   * [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
    */
   pragma[noopt]
   final predicate hasLocationInfo(

cpp/ql/src/experimental/Security/CWE/CWE-1041/FindWrapperFunctions.cpp

@@ -0,0 +1,18 @@
+...
+int myFclose(FILE * fmy)
+{
+  if(!fclose(fmy)) {
+    fmy = NULL;
+    return 0;
+  }
+  return -1;
+}
+...
+  fe = fopen("myFile.txt", "wt"); 
+  ...
+  fclose(fe); // BAD
+...
+  fe = fopen("myFile.txt", "wt"); 
+  ...
+  myFclose(fe); // GOOD
+...

cpp/ql/src/experimental/Security/CWE/CWE-1041/FindWrapperFunctions.qhelp

@@ -0,0 +1,23 @@
+<!DOCTYPE qhelp PUBLIC
+  "-//Semmle//qhelp//EN"
+  "qhelp.dtd">
+<qhelp>
+<overview>
+<p>The presence of a shell function with additional check indicates the possible risks of the call. Use this check everywhere.</p>
+
+</overview>
+
+<example>
+<p>The following example demonstrates fallacious and fixed methods of using wrapper functions.</p>
+<sample src="FindWrapperFunctions.cpp" />
+
+</example>
+<references>
+
+<li>
+  CERT C Coding Standard:
+  <a href="https://wiki.sei.cmu.edu/confluence/display/java/JNI00-J.+Define+wrappers+around+native+methods">JNI00-J. Define wrappers around native methods</a>.
+</li>
+
+</references>
+</qhelp>

cpp/ql/src/experimental/Security/CWE/CWE-1041/FindWrapperFunctions.ql

@@ -0,0 +1,141 @@
+/**
+ * @name Missed opportunity to call wrapper function
+ * @description If a wrapper function is defined for a given function, any call to the given function should be via the wrapper function.
+ * @kind problem
+ * @id cpp/call-to-function-without-wrapper
+ * @problem.severity warning
+ * @precision medium
+ * @tags correctness
+ *       maintainability
+ *       security
+ *       external/cwe/cwe-1041
+ */
+
+import cpp
+import semmle.code.cpp.valuenumbering.GlobalValueNumbering
+import semmle.code.cpp.commons.Assertions
+
+/**
+ * A function call that is used in error situations (logging, throwing an exception, abnormal termination).
+ */
+class CallUsedToHandleErrors extends FunctionCall {
+  CallUsedToHandleErrors() {
+    // call that is known to not return
+    not exists(this.(ControlFlowNode).getASuccessor())
+    or
+    // call throwing an exception
+    exists(ThrowExpr tex | tex = this.(ControlFlowNode).getASuccessor())
+    or
+    // call logging a message, possibly an error
+    exists(FormattingFunction ff | ff = this.(ControlFlowNode).getASuccessor())
+    or
+    // enabling recursive search
+    exists(CallUsedToHandleErrors fr | getTarget() = fr.getEnclosingFunction())
+  }
+}
+
+/** Holds if the conditions for a call outside the wrapper function are met. */
+predicate conditionsOutsideWrapper(FunctionCall fcp) {
+  fcp.getNumberOfArguments() > 0 and
+  not exists(ConditionalStmt cdtmp | fcp.getEnclosingStmt().getParentStmt*() = cdtmp) and
+  not exists(Loop lptmp | fcp.getEnclosingStmt().getParentStmt*() = lptmp) and
+  not exists(ReturnStmt rttmp | fcp.getEnclosingStmt().getParentStmt*() = rttmp) and
+  not exists(FunctionCall fctmp2 | fcp = fctmp2.getAnArgument().getAChild*()) and
+  not exists(Assignment astmp | fcp = astmp.getRValue().getAChild*()) and
+  not exists(Initializer intmp | fcp = intmp.getExpr().getAChild*()) and
+  not exists(Assertion astmp | fcp = astmp.getAsserted().getAChild*()) and
+  not exists(Operation optmp | fcp = optmp.getAChild*()) and
+  not exists(ArrayExpr aetmp | fcp = aetmp.getAChild*()) and
+  not exists(ExprCall ectmp | fcp = ectmp.getAnArgument().getAChild*())
+}
+
+/** Holds if the conditions for calling `fcp` inside the `fnp` wrapper function are met. */
+pragma[inline]
+predicate conditionsInsideWrapper(FunctionCall fcp, Function fnp) {
+  not exists(FunctionCall fctmp2 |
+    fctmp2.getEnclosingFunction() = fnp and fcp = fctmp2.getAnArgument().getAChild*()
+  ) and
+  not fcp instanceof CallUsedToHandleErrors and
+  not fcp.getAnArgument().isConstant() and
+  fcp.getEnclosingFunction() = fnp and
+  fnp.getNumberOfParameters() > 0 and
+  // the call arguments must be passed through the arguments of the wrapper function
+  forall(int i | i in [0 .. fcp.getNumberOfArguments() - 1] |
+    globalValueNumber(fcp.getArgument(i)) = globalValueNumber(fnp.getAParameter().getAnAccess())
+  ) and
+  // there should be no more than one required call inside the wrapper function
+  not exists(FunctionCall fctmp |
+    fctmp.getTarget() = fcp.getTarget() and
+    fctmp.getFile() = fcp.getFile() and
+    fctmp != fcp and
+    fctmp.getEnclosingFunction() = fnp
+  ) and
+  // inside the wrapper function there should be no calls without paths to the desired function
+  not exists(FunctionCall fctmp |
+    fctmp.getEnclosingFunction() = fnp and
+    fctmp.getFile() = fcp.getFile() and
+    fctmp != fcp and
+    (
+      fctmp = fcp.getAPredecessor+()
+      or
+      not exists(FunctionCall fctmp1 |
+        fctmp1 = fcp and
+        (
+          fctmp.getASuccessor+() = fctmp1 or
+          fctmp.getAPredecessor+() = fctmp1
+        )
+      )
+    )
+  )
+}
+
+/** Holds if the conditions for the wrapper function are met. */
+pragma[inline]
+predicate conditionsForWrapper(FunctionCall fcp, Function fnp) {
+  not exists(ExprCall ectmp | fnp = ectmp.getEnclosingFunction()) and
+  not exists(Loop lp | lp.getEnclosingFunction() = fnp) and
+  not exists(SwitchStmt sw | sw.getEnclosingFunction() = fnp) and
+  not fnp instanceof Operator and
+  // inside the wrapper function there should be checks of arguments or the result,
+  // perhaps by means of passing the latter as an argument to some function
+  (
+    exists(IfStmt ifs |
+      ifs.getEnclosingFunction() = fnp and
+      (
+        globalValueNumber(ifs.getCondition().getAChild*()) = globalValueNumber(fcp.getAnArgument()) and
+        ifs.getASuccessor*() = fcp
+        or
+        ifs.getCondition().getAChild() = fcp
+      )
+    )
+    or
+    exists(FunctionCall fctmp |
+      fctmp.getEnclosingFunction() = fnp and
+      globalValueNumber(fctmp.getAnArgument().getAChild*()) = globalValueNumber(fcp)
+    )
+  ) and
+  // inside the wrapper function there must be a function call to handle the error
+  exists(CallUsedToHandleErrors fctmp |
+    fctmp.getEnclosingFunction() = fnp and
+    forall(int i | i in [0 .. fnp.getNumberOfParameters() - 1] |
+      fnp.getParameter(i).getAnAccess().getTarget() =
+        fcp.getAnArgument().(VariableAccess).getTarget() or
+      fnp.getParameter(i).getUnspecifiedType() instanceof Class or
+      fnp.getParameter(i).getUnspecifiedType().(ReferenceType).getBaseType() instanceof Class or
+      fnp.getParameter(i).getAnAccess().getTarget() =
+        fctmp.getAnArgument().(VariableAccess).getTarget()
+    )
+  )
+}
+
+from FunctionCall fc, Function fn
+where
+  exists(FunctionCall fctmp |
+    conditionsInsideWrapper(fctmp, fn) and
+    conditionsForWrapper(fctmp, fn) and
+    conditionsOutsideWrapper(fc) and
+    fctmp.getTarget() = fc.getTarget() and
+    fc.getEnclosingFunction() != fn and
+    fc.getEnclosingFunction().getMetrics().getNumberOfCalls() > fn.getMetrics().getNumberOfCalls()
+  )
+select fc, "Consider changing the call to $@", fn, fn.getName()

cpp/ql/src/external/CodeDuplication.qll

@@ -61,7 +61,7 @@ class Copy extends @duplication_or_similarity {
    * The location spans column `startcolumn` of line `startline` to
    * column `endcolumn` of line `endline` in file `filepath`.
    * For more information, see
-   * [Locations](https://help.semmle.com/QL/learn-ql/ql/locations.html).
+   * [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
    */
   predicate hasLocationInfo(
     string filepath, int startline, int startcolumn, int endline, int endcolumn

cpp/ql/src/external/DefectFilter.qll

@@ -8,7 +8,7 @@ import cpp
  * column `startcolumn` of line `startline` to column `endcolumn` of line `endline`
  * in file `filepath`.
  *
- * For more information, see [Locations](https://help.semmle.com/QL/learn-ql/ql/locations.html).
+ * For more information, see [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
  */
 external predicate defectResults(
   int id, string queryPath, string file, int startline, int startcol, int endline, int endcol,

cpp/ql/src/external/MetricFilter.qll

@@ -8,7 +8,7 @@ import cpp
  * column `startcolumn` of line `startline` to column `endcolumn` of line `endline`
  * in file `filepath`.
  *
- * For more information, see [Locations](https://help.semmle.com/QL/learn-ql/ql/locations.html).
+ * For more information, see [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
  */
 external predicate metricResults(
   int id, string queryPath, string file, int startline, int startcol, int endline, int endcol,

cpp/ql/src/jsf/4.10 Classes/AV Rule 79.ql

@@ -230,13 +230,13 @@ predicate leakedInSameMethod(Resource r, Expr acquire) {
         )
       )
     )
-    or
+  )
-    exists(FunctionAccess fa, string kind |
+  or
-      // the address of a function that releases `r` is taken (and likely
+  exists(FunctionAccess fa, string kind |
-      // used to release `r` at some point).
+    // the address of a function that releases `r` is taken (and likely
-      r.acquisitionWithRequiredKind(acquire, kind) and
+    // used to release `r` at some point).
-      fa.getTarget() = r.getAReleaseExpr(kind).getEnclosingFunction()
+    r.acquisitionWithRequiredKind(acquire, kind) and
-    )
+    fa.getTarget() = r.getAReleaseExpr(kind).getEnclosingFunction()
   )
 }
 

cpp/ql/test/TestUtilities/InlineExpectationsTest.qll

@@ -4,7 +4,7 @@
  * (in that the `.expected` file should always be empty).
  *
  * To add this framework to a new language:
- * - Add a file `InlineExpectationsTestPrivate.qll` that defines a `LineComment` class. This class
+ * - Add a file `InlineExpectationsTestPrivate.qll` that defines a `ExpectationComment` class. This class
  *   must support a `getContents` method that returns the contents of the given comment, _excluding_
  *   the comment indicator itself. It should also define `toString` and `getLocation` as usual.
  *
@@ -60,8 +60,8 @@
  *
  * Example:
  * ```cpp
- * int i = x + 5;  // $const=5
+ * int i = x + 5;  // $ const=5
- * int j = y + (7 - 3)  // $const=7 const=3 const=4  // The result of the subtraction is a constant.
+ * int j = y + (7 - 3)  // $ const=7 const=3 const=4  // The result of the subtraction is a constant.
  * ```
  *
  * For tests that contain known missing and spurious results, it is possible to further
@@ -194,7 +194,7 @@ private int getEndOfColumnPosition(int start, string content) {
 }
 
 private predicate getAnExpectation(
-  LineComment comment, TColumn column, string expectation, string tags, string value
+  ExpectationComment comment, TColumn column, string expectation, string tags, string value
 ) {
   exists(string content |
     content = comment.getContents().regexpCapture(expectationCommentPattern(), 1) and
@@ -247,14 +247,14 @@ private newtype TFailureLocatable =
   ) {
     test.hasActualResult(location, element, tag, value)
   } or
-  TValidExpectation(LineComment comment, string tag, string value, string knownFailure) {
+  TValidExpectation(ExpectationComment comment, string tag, string value, string knownFailure) {
     exists(TColumn column, string tags |
       getAnExpectation(comment, column, _, tags, value) and
       tag = tags.splitAt(",") and
       knownFailure = getColumnString(column)
     )
   } or
-  TInvalidExpectation(LineComment comment, string expectation) {
+  TInvalidExpectation(ExpectationComment comment, string expectation) {
     getAnExpectation(comment, _, expectation, _, _) and
     not expectation.regexpMatch(expectationPattern())
   }
@@ -292,7 +292,7 @@ class ActualResult extends FailureLocatable, TActualResult {
 }
 
 abstract private class Expectation extends FailureLocatable {
-  LineComment comment;
+  ExpectationComment comment;
 
   override string toString() { result = comment.toString() }
 

cpp/ql/test/TestUtilities/InlineExpectationsTestPrivate.qll

@@ -1,16 +1,16 @@
 import cpp
 
-private newtype TLineComment = MkLineComment(CppStyleComment c)
+private newtype TExpectationComment = MkExpectationComment(CppStyleComment c)
 
 /**
  * Represents a line comment in the CPP style.
  * Unlike the `CppStyleComment` class, however, the string returned by `getContents` does _not_
  * include the preceding comment marker (`//`).
  */
-class LineComment extends TLineComment {
+class ExpectationComment extends TExpectationComment {
   CppStyleComment comment;
 
-  LineComment() { this = MkLineComment(comment) }
+  ExpectationComment() { this = MkExpectationComment(comment) }
 
   /** Returns the contents of the given comment, _without_ the preceding comment marker (`//`). */
   string getContents() { result = comment.getContents().suffix(2) }

cpp/ql/test/experimental/query-tests/Security/CWE/CWE-1041/semmle/tests/FindWrapperFunctions.expected

@@ -0,0 +1 @@
+| test.cpp:23:3:23:8 | call to fclose | Consider changing the call to $@ | test.cpp:9:6:9:13 | myFclose | myFclose |

cpp/ql/test/experimental/query-tests/Security/CWE/CWE-1041/semmle/tests/FindWrapperFunctions.qlref

@@ -0,0 +1 @@
+experimental/Security/CWE/CWE-1041/FindWrapperFunctions.ql

cpp/ql/test/experimental/query-tests/Security/CWE/CWE-1041/semmle/tests/test.cpp

@@ -0,0 +1,27 @@
+#define NULL (0)
+typedef int FILE;
+FILE *fopen(const char *filename, const char *mode);
+int fclose(FILE *stream);
+extern FILE * fe;
+extern int printf(const char *fmt, ...);
+void exit(int status);
+
+void myFclose(FILE * fmy)
+{
+  int i;
+  if(fmy) {
+    i = fclose(fmy);
+    fmy = NULL;
+    printf("close end is code %d",i);
+    if(i!=0) exit(1);
+  }
+}
+
+int main(int argc, char *argv[])
+{
+  fe = fopen("myFile.txt", "wt");
+  fclose(fe); // BAD
+  fe = fopen("myFile.txt", "wt");
+  myFclose(fe); // GOOD
+  return 0;
+}

cpp/ql/test/include/iterator.h

@@ -0,0 +1,97 @@
+#if !defined(CODEQL_ITERATOR_H)
+#define CODEQL_ITERATOR_H
+
+typedef unsigned long size_t;
+
+#include "type_traits.h"
+
+namespace std {
+	struct ptrdiff_t;
+
+	template<class I> struct iterator_traits;
+
+	template <class Category,
+			  class value_type,
+			  class difference_type = ptrdiff_t,
+			  class pointer_type = value_type*,
+			  class reference_type = value_type&>
+	struct iterator {
+		typedef Category iterator_category;
+
+		iterator();
+		iterator(iterator<Category, remove_const_t<value_type> > const &other); // non-const -> const conversion constructor
+
+		iterator &operator++();
+		iterator operator++(int);
+		iterator &operator--();
+		iterator operator--(int);
+		bool operator==(iterator other) const;
+		bool operator!=(iterator other) const;
+		reference_type operator*() const;
+		pointer_type operator->() const;
+		iterator operator+(int);
+		iterator operator-(int);
+		iterator &operator+=(int);
+		iterator &operator-=(int);
+		int operator-(iterator);
+		reference_type operator[](int);
+	};
+
+	struct input_iterator_tag {};
+	struct forward_iterator_tag : public input_iterator_tag {};
+	struct bidirectional_iterator_tag : public forward_iterator_tag {};
+	struct random_access_iterator_tag : public bidirectional_iterator_tag {};
+
+	struct output_iterator_tag {};
+
+	template<class Container>
+	class back_insert_iterator {
+	protected:
+		Container* container = nullptr;
+	public:
+		using iterator_category = output_iterator_tag;
+		using value_type = void;
+		using difference_type = ptrdiff_t;
+		using pointer = void;
+		using reference = void;
+		using container_type = Container;
+		constexpr back_insert_iterator() noexcept = default;
+		constexpr explicit back_insert_iterator(Container& x);
+		back_insert_iterator& operator=(const typename Container::value_type& value);
+		back_insert_iterator& operator=(typename Container::value_type&& value);
+		back_insert_iterator& operator*();
+		back_insert_iterator& operator++();
+		back_insert_iterator operator++(int);
+	};
+
+	template<class Container>
+	constexpr back_insert_iterator<Container> back_inserter(Container& x) {
+		return back_insert_iterator<Container>(x);
+	}
+
+	template<class Container>
+	class front_insert_iterator {
+	protected:
+		Container* container = nullptr;
+	public:
+		using iterator_category = output_iterator_tag;
+		using value_type = void;
+		using difference_type = ptrdiff_t;
+		using pointer = void;
+		using reference = void;
+		using container_type = Container;
+		constexpr front_insert_iterator() noexcept = default;
+		constexpr explicit front_insert_iterator(Container& x);
+		constexpr front_insert_iterator& operator=(const typename Container::value_type& value);
+		constexpr front_insert_iterator& operator=(typename Container::value_type&& value);
+		constexpr front_insert_iterator& operator*();
+		constexpr front_insert_iterator& operator++();
+		constexpr front_insert_iterator operator++(int);
+	};
+	template<class Container>
+	constexpr front_insert_iterator<Container> front_inserter(Container& x) {
+		return front_insert_iterator<Container>(x);
+	}
+}
+
+#endif