workaround.

Kotlin: Fix for 2.1.0-Beta2

.bazelrc

@@ -25,5 +25,6 @@ common --registry=https://bcr.bazel.build
 
 common --@rules_dotnet//dotnet/settings:strict_deps=false
 common --experimental_isolated_extension_usages
+common --incompatible_use_plus_in_repo_names
 
 try-import %workspace%/local.bazelrc

.github/workflows/go-tests-other-os.yml

@@ -10,26 +10,35 @@ on:
       - MODULE.bazel
       - .bazelrc
       - misc/bazel/**
-
+  push:
+    branches:
+      - criemen/debug
 permissions:
   contents: read
 
 jobs:
-  test-mac:
-    name: Test MacOS
-    runs-on: macos-latest
-    steps:
-      - name: Check out code
-        uses: actions/checkout@v4
-      - name: Run tests
-        uses: ./go/actions/test
-
   test-win:
-    if: github.repository_owner == 'github'
     name: Test Windows
-    runs-on: windows-latest-xl
+    runs-on: windows-latest
     steps:
       - name: Check out code
         uses: actions/checkout@v4
-      - name: Run tests
+        with:
-        uses: ./go/actions/test
+          sparse-checkout: go
+
+      - name: dump env bash
+        shell: bash
+        run: |
+          set -x
+          cd go
+          type -a bazel
+          where bazel
+          echo $PATH
+          bazel version
+      - name: Test
+        shell: bash
+        run: |
+          cd go
+          make gen
+      # - name: Run tests
+      #   uses: ./go/actions/test

Cargo.lock

@@ -112,6 +112,16 @@ version = "0.7.6"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "7c02d123df017efcdfbd739ef81735b36c5ba83ec3c59c80a9d7ecc718f92e50"
 
+[[package]]
+name = "ast-generator"
+version = "0.1.0"
+dependencies = [
+ "itertools 0.10.5",
+ "proc-macro2",
+ "quote",
+ "ungrammar",
+]
+
 [[package]]
 name = "atomic"
 version = "0.6.0"
@@ -380,11 +390,13 @@ dependencies = [
  "ra_ap_base_db",
  "ra_ap_hir",
  "ra_ap_hir_def",
+ "ra_ap_hir_expand",
  "ra_ap_ide_db",
  "ra_ap_load-cargo",
  "ra_ap_parser",
  "ra_ap_paths",
  "ra_ap_project_model",
+ "ra_ap_span",
  "ra_ap_syntax",
  "ra_ap_vfs",
  "rust-extractor-macros",
@@ -681,16 +693,6 @@ version = "0.4.7"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "7ab85b9b05e3978cc9a9cf8fea7f01b494e1a09ed3037e16ba39edc7a29eb61a"
 
-[[package]]
-name = "generate-schema"
-version = "0.1.0"
-dependencies = [
- "itertools 0.10.5",
- "proc-macro2",
- "quote",
- "ungrammar",
-]
-
 [[package]]
 name = "getrandom"
 version = "0.2.15"

Cargo.toml

@@ -7,7 +7,7 @@ members = [
     "ruby/extractor",
     "rust/extractor",
     "rust/extractor/macros",
-    "rust/generate-schema",
+    "rust/ast-generator",
 ]
 
 [patch.crates-io]

MODULE.bazel

@@ -23,11 +23,11 @@ bazel_dep(name = "bazel_skylib", version = "1.6.1")
 bazel_dep(name = "abseil-cpp", version = "20240116.0", repo_name = "absl")
 bazel_dep(name = "nlohmann_json", version = "3.11.3", repo_name = "json")
 bazel_dep(name = "fmt", version = "10.0.0")
-bazel_dep(name = "rules_kotlin", version = "1.9.4-codeql.1")
+bazel_dep(name = "rules_kotlin", version = "2.0.0-codeql.1")
 bazel_dep(name = "gazelle", version = "0.38.0")
 bazel_dep(name = "rules_dotnet", version = "0.15.1")
 bazel_dep(name = "googletest", version = "1.14.0.bcr.1")
-bazel_dep(name = "rules_rust", version = "0.50.0")
+bazel_dep(name = "rules_rust", version = "0.52.2")
 
 bazel_dep(name = "buildifier_prebuilt", version = "6.4.0", dev_dependency = True)
 
@@ -61,7 +61,7 @@ r.from_cargo(
         "//ruby/extractor:Cargo.toml",
         "//rust/extractor:Cargo.toml",
         "//rust/extractor/macros:Cargo.toml",
-        "//rust/generate-schema:Cargo.toml",
+        "//rust/ast-generator:Cargo.toml",
         "//shared/tree-sitter-extractor:Cargo.toml",
     ],
 )

cpp/downgrades/6f5d51e89e762fe4609fd4ac8ee3afb04221e873/exprs.ql

@@ -0,0 +1,15 @@
+class Expr extends @expr {
+  string toString() { none() }
+}
+
+class Location extends @location_expr {
+  string toString() { none() }
+}
+
+predicate isExprRequires(Expr expr) { exists(int kind | exprs(expr, kind, _) | kind = 390) }
+
+from Expr expr, int kind, int kind_new, Location location
+where
+  exprs(expr, kind, location) and
+  if isExprRequires(expr) then kind_new = 1 else kind_new = kind
+select expr, kind_new, location

cpp/downgrades/6f5d51e89e762fe4609fd4ac8ee3afb04221e873/upgrade.properties

@@ -0,0 +1,3 @@
+description: Add requires expr
+compatibility: partial
+exprs.rel: run exprs.qlo

cpp/downgrades/e51fad7a2436caefab0c6bd52f05e28e7cce4d92/exprs.ql

@@ -0,0 +1,17 @@
+class Expr extends @expr {
+  string toString() { none() }
+}
+
+class Location extends @location_expr {
+  string toString() { none() }
+}
+
+predicate isExprRequirement(Expr expr) {
+  exists(int kind | exprs(expr, kind, _) | kind = [391, 392, 393])
+}
+
+from Expr expr, int kind, int kind_new, Location location
+where
+  exprs(expr, kind, location) and
+  if isExprRequirement(expr) then kind_new = 1 else kind_new = kind
+select expr, kind_new, location

cpp/downgrades/e51fad7a2436caefab0c6bd52f05e28e7cce4d92/params.ql

@@ -0,0 +1,17 @@
+class Parameter extends @parameter {
+  string toString() { none() }
+}
+
+class ParameterizedElement extends @parameterized_element {
+  string toString() { none() }
+}
+
+class Type extends @type {
+  string toString() { none() }
+}
+
+from Parameter param, ParameterizedElement pe, int index, Type type
+where
+  params(param, pe, index, type) and
+  not pe instanceof @requires_expr
+select param, pe, index, type

cpp/downgrades/e51fad7a2436caefab0c6bd52f05e28e7cce4d92/upgrade.properties

@@ -0,0 +1,5 @@
+description: Support C++20 requires expressions
+compatibility: partial
+compound_requirement_is_noexcept.rel: delete
+exprs.rel: run exprs.qlo
+params.rel: run params.qlo

cpp/ql/lib/CHANGELOG.md

@@ -1,3 +1,10 @@
+## 2.0.2
+
+### Minor Analysis Improvements
+
+* Added taint flow model for `fopen` and related functions.
+* The `SimpleRangeAnalysis` library (`semmle.code.cpp.rangeanalysis.SimpleRangeAnalysis`) now generates more precise ranges for calls to `fgetc` and `getc`.
+
 ## 2.0.1
 
 No user-facing changes.

cpp/ql/lib/change-notes/2014-10-11-requires-expressions.md

@@ -0,0 +1,4 @@
+---
+category: feature
+---
+* Added classes `RequiresExpr`, `SimpleRequirementExpr`, `TypeRequirementExpr`, `CompoundRequirementExpr`, and `NestedRequirementExpr` to represent C++20 requires expressions and the simple, type, compound, and nested requirements that can occur in `requires` expressions.

cpp/ql/lib/change-notes/released/2.0.2.md

@@ -0,0 +1,6 @@
+## 2.0.2
+
+### Minor Analysis Improvements
+
+* Added taint flow model for `fopen` and related functions.
+* The `SimpleRangeAnalysis` library (`semmle.code.cpp.rangeanalysis.SimpleRangeAnalysis`) now generates more precise ranges for calls to `fgetc` and `getc`.

cpp/ql/lib/codeql-pack.release.yml

@@ -1,2 +1,2 @@
 ---
-lastReleaseVersion: 2.0.1
+lastReleaseVersion: 2.0.2

cpp/ql/lib/cpp.qll

@@ -17,6 +17,7 @@ import semmle.code.cpp.File
 import semmle.code.cpp.Linkage
 import semmle.code.cpp.Location
 import semmle.code.cpp.Compilation
+import semmle.code.cpp.Concept
 import semmle.code.cpp.Element
 import semmle.code.cpp.Namespace
 import semmle.code.cpp.Specifier

cpp/ql/lib/qlpack.yml

@@ -1,5 +1,5 @@
 name: codeql/cpp-all
-version: 2.0.1
+version: 2.0.3-dev
 groups: cpp
 dbscheme: semmlecode.cpp.dbscheme
 extractor: cpp

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

@@ -0,0 +1,161 @@
+/**
+ * Provides classes for working with C++ concepts.
+ */
+
+import semmle.code.cpp.exprs.Expr
+
+/**
+ * A C++ requires expression.
+ *
+ * For example, with `T` and `U` template parameters:
+ * ```cpp
+ * requires (T x, U y) { x + y; };
+ * ```
+ */
+class RequiresExpr extends Expr, @requires_expr {
+  override string toString() {
+    if exists(this.getAParameter())
+    then result = "requires(...) { ... }"
+    else result = "requires { ... }"
+  }
+
+  override string getAPrimaryQlClass() { result = "RequiresExpr" }
+
+  /**
+   * Gets a requirement in this requires expression.
+   */
+  RequirementExpr getARequirement() { result = this.getAChild() }
+
+  /**
+   * Gets the nth requirement in this requires expression.
+   */
+  RequirementExpr getRequirement(int n) { result = this.getChild(n) }
+
+  /**
+   * Gets the number of requirements in this requires expression.
+   */
+  int getNumberOfRequirements() { result = count(this.getARequirement()) }
+
+  /**
+   * Gets a parameter of this requires expression, if any.
+   */
+  Parameter getAParameter() { result.getRequiresExpr() = underlyingElement(this) }
+
+  /**
+   * Gets the the nth parameter of this requires expression.
+   */
+  Parameter getParameter(int n) {
+    result.getRequiresExpr() = underlyingElement(this) and result.getIndex() = n
+  }
+
+  /**
+   * Gets the number of parameters of this requires expression.
+   */
+  int getNumberOfParameters() { result = count(this.getAParameter()) }
+}
+
+/**
+ * A C++ requirement in a requires expression.
+ */
+class RequirementExpr extends Expr { }
+
+/**
+ * A C++ simple requirement in a requires expression.
+ *
+ * For example, if:
+ * ```cpp
+ * requires(T x, U y) { x + y; };
+ * ```
+ * with `T` and `U` template parameters, then `x + y;` is a simple requirement.
+ */
+class SimpleRequirementExpr extends RequirementExpr {
+  SimpleRequirementExpr() {
+    this.getParent() instanceof RequiresExpr and
+    not this instanceof TypeRequirementExpr and
+    not this instanceof CompoundRequirementExpr and
+    not this instanceof NestedRequirementExpr
+  }
+
+  override string getAPrimaryQlClass() { result = "SimpleRequirementExpr" }
+}
+
+/**
+ * A C++ type requirement in a requires expression.
+ *
+ * For example, if:
+ * ```cpp
+ * requires { typename T::a_field; };
+ * ```
+ * with `T` a template parameter, then `typename T::a_field;` is a type requirement.
+ */
+class TypeRequirementExpr extends RequirementExpr, TypeName {
+  TypeRequirementExpr() { this.getParent() instanceof RequiresExpr }
+
+  override string getAPrimaryQlClass() { result = "TypeRequirementExpr" }
+}
+
+/**
+ * A C++ compound requirement in a requires expression.
+ *
+ * For example, if:
+ * ```cpp
+ * requires(T x) { { x } noexcept -> std::same_as<int>; };
+ * ```
+ * with `T` a template parameter, then `{ x } noexcept -> std::same_as<int>;` is
+ * a compound requirement.
+ */
+class CompoundRequirementExpr extends RequirementExpr, @compound_requirement {
+  override string toString() {
+    if exists(this.getReturnTypeRequirement())
+    then result = "{ ... } -> ..."
+    else result = "{ ... }"
+  }
+
+  override string getAPrimaryQlClass() { result = "CompoundRequirementExpr" }
+
+  /**
+   * Gets the expression from the compound requirement.
+   */
+  Expr getExpr() { result = this.getChild(0) }
+
+  /**
+   * Gets the return type requirement from the compound requirement, if any.
+   */
+  Expr getReturnTypeRequirement() { result = this.getChild(1) }
+
+  /**
+   * Holds if the expression from the compound requirement must not be
+   * potentially throwing.
+   */
+  predicate isNoExcept() { compound_requirement_is_noexcept(underlyingElement(this)) }
+}
+
+/**
+ * A C++ nested requirement in a requires expression.
+ *
+ * For example, if:
+ * ```cpp
+ * requires { requires std::is_same<T, int>::value; };
+ * ```
+ * with `T` a template parameter, then `requires std::is_same<T, int>::value;` is
+ * a nested requirement.
+ */
+class NestedRequirementExpr extends Expr, @nested_requirement {
+  override string toString() { result = "requires ..." }
+
+  override string getAPrimaryQlClass() { result = "NestedRequirementExpr" }
+
+  /**
+   * Gets the constraint from the nested requirement.
+   */
+  Expr getConstraint() { result = this.getChild(0) }
+}
+
+/**
+ * A C++ concept id expression.
+ */
+class ConceptIdExpr extends RequirementExpr, @concept_id {
+  override string toString() { result = "concept<...>" }
+
+  override string getAPrimaryQlClass() { result = "ConceptIdExpr" }
+}

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

@@ -500,6 +500,17 @@ class Function extends Declaration, ControlFlowNode, AccessHolder, @function {
    * Gets the nearest enclosing AccessHolder.
    */
   override AccessHolder getEnclosingAccessHolder() { result = this.getDeclaringType() }
+
+  /**
+   * Holds if this function has extraction errors that create an `ErrorExpr`.
+   */
+  predicate hasErrors() {
+    exists(ErrorExpr e |
+      e.getEnclosingFunction() = this and
+      // Exclude the first allocator call argument because it is always extracted as `ErrorExpr`.
+      not exists(NewOrNewArrayExpr new | e = new.getAllocatorCall().getArgument(0))
+    )
+  }
 }
 
 pragma[noinline]

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

@@ -7,8 +7,8 @@ import semmle.code.cpp.Declaration
 private import semmle.code.cpp.internal.ResolveClass
 
 /**
- * A C/C++ function parameter or catch block parameter. For example the
+ * A C/C++ function parameter, catch block parameter, or requires expression parameter.
- * function parameter `p` and the catch block parameter `e` in the following
+ * For example the function parameter `p` and the catch block parameter `e` in the following
  * code:
  * ```
  * void myFunction(int p) {
@@ -20,8 +20,8 @@ private import semmle.code.cpp.internal.ResolveClass
  * }
  * ```
  *
- * For catch block parameters, there is a one-to-one correspondence between
+ * For catch block parameters and expression , there is a one-to-one
- * the `Parameter` and its `ParameterDeclarationEntry`.
+ * correspondence between the `Parameter` and its `VariableDeclarationEntry`.
  *
  * For function parameters, there is a one-to-many relationship between
  * `Parameter` and `ParameterDeclarationEntry`, because one function can
@@ -118,6 +118,12 @@ class Parameter extends LocalScopeVariable, @parameter {
    */
   BlockStmt getCatchBlock() { params(underlyingElement(this), unresolveElement(result), _, _) }
 
+  /**
+   * Gets the requires expression to which the parameter belongs, if it is a
+   * requires expression parameter.
+   */
+  RequiresExpr getRequiresExpr() { params(underlyingElement(this), unresolveElement(result), _, _) }
+
   /**
    * Gets the zero-based index of this parameter.
    *

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

@@ -80,6 +80,10 @@ private Declaration getAnEnclosingDeclaration(Locatable ast) {
   or
   result = ast.(Parameter).getFunction()
   or
+  result = ast.(Parameter).getCatchBlock().getEnclosingFunction()
+  or
+  result = ast.(Parameter).getRequiresExpr().getEnclosingFunction()
+  or
   result = ast.(Expr).getEnclosingDeclaration()
   or
   result = ast.(Initializer).getDeclaration()
@@ -99,7 +103,10 @@ private newtype TPrintAstNode =
     stmt.getADeclarationEntry() = entry and
     shouldPrintDeclaration(stmt.getEnclosingFunction())
   } or
-  TParametersNode(Function func) { shouldPrintDeclaration(func) } or
+  TFunctionParametersNode(Function func) { shouldPrintDeclaration(func) } or
+  TRequiresExprParametersNode(RequiresExpr req) {
+    shouldPrintDeclaration(getAnEnclosingDeclaration(req))
+  } or
   TConstructorInitializersNode(Constructor ctor) {
     ctor.hasEntryPoint() and
     shouldPrintDeclaration(ctor)
@@ -303,14 +310,14 @@ class ExprNode extends AstNode {
 
   ExprNode() { expr = ast }
 
-  override AstNode getChildInternal(int childIndex) {
+  override PrintAstNode getChildInternal(int childIndex) {
-    result.getAst() = expr.getChild(childIndex)
+    result.(AstNode).getAst() = expr.getChild(childIndex)
     or
     childIndex = max(int index | exists(expr.getChild(index)) or index = 0) + 1 and
-    result.getAst() = expr.(ConditionDeclExpr).getInitializingExpr()
+    result.(AstNode).getAst() = expr.(ConditionDeclExpr).getInitializingExpr()
     or
     exists(int destructorIndex |
-      result.getAst() = expr.getImplicitDestructorCall(destructorIndex) and
+      result.(AstNode).getAst() = expr.getImplicitDestructorCall(destructorIndex) and
       childIndex = destructorIndex + max(int index | exists(expr.getChild(index)) or index = 0) + 2
     )
   }
@@ -329,7 +336,8 @@ class ExprNode extends AstNode {
   }
 
   override string getChildAccessorPredicateInternal(int childIndex) {
-    result = getChildAccessorWithoutConversions(ast, this.getChildInternal(childIndex).getAst())
+    result =
+      getChildAccessorWithoutConversions(ast, this.getChildInternal(childIndex).(AstNode).getAst())
   }
 
   /**
@@ -409,6 +417,26 @@ class StmtExprNode extends ExprNode {
   }
 }
 
+/**
+ * A node representing a `RequiresExpr`
+ */
+class RequiresExprNode extends ExprNode {
+  override RequiresExpr expr;
+
+  override PrintAstNode getChildInternal(int childIndex) {
+    result = super.getChildInternal(childIndex)
+    or
+    childIndex = -1 and
+    result.(RequiresExprParametersNode).getRequiresExpr() = expr
+  }
+
+  override string getChildAccessorPredicateInternal(int childIndex) {
+    result = super.getChildAccessorPredicateInternal(childIndex)
+    or
+    childIndex = -1 and result = "<params>"
+  }
+}
+
 /**
  * A node representing a `DeclarationEntry`.
  */
@@ -510,6 +538,22 @@ class DeclStmtNode extends StmtNode {
   }
 }
 
+/**
+ * A node representing a `Handler`.
+ */
+class HandlerNode extends ChildStmtNode {
+  Handler handler;
+
+  HandlerNode() { handler = stmt }
+
+  override BaseAstNode getChildInternal(int childIndex) {
+    result = super.getChildInternal(childIndex)
+    or
+    childIndex = -1 and
+    result.getAst() = handler.getParameter()
+  }
+}
+
 /**
  * A node representing a `Parameter`.
  */
@@ -552,10 +596,10 @@ class InitializerNode extends AstNode {
 /**
  * A node representing the parameters of a `Function`.
  */
-class ParametersNode extends PrintAstNode, TParametersNode {
+class FunctionParametersNode extends PrintAstNode, TFunctionParametersNode {
   Function func;
 
-  ParametersNode() { this = TParametersNode(func) }
+  FunctionParametersNode() { this = TFunctionParametersNode(func) }
 
   final override string toString() { result = "" }
 
@@ -576,6 +620,33 @@ class ParametersNode extends PrintAstNode, TParametersNode {
   final Function getFunction() { result = func }
 }
 
+/**
+ * A node representing the parameters of a `RequiresExpr`.
+ */
+class RequiresExprParametersNode extends PrintAstNode, TRequiresExprParametersNode {
+  RequiresExpr req;
+
+  RequiresExprParametersNode() { this = TRequiresExprParametersNode(req) }
+
+  final override string toString() { result = "" }
+
+  final override Location getLocation() { result = getRepresentativeLocation(req) }
+
+  override AstNode getChildInternal(int childIndex) {
+    result.getAst() = req.getParameter(childIndex)
+  }
+
+  override string getChildAccessorPredicateInternal(int childIndex) {
+    exists(this.getChildInternal(childIndex)) and
+    result = "getParameter(" + childIndex.toString() + ")"
+  }
+
+  /**
+   * Gets the `RequiresExpr` for which this node represents the parameters.
+   */
+  final RequiresExpr getRequiresExpr() { result = req }
+}
+
 /**
  * A node representing the initializer list of a `Constructor`.
  */
@@ -679,7 +750,7 @@ class FunctionNode extends FunctionOrGlobalOrNamespaceVariableNode {
 
   override PrintAstNode getChildInternal(int childIndex) {
     childIndex = 0 and
-    result.(ParametersNode).getFunction() = func
+    result.(FunctionParametersNode).getFunction() = func
     or
     childIndex = 1 and
     result.(ConstructorInitializersNode).getConstructor() = func
@@ -754,6 +825,8 @@ private predicate namedStmtChildPredicates(Locatable s, Element e, string pred)
     or
     s.(ConstexprIfStmt).getElse() = e and pred = "getElse()"
     or
+    s.(Handler).getParameter() = e and pred = "getParameter()"
+    or
     s.(IfStmt).getInitialization() = e and pred = "getInitialization()"
     or
     s.(IfStmt).getCondition() = e and pred = "getCondition()"
@@ -901,6 +974,11 @@ private predicate namedExprChildPredicates(Expr expr, Element ele, string pred)
     or
     expr.(CommaExpr).getRightOperand() = ele and pred = "getRightOperand()"
     or
+    expr.(CompoundRequirementExpr).getExpr() = ele and pred = "getExpr()"
+    or
+    expr.(CompoundRequirementExpr).getReturnTypeRequirement() = ele and
+    pred = "getReturnTypeRequirement()"
+    or
     expr.(ConditionDeclExpr).getVariableAccess() = ele and pred = "getVariableAccess()"
     or
     expr.(ConstructorFieldInit).getExpr() = ele and pred = "getExpr()"
@@ -921,6 +999,8 @@ private predicate namedExprChildPredicates(Expr expr, Element ele, string pred)
     or
     expr.(LambdaExpression).getInitializer() = ele and pred = "getInitializer()"
     or
+    expr.(NestedRequirementExpr).getConstraint() = ele and pred = "getConstraint()"
+    or
     expr.(NewOrNewArrayExpr).getAllocatorCall() = ele and pred = "getAllocatorCall()"
     or
     expr.(NewOrNewArrayExpr).getAlignmentArgument() = ele and pred = "getAlignmentArgument()"
@@ -960,6 +1040,11 @@ private predicate namedExprChildPredicates(Expr expr, Element ele, string pred)
     or
     expr.(UnaryOperation).getOperand() = ele and pred = "getOperand()"
     or
+    exists(int n |
+      expr.(RequiresExpr).getRequirement(n) = ele and
+      pred = "getRequirement(" + n + ")"
+    )
+    or
     expr.(SizeofExprOperator).getExprOperand() = ele and pred = "getExprOperand()"
     or
     expr.(StmtExpr).getStmt() = ele and pred = "getStmt()"

cpp/ql/lib/semmle/code/cpp/internal/QualifiedName.qll

@@ -181,12 +181,7 @@ class VariableDeclarationEntry extends @var_decl {
   string getName() { var_decls(this, _, _, result, _) and result != "" }
 }
 
-class Parameter extends LocalScopeVariable, @parameter {
+class Parameter extends LocalScopeVariable, @parameter { }
-  @functionorblock function;
-  int index;
-
-  Parameter() { params(this, function, index, _) }
-}
 
 class GlobalOrNamespaceVariable extends Variable, @globalvariable { }
 

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

@@ -6,6 +6,112 @@
  * uses, however, it is better to write a query that imports `PrintIR.qll`, extends
  * `PrintIRConfiguration`, and overrides `shouldPrintDeclaration()` to select a subset of declarations
  * to dump.
+ *
+ * Anatomy of a printed IR instruction
+ *
+ * An instruction:
+ *
+ * ```
+ * # 2281|     v2281_19(void) = Call[~String]     : func:r2281_18, this:r2281_17
+ * ```
+ *
+ * The prefix `# 2281|` specifies that this instruction was generated by the C++ source code on line 2281.
+ * Scrolling up in the printed output, one will eventually find the name of the file to which the line
+ * belongs.
+ *
+ * `v2281_19(void)` is the result of the instruction. Here, `v` means this is a void result or operand (so
+ * there should be no later uses of the result; see below for other possible values). The `2281_19` is a
+ * unique ID for the result. This is usually just the line number plus a small integer suffix to make it
+ * unique within the function. The type of the result is `void`. In this case, it is `void`, because
+ * `~String` returns `void`. The type of the result is usually just the name of the appropriate C++ type,
+ * but it will sometimes be a type like `glval<int>`, which means result holds a glvalue, which at the
+ * IR level works like a pointer. In other words, in the source code the type was `int`, but it is really
+ * more like an `int*`. We see this, for example, in `x = y;`, where `x` is a glvalue.
+ *
+ * `Call` is the opcode of the instruction. Common opcodes include:
+ *
+ * * Arithmetic operations: `Add`, `Sub`, `Mul`, etc.
+ * * Memory access operations: `Load`, `Store`.
+ * * Function calls: `Call`.
+ * * Literals: `Constant`.
+ * * Variable addresses: `VariableAddress`.
+ * * Function entry points: `EnterFunction`.
+ * * Return from a function: `Return`, `ReturnVoid`. Note that the value being returned is set separately by a
+ *   `Store` to a special `#return` variable.
+ * * Stack unwinding for C++ function that throw and where the exception escapes the function: `Unwind`.
+ * * Common exit point for `Unwind` and `Return`: `ExitFunction`.
+ * * SSA-related opcodes: `Phi`, `Chi`.
+ *
+ * `[~String]` denotes additional information. The information might be present earlier in the IR, as is the case
+ * for `Call`, where it is the name of the called function. This is also the case for `Load` and `Store`, where it
+ * is the name of the variable that loaded or stored (if known). In the case of `Constant`, `FieldAddress`, and
+ * `VariableAddress`, the information between brackets does not occur earlier.
+ *
+ * `func:r2281_18` and `this:r28281_17` are the operands of the instruction. The `func:` prefix denotes the operand
+ * that holds the address of the called function. The `this:` prefix denotes the argument to the special `this`
+ * parameter of an instance member function. `r2281_18`, `r2281_17` are the unique IDs of the operands. Each of these
+ * matches the ID of a previously seen result, showing where that value came from. The `r` means that these are
+ * "register" operands (see below).
+ *
+ * Result and operand kinds:
+ *
+ * Every result and operand is one of these three kinds:
+ *
+ * * `r` "register". These operands are not stored in any particular memory location. We can think of them as
+ *   temporary values created during the evaluation of an expression. A register operand almost always has one
+ *   use, often in the same block as its definition.
+ * *  `m` "memory". These operands represents accesses to a specific memory location. The location could be a
+ *    local variable, a global variable, a field of an object, an element of an array, or any memory that we happen
+ *    to have a pointer to. These only occur as the result of a `Store`, the source operand of a `Load` or on the
+ *    SSA instructions (`Phi`, `Chi`).
+ * * `v` "void". Really just a register operand, but we mark register operands of type void with this special prefix
+ *   so we know that there is no actual value there.
+ *
+ * Branches in the IR:
+ *
+ * The IR is divided into basic blocks. At the end of each block, there are one or more edges showing the possible
+ * control flow successors of the block.
+ *
+ * ```
+ * #   44|     v44_3(void)    = ConditionalBranch : r44_2
+ * #-----|   False -> Block 4
+ * #-----|   True -> Block 3
+ * ```
+ * Here we have a block that ends with a conditional branch. The two edges show where the control flows to depending
+ * on whether the condition is true or false.
+ *
+ * SSA instructions:
+ *
+ * We use `Phi` instructions in SSA to create a single definition for a variable that might be assigned on multiple
+ * control flow paths. The `Phi` instruction merges the potential values of that variable from each predecessor edge,
+ * and the resulting definition is then used wherever that variable is accessed later on.
+ *
+ * When dealing with aliased memory, we use the `Chi` instruction to create a single definition for memory that might
+ * or might not have been updated by a store, depending on the actual address that was written to. For example, take:
+ *
+ * ```cpp
+ * int x = 5;
+ * int y = 7;
+ * int* p = condition ? &x : &y;
+ * *p = 6;
+ * return x;
+ * ```
+ *
+ * At the point where we store to `*p`, we do not know whether `p` points to `x` or `y`. Thus, we do not know whether
+ * `return x;` is going to return the value that `x` was originally initialized to (5), or whether it will return 6,
+ *  because it was overwritten by `*p = 6;`. We insert a `Chi` instruction immediately after the store to `*p`:
+ *
+ * ```
+ * r2(int)     = Constant[6]
+ * r3(int*)    = <<value of p>>
+ * m4(int)     = Store           : &r3, r2  // Stores the constant 6 to *p
+ * m5(unknown) = Chi             : total:m1, partial:m4
+ * ```
+ * The `partial:` operand represents the memory that was just stored. The `total:` operand represents the previous
+ * contents of all of the memory that `p` might have pointed to (in this case, both `x` and `y`). The result of the
+ * `Chi` represents the new contents of whatever memory the `total:` operand referred to. We usually do not know exactly
+ * which parts of that memory were overwritten, but it does model that any of that memory could have been modified, so
+ * that later instructions do not assume that the memory was unchanged.
  */
 
 private import internal.IRInternal

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

@@ -6,6 +6,112 @@
  * uses, however, it is better to write a query that imports `PrintIR.qll`, extends
  * `PrintIRConfiguration`, and overrides `shouldPrintDeclaration()` to select a subset of declarations
  * to dump.
+ *
+ * Anatomy of a printed IR instruction
+ *
+ * An instruction:
+ *
+ * ```
+ * # 2281|     v2281_19(void) = Call[~String]     : func:r2281_18, this:r2281_17
+ * ```
+ *
+ * The prefix `# 2281|` specifies that this instruction was generated by the C++ source code on line 2281.
+ * Scrolling up in the printed output, one will eventually find the name of the file to which the line
+ * belongs.
+ *
+ * `v2281_19(void)` is the result of the instruction. Here, `v` means this is a void result or operand (so
+ * there should be no later uses of the result; see below for other possible values). The `2281_19` is a
+ * unique ID for the result. This is usually just the line number plus a small integer suffix to make it
+ * unique within the function. The type of the result is `void`. In this case, it is `void`, because
+ * `~String` returns `void`. The type of the result is usually just the name of the appropriate C++ type,
+ * but it will sometimes be a type like `glval<int>`, which means result holds a glvalue, which at the
+ * IR level works like a pointer. In other words, in the source code the type was `int`, but it is really
+ * more like an `int*`. We see this, for example, in `x = y;`, where `x` is a glvalue.
+ *
+ * `Call` is the opcode of the instruction. Common opcodes include:
+ *
+ * * Arithmetic operations: `Add`, `Sub`, `Mul`, etc.
+ * * Memory access operations: `Load`, `Store`.
+ * * Function calls: `Call`.
+ * * Literals: `Constant`.
+ * * Variable addresses: `VariableAddress`.
+ * * Function entry points: `EnterFunction`.
+ * * Return from a function: `Return`, `ReturnVoid`. Note that the value being returned is set separately by a
+ *   `Store` to a special `#return` variable.
+ * * Stack unwinding for C++ function that throw and where the exception escapes the function: `Unwind`.
+ * * Common exit point for `Unwind` and `Return`: `ExitFunction`.
+ * * SSA-related opcodes: `Phi`, `Chi`.
+ *
+ * `[~String]` denotes additional information. The information might be present earlier in the IR, as is the case
+ * for `Call`, where it is the name of the called function. This is also the case for `Load` and `Store`, where it
+ * is the name of the variable that loaded or stored (if known). In the case of `Constant`, `FieldAddress`, and
+ * `VariableAddress`, the information between brackets does not occur earlier.
+ *
+ * `func:r2281_18` and `this:r28281_17` are the operands of the instruction. The `func:` prefix denotes the operand
+ * that holds the address of the called function. The `this:` prefix denotes the argument to the special `this`
+ * parameter of an instance member function. `r2281_18`, `r2281_17` are the unique IDs of the operands. Each of these
+ * matches the ID of a previously seen result, showing where that value came from. The `r` means that these are
+ * "register" operands (see below).
+ *
+ * Result and operand kinds:
+ *
+ * Every result and operand is one of these three kinds:
+ *
+ * * `r` "register". These operands are not stored in any particular memory location. We can think of them as
+ *   temporary values created during the evaluation of an expression. A register operand almost always has one
+ *   use, often in the same block as its definition.
+ * *  `m` "memory". These operands represents accesses to a specific memory location. The location could be a
+ *    local variable, a global variable, a field of an object, an element of an array, or any memory that we happen
+ *    to have a pointer to. These only occur as the result of a `Store`, the source operand of a `Load` or on the
+ *    SSA instructions (`Phi`, `Chi`).
+ * * `v` "void". Really just a register operand, but we mark register operands of type void with this special prefix
+ *   so we know that there is no actual value there.
+ *
+ * Branches in the IR:
+ *
+ * The IR is divided into basic blocks. At the end of each block, there are one or more edges showing the possible
+ * control flow successors of the block.
+ *
+ * ```
+ * #   44|     v44_3(void)    = ConditionalBranch : r44_2
+ * #-----|   False -> Block 4
+ * #-----|   True -> Block 3
+ * ```
+ * Here we have a block that ends with a conditional branch. The two edges show where the control flows to depending
+ * on whether the condition is true or false.
+ *
+ * SSA instructions:
+ *
+ * We use `Phi` instructions in SSA to create a single definition for a variable that might be assigned on multiple
+ * control flow paths. The `Phi` instruction merges the potential values of that variable from each predecessor edge,
+ * and the resulting definition is then used wherever that variable is accessed later on.
+ *
+ * When dealing with aliased memory, we use the `Chi` instruction to create a single definition for memory that might
+ * or might not have been updated by a store, depending on the actual address that was written to. For example, take:
+ *
+ * ```cpp
+ * int x = 5;
+ * int y = 7;
+ * int* p = condition ? &x : &y;
+ * *p = 6;
+ * return x;
+ * ```
+ *
+ * At the point where we store to `*p`, we do not know whether `p` points to `x` or `y`. Thus, we do not know whether
+ * `return x;` is going to return the value that `x` was originally initialized to (5), or whether it will return 6,
+ *  because it was overwritten by `*p = 6;`. We insert a `Chi` instruction immediately after the store to `*p`:
+ *
+ * ```
+ * r2(int)     = Constant[6]
+ * r3(int*)    = <<value of p>>
+ * m4(int)     = Store           : &r3, r2  // Stores the constant 6 to *p
+ * m5(unknown) = Chi             : total:m1, partial:m4
+ * ```
+ * The `partial:` operand represents the memory that was just stored. The `total:` operand represents the previous
+ * contents of all of the memory that `p` might have pointed to (in this case, both `x` and `y`). The result of the
+ * `Chi` represents the new contents of whatever memory the `total:` operand referred to. We usually do not know exactly
+ * which parts of that memory were overwritten, but it does model that any of that memory could have been modified, so
+ * that later instructions do not assume that the memory was unchanged.
  */
 
 private import internal.IRInternal

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

@@ -6,6 +6,112 @@
  * uses, however, it is better to write a query that imports `PrintIR.qll`, extends
  * `PrintIRConfiguration`, and overrides `shouldPrintDeclaration()` to select a subset of declarations
  * to dump.
+ *
+ * Anatomy of a printed IR instruction
+ *
+ * An instruction:
+ *
+ * ```
+ * # 2281|     v2281_19(void) = Call[~String]     : func:r2281_18, this:r2281_17
+ * ```
+ *
+ * The prefix `# 2281|` specifies that this instruction was generated by the C++ source code on line 2281.
+ * Scrolling up in the printed output, one will eventually find the name of the file to which the line
+ * belongs.
+ *
+ * `v2281_19(void)` is the result of the instruction. Here, `v` means this is a void result or operand (so
+ * there should be no later uses of the result; see below for other possible values). The `2281_19` is a
+ * unique ID for the result. This is usually just the line number plus a small integer suffix to make it
+ * unique within the function. The type of the result is `void`. In this case, it is `void`, because
+ * `~String` returns `void`. The type of the result is usually just the name of the appropriate C++ type,
+ * but it will sometimes be a type like `glval<int>`, which means result holds a glvalue, which at the
+ * IR level works like a pointer. In other words, in the source code the type was `int`, but it is really
+ * more like an `int*`. We see this, for example, in `x = y;`, where `x` is a glvalue.
+ *
+ * `Call` is the opcode of the instruction. Common opcodes include:
+ *
+ * * Arithmetic operations: `Add`, `Sub`, `Mul`, etc.
+ * * Memory access operations: `Load`, `Store`.
+ * * Function calls: `Call`.
+ * * Literals: `Constant`.
+ * * Variable addresses: `VariableAddress`.
+ * * Function entry points: `EnterFunction`.
+ * * Return from a function: `Return`, `ReturnVoid`. Note that the value being returned is set separately by a
+ *   `Store` to a special `#return` variable.
+ * * Stack unwinding for C++ function that throw and where the exception escapes the function: `Unwind`.
+ * * Common exit point for `Unwind` and `Return`: `ExitFunction`.
+ * * SSA-related opcodes: `Phi`, `Chi`.
+ *
+ * `[~String]` denotes additional information. The information might be present earlier in the IR, as is the case
+ * for `Call`, where it is the name of the called function. This is also the case for `Load` and `Store`, where it
+ * is the name of the variable that loaded or stored (if known). In the case of `Constant`, `FieldAddress`, and
+ * `VariableAddress`, the information between brackets does not occur earlier.
+ *
+ * `func:r2281_18` and `this:r28281_17` are the operands of the instruction. The `func:` prefix denotes the operand
+ * that holds the address of the called function. The `this:` prefix denotes the argument to the special `this`
+ * parameter of an instance member function. `r2281_18`, `r2281_17` are the unique IDs of the operands. Each of these
+ * matches the ID of a previously seen result, showing where that value came from. The `r` means that these are
+ * "register" operands (see below).
+ *
+ * Result and operand kinds:
+ *
+ * Every result and operand is one of these three kinds:
+ *
+ * * `r` "register". These operands are not stored in any particular memory location. We can think of them as
+ *   temporary values created during the evaluation of an expression. A register operand almost always has one
+ *   use, often in the same block as its definition.
+ * *  `m` "memory". These operands represents accesses to a specific memory location. The location could be a
+ *    local variable, a global variable, a field of an object, an element of an array, or any memory that we happen
+ *    to have a pointer to. These only occur as the result of a `Store`, the source operand of a `Load` or on the
+ *    SSA instructions (`Phi`, `Chi`).
+ * * `v` "void". Really just a register operand, but we mark register operands of type void with this special prefix
+ *   so we know that there is no actual value there.
+ *
+ * Branches in the IR:
+ *
+ * The IR is divided into basic blocks. At the end of each block, there are one or more edges showing the possible
+ * control flow successors of the block.
+ *
+ * ```
+ * #   44|     v44_3(void)    = ConditionalBranch : r44_2
+ * #-----|   False -> Block 4
+ * #-----|   True -> Block 3
+ * ```
+ * Here we have a block that ends with a conditional branch. The two edges show where the control flows to depending
+ * on whether the condition is true or false.
+ *
+ * SSA instructions:
+ *
+ * We use `Phi` instructions in SSA to create a single definition for a variable that might be assigned on multiple
+ * control flow paths. The `Phi` instruction merges the potential values of that variable from each predecessor edge,
+ * and the resulting definition is then used wherever that variable is accessed later on.
+ *
+ * When dealing with aliased memory, we use the `Chi` instruction to create a single definition for memory that might
+ * or might not have been updated by a store, depending on the actual address that was written to. For example, take:
+ *
+ * ```cpp
+ * int x = 5;
+ * int y = 7;
+ * int* p = condition ? &x : &y;
+ * *p = 6;
+ * return x;
+ * ```
+ *
+ * At the point where we store to `*p`, we do not know whether `p` points to `x` or `y`. Thus, we do not know whether
+ * `return x;` is going to return the value that `x` was originally initialized to (5), or whether it will return 6,
+ *  because it was overwritten by `*p = 6;`. We insert a `Chi` instruction immediately after the store to `*p`:
+ *
+ * ```
+ * r2(int)     = Constant[6]
+ * r3(int*)    = <<value of p>>
+ * m4(int)     = Store           : &r3, r2  // Stores the constant 6 to *p
+ * m5(unknown) = Chi             : total:m1, partial:m4
+ * ```
+ * The `partial:` operand represents the memory that was just stored. The `total:` operand represents the previous
+ * contents of all of the memory that `p` might have pointed to (in this case, both `x` and `y`). The result of the
+ * `Chi` represents the new contents of whatever memory the `total:` operand referred to. We usually do not know exactly
+ * which parts of that memory were overwritten, but it does model that any of that memory could have been modified, so
+ * that later instructions do not assume that the memory was unchanged.
  */
 
 private import internal.IRInternal

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

@@ -7,7 +7,7 @@ import semmle.code.cpp.models.interfaces.Alias
 import semmle.code.cpp.models.interfaces.SideEffect
 
 /** The function `fopen` and friends. */
-private class Fopen extends Function, AliasFunction, SideEffectFunction {
+private class Fopen extends Function, AliasFunction, SideEffectFunction, TaintFunction {
   Fopen() {
     this.hasGlobalOrStdName(["fopen", "fopen_s", "freopen"])
     or
@@ -47,4 +47,22 @@ private class Fopen extends Function, AliasFunction, SideEffectFunction {
     i = 0 and
     buffer = true
   }
+
+  override predicate hasTaintFlow(FunctionInput input, FunctionOutput output) {
+    (
+      this.hasGlobalOrStdName(["fopen", "freopen"]) or
+      this.hasGlobalName(["_wfopen", "_fsopen", "_wfsopen"])
+    ) and
+    input.isParameterDeref(0) and
+    output.isReturnValueDeref()
+    or
+    // The out parameter is a pointer to a `FILE*`.
+    this.hasGlobalOrStdName("fopen_s") and
+    input.isParameterDeref(1) and
+    output.isParameterDeref(0, 2)
+    or
+    this.hasGlobalName(["_open", "_wopen"]) and
+    input.isParameterDeref(0) and
+    output.isReturnValue()
+  }
 }

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

@@ -192,6 +192,37 @@ private class UnsignedMulExpr extends MulExpr {
   }
 }
 
+/**
+ * Gets the value of the `EOF` macro.
+ *
+ * This is typically `"-1"`, but this is not guaranteed to be the case on all
+ * systems.
+ */
+private int getEofValue() {
+  exists(MacroInvocation mi |
+    mi.getMacroName() = "EOF" and
+    result = unique( | | mi.getExpr().getValue().toInt())
+  )
+}
+
+/** Get standard `getc` function or related variants. */
+private class Getc extends Function {
+  Getc() { this.hasGlobalOrStdOrBslName(["fgetc", "getc"]) }
+}
+
+/** A call to `getc` */
+private class CallToGetc extends FunctionCall {
+  CallToGetc() { this.getTarget() instanceof Getc }
+}
+
+/**
+ * A call to `getc` that we can analyze because we know
+ * the value of the `EOF` macro.
+ */
+private class AnalyzableCallToGetc extends CallToGetc {
+  AnalyzableCallToGetc() { exists(getEofValue()) }
+}
+
 /**
  * Holds if `expr` is effectively a multiplication of `operand` with the
  * positive constant `positive`.
@@ -287,6 +318,8 @@ private predicate analyzableExpr(Expr e) {
     or
     e instanceof RemExpr
     or
+    e instanceof AnalyzableCallToGetc
+    or
     // A conversion is analyzable, provided that its child has an arithmetic
     // type. (Sometimes the child is a reference type, and so does not get
     // any bounds.) Rather than checking whether the type of the child is
@@ -861,6 +894,14 @@ private float getLowerBoundsImpl(Expr expr) {
       )
     )
     or
+    exists(AnalyzableCallToGetc getc |
+      expr = getc and
+      // from https://en.cppreference.com/w/c/io/fgetc:
+      // On success, returns the obtained character as an unsigned char
+      // converted to an int. On failure, returns EOF.
+      result = min([typeLowerBound(any(UnsignedCharType pct)), getEofValue()])
+    )
+    or
     // If the conversion is to an arithmetic type then we just return the
     // lower bound of the child. We do not need to handle truncation and
     // overflow here, because that is done in `getTruncatedLowerBounds`.
@@ -1055,6 +1096,14 @@ private float getUpperBoundsImpl(Expr expr) {
       )
     )
     or
+    exists(AnalyzableCallToGetc getc |
+      expr = getc and
+      // from https://en.cppreference.com/w/c/io/fgetc:
+      // On success, returns the obtained character as an unsigned char
+      // converted to an int. On failure, returns EOF.
+      result = max([typeUpperBound(any(UnsignedCharType pct)), getEofValue()])
+    )
+    or
     // If the conversion is to an arithmetic type then we just return the
     // upper bound of the child. We do not need to handle truncation and
     // overflow here, because that is done in `getTruncatedUpperBounds`.

cpp/ql/lib/semmlecode.cpp.dbscheme

@@ -534,7 +534,7 @@ static_asserts(
 #keyset[function, index, type_id]
 params(
     int id: @parameter,
-    int function: @functionorblock ref,
+    int function: @parameterized_element ref,
     int index: int ref,
     int type_id: @type ref
 );
@@ -1790,6 +1790,10 @@ case @expr.kind of
 | 387 = @istriviallyrelocatable
 | 388 = @datasizeof
 | 389 = @c11_generic
+| 390 = @requires_expr
+| 391 = @nested_requirement
+| 392 = @compound_requirement
+| 393 = @concept_id
 ;
 
 @var_args_expr = @vastartexpr
@@ -1908,6 +1912,10 @@ case @expr.kind of
             | @istriviallyrelocatable
             ;
 
+compound_requirement_is_noexcept(
+    int expr: @compound_requirement ref
+);
+
 new_allocated_type(
     unique int expr: @new_expr ref,
     int type_id: @type ref
@@ -2167,11 +2175,11 @@ stmt_decl_entry_bind(
     int decl_entry: @element ref
 );
 
-@functionorblock = @function | @stmt_block;
+@parameterized_element = @function | @stmt_block | @requires_expr;
 
 blockscope(
     unique int block: @stmt_block ref,
-    int enclosing: @functionorblock ref
+    int enclosing: @parameterized_element ref
 );
 
 @jump = @stmt_goto | @stmt_break | @stmt_continue;

cpp/ql/lib/upgrades/6f5d51e89e762fe4609fd4ac8ee3afb04221e873/upgrade.properties

@@ -0,0 +1,2 @@
+description: Support C++20 requires expressions
+compatibility: backwards

cpp/ql/lib/upgrades/7ff6a6e53dbcff09d1b9b758b594bc6d17366863/upgrade.properties

@@ -0,0 +1,2 @@
+description: Add requires expressions
+compatibility: full

cpp/ql/src/Best Practices/Unused Entities/UnusedLocals.ql

@@ -57,5 +57,5 @@ where
   not declarationHasSideEffects(v) and
   not exists(AsmStmt s | f = s.getEnclosingFunction()) and
   not v.getAnAttribute().getName() = "unused" and
-  not any(ErrorExpr e).getEnclosingFunction() = f // unextracted expr may use `v`
+  not f.hasErrors() // Unextracted expressions may use `v`
 select v, "Variable " + v.getName() + " is not used."

cpp/ql/src/CHANGELOG.md

@@ -1,3 +1,12 @@
+## 1.2.5
+
+### Minor Analysis Improvements
+
+* The `cpp/unclear-array-index-validation` ("Unclear validation of array index") query has been improved to reduce false positives and increase true positives.
+* Fixed false positives in the `cpp/uninitialized-local` ("Potentially uninitialized local variable") query if there are extraction errors in the function.
+* The `cpp/incorrect-string-type-conversion` query now produces fewer false positives caused by failure to detect byte arrays.
+* The `cpp/incorrect-string-type-conversion` query now produces fewer false positives caused by failure to recognize dynamic checks prior to possible dangerous widening.
+
 ## 1.2.4
 
 ### Minor Analysis Improvements

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

@@ -29,7 +29,7 @@ class ReturnStackAllocatedMemoryConfig extends MustFlowConfiguration {
   override predicate isSource(Instruction source) {
     exists(Function func |
       // Rule out FPs caused by extraction errors.
-      not any(ErrorExpr e).getEnclosingFunction() = func and
+      not func.hasErrors() and
       not intentionallyReturnsStackPointer(func) and
       func = source.getEnclosingFunction()
     |

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

@@ -65,6 +65,7 @@ predicate isSinkImpl(Instruction sink, VariableAccess va) {
   exists(LoadInstruction load |
     va = load.getUnconvertedResultExpression() and
     not va = commonException() and
+    not va.getTarget().(LocalVariable).getFunction().hasErrors() and
     sink = load.getSourceValue()
   )
 }

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

@@ -24,7 +24,7 @@ predicate instructionHasVariable(VariableAddressInstruction vai, StackVariable v
   // Pointer-to-member types aren't properly handled in the dbscheme.
   not vai.getResultType() instanceof PointerToMemberType and
   // Rule out FPs caused by extraction errors.
-  not any(ErrorExpr e).getEnclosingFunction() = f
+  not f.hasErrors()
 }
 
 /**

cpp/ql/src/Security/CWE/CWE-129/ImproperArrayIndexValidation.ql

@@ -14,102 +14,56 @@
 
 import cpp
 import semmle.code.cpp.controlflow.IRGuards
-import semmle.code.cpp.security.FlowSources
+import semmle.code.cpp.security.FlowSources as FS
-import semmle.code.cpp.ir.dataflow.TaintTracking
+import semmle.code.cpp.dataflow.new.TaintTracking
-import semmle.code.cpp.rangeanalysis.RangeAnalysisUtils
+import semmle.code.cpp.rangeanalysis.SimpleRangeAnalysis
 import ImproperArrayIndexValidation::PathGraph
-import semmle.code.cpp.security.Security
 
-predicate hasUpperBound(VariableAccess offsetExpr) {
+predicate isFlowSource(FS::FlowSource source, string sourceType) {
-  exists(BasicBlock controlled, StackVariable offsetVar, SsaDefinition def |
+  sourceType = source.getSourceType()
-    controlled.contains(offsetExpr) and
+}
-    linearBoundControls(controlled, def, offsetVar) and
+
-    offsetExpr = def.getAUse(offsetVar)
+predicate guardChecks(IRGuardCondition g, Expr e, boolean branch) {
+  exists(Operand op | op.getDef().getConvertedResultExpression() = e |
+    // `op < k` is true and `k > 0`
+    g.comparesLt(op, any(int k | k > 0), true, any(BooleanValue bv | bv.getValue() = branch))
+    or
+    // `op < _ + k` is true and `k > 0`.
+    g.comparesLt(op, _, any(int k | k > 0), true, branch)
+    or
+    // op == k
+    g.comparesEq(op, _, true, any(BooleanValue bv | bv.getValue() = branch))
+    or
+    // op == _ + k
+    g.comparesEq(op, _, _, true, branch)
   )
 }
 
-pragma[noinline]
+/**
-predicate linearBoundControls(BasicBlock controlled, SsaDefinition def, StackVariable offsetVar) {
+ * Holds if `arrayExpr` accesses an `ArrayType` with a constant size `N`, and
-  exists(GuardCondition guard, boolean branch |
+ * the value of `offsetExpr` is known to be smaller than `N`.
-    guard.controls(controlled, branch) and
+ */
-    cmpWithLinearBound(guard, def.getAUse(offsetVar), Lesser(), branch)
+predicate offsetIsAlwaysInBounds(ArrayExpr arrayExpr, VariableAccess offsetExpr) {
+  exists(ArrayType arrayType |
+    arrayType = arrayExpr.getArrayBase().getUnspecifiedType() and
+    arrayType.getArraySize() > upperBound(offsetExpr.getFullyConverted())
   )
 }
 
-predicate readsVariable(LoadInstruction load, Variable var) {
-  load.getSourceAddress().(VariableAddressInstruction).getAstVariable() = var
-}
-
-predicate hasUpperBoundsCheck(Variable var) {
-  exists(RelationalOperation oper, VariableAccess access |
-    oper.getAnOperand() = access and
-    access.getTarget() = var and
-    // Comparing to 0 is not an upper bound check
-    not oper.getAnOperand().getValue() = "0"
-  )
-}
-
-predicate nodeIsBarrierEqualityCandidate(DataFlow::Node node, Operand access, Variable checkedVar) {
-  readsVariable(node.asInstruction(), checkedVar) and
-  any(IRGuardCondition guard).ensuresEq(access, _, _, node.asInstruction().getBlock(), true)
-}
-
-predicate isFlowSource(FlowSource source, string sourceType) { sourceType = source.getSourceType() }
-
-predicate predictableInstruction(Instruction instr) {
-  instr instanceof ConstantInstruction
-  or
-  instr instanceof StringConstantInstruction
-  or
-  // This could be a conversion on a string literal
-  predictableInstruction(instr.(UnaryInstruction).getUnary())
-}
-
 module ImproperArrayIndexValidationConfig implements DataFlow::ConfigSig {
   predicate isSource(DataFlow::Node source) { isFlowSource(source, _) }
 
   predicate isBarrier(DataFlow::Node node) {
-    hasUpperBound(node.asExpr())
+    node = DataFlow::BarrierGuard<guardChecks/3>::getABarrierNode()
-    or
-    // These barriers are ported from `DefaultTaintTracking` because this query is quite noisy
-    // otherwise.
-    exists(Variable checkedVar |
-      readsVariable(node.asInstruction(), checkedVar) and
-      hasUpperBoundsCheck(checkedVar)
-    )
-    or
-    exists(Variable checkedVar, Operand access |
-      readsVariable(access.getDef(), checkedVar) and
-      nodeIsBarrierEqualityCandidate(node, access, checkedVar)
-    )
-    or
-    // Don't use dataflow into binary instructions if both operands are unpredictable
-    exists(BinaryInstruction iTo |
-      iTo = node.asInstruction() and
-      not predictableInstruction(iTo.getLeft()) and
-      not predictableInstruction(iTo.getRight()) and
-      // propagate taint from either the pointer or the offset, regardless of predictability
-      not iTo instanceof PointerArithmeticInstruction
-    )
-    or
-    // don't use dataflow through calls to pure functions if two or more operands
-    // are unpredictable
-    exists(Instruction iFrom1, Instruction iFrom2, CallInstruction iTo |
-      iTo = node.asInstruction() and
-      isPureFunction(iTo.getStaticCallTarget().getName()) and
-      iFrom1 = iTo.getAnArgument() and
-      iFrom2 = iTo.getAnArgument() and
-      not predictableInstruction(iFrom1) and
-      not predictableInstruction(iFrom2) and
-      iFrom1 != iFrom2
-    )
   }
 
+  predicate isBarrierOut(DataFlow::Node node) { isSink(node) }
+
   predicate isSink(DataFlow::Node sink) {
     exists(ArrayExpr arrayExpr, VariableAccess offsetExpr |
       offsetExpr = arrayExpr.getArrayOffset() and
       sink.asExpr() = offsetExpr and
-      not hasUpperBound(offsetExpr)
+      not offsetIsAlwaysInBounds(arrayExpr, offsetExpr)
     )
   }
 }

cpp/ql/src/Security/CWE/CWE-704/WcharCharConversion.ql

@@ -13,23 +13,85 @@
  */
 
 import cpp
+import semmle.code.cpp.controlflow.Guards
 
 class WideCharPointerType extends PointerType {
   WideCharPointerType() { this.getBaseType() instanceof WideCharType }
 }
 
+/**
+ * Given type `t`, recurses through and returns all
+ * intermediate base types, including `t`.
+ */
+Type getABaseType(Type t) {
+  result = t
+  or
+  result = getABaseType(t.(DerivedType).getBaseType())
+  or
+  result = getABaseType(t.(TypedefType).getBaseType())
+}
+
 /**
  * A type that may also be `CharPointerType`, but that are likely used as arbitrary buffers.
  */
 class UnlikelyToBeAStringType extends Type {
   UnlikelyToBeAStringType() {
-    this.(PointerType).getBaseType().(CharType).isUnsigned() or
+    exists(Type targ | getABaseType(this) = targ |
-    this.(PointerType).getBaseType().getName().toLowerCase().matches("%byte") or
+      // NOTE: not using CharType isUnsigned, but rather look for any explicitly declared unsigned
-    this.getName().toLowerCase().matches("%byte") or
+      // char types. Assuming these are used for buffers, not strings.
-    this.(PointerType).getBaseType().hasName("uint8_t")
+      targ.(CharType).getName().toLowerCase().matches("unsigned%") or
+      targ.getName().toLowerCase().matches(["uint8_t", "%byte%"])
+    )
   }
 }
 
+// Types that can be wide depending on the UNICODE macro
+// see https://learn.microsoft.com/en-us/windows/win32/winprog/windows-data-types
+class UnicodeMacroDependentWidthType extends Type {
+  UnicodeMacroDependentWidthType() {
+    exists(Type targ | getABaseType(this) = targ |
+      targ.getName() in [
+          "LPCTSTR",
+          "LPTSTR",
+          "PCTSTR",
+          "PTSTR",
+          "TBYTE",
+          "TCHAR"
+        ]
+    )
+  }
+}
+
+class UnicodeMacro extends Macro {
+  UnicodeMacro() { this.getName().toLowerCase().matches("%unicode%") }
+}
+
+class UnicodeMacroInvocation extends MacroInvocation {
+  UnicodeMacroInvocation() { this.getMacro() instanceof UnicodeMacro }
+}
+
+/**
+ * Holds when a expression whose type is UnicodeMacroDependentWidthType and
+ * is observed to be guarded by a check involving a bitwise-and operation
+ * with a UnicodeMacroInvocation.
+ * Such expressions are assumed to be checked dynamically, i.e.,
+ * the flag would indicate if UNICODE typing is set correctly to allow
+ * or disallow a widening cast.
+ */
+predicate isLikelyDynamicallyChecked(Expr e) {
+  e.getType() instanceof UnicodeMacroDependentWidthType and
+  exists(GuardCondition gc, BitwiseAndExpr bai, UnicodeMacroInvocation umi |
+    bai.getAnOperand() = umi.getExpr()
+  |
+    // bai == 0 is false when reaching `e.getBasicBlock()`.
+    // That is, bai != 0 when reaching `e.getBasicBlock()`.
+    gc.ensuresEq(bai, 0, e.getBasicBlock(), false)
+    or
+    // bai == k and k != 0 is true when reaching `e.getBasicBlock()`.
+    gc.ensuresEq(bai, any(int k | k != 0), e.getBasicBlock(), true)
+  )
+}
+
 from Expr e1, Cast e2
 where
   e2 = e1.getConversion() and
@@ -42,7 +104,11 @@ where
   not e1.getType() instanceof UnlikelyToBeAStringType and
   // Avoid castings from 'new' expressions as typically these will be safe
   // Example: `__Type* ret = reinterpret_cast<__Type*>(New(m_pmo) char[num * sizeof(__Type)]);`
-  not exists(NewOrNewArrayExpr newExpr | newExpr.getAChild*() = e1)
+  not exists(NewOrNewArrayExpr newExpr | newExpr.getAChild*() = e1) and
+  // Avoid cases where the cast is guarded by a check to determine if
+  // unicode encoding is enabled in such a way to disallow the dangerous cast
+  // at runtime.
+  not isLikelyDynamicallyChecked(e1)
 select e1,
   "Conversion from " + e1.getType().toString() + " to " + e2.getType().toString() +
     ". Use of invalid string can lead to undefined behavior."

cpp/ql/src/change-notes/released/1.2.5.md

@@ -0,0 +1,8 @@
+## 1.2.5
+
+### Minor Analysis Improvements
+
+* The `cpp/unclear-array-index-validation` ("Unclear validation of array index") query has been improved to reduce false positives and increase true positives.
+* Fixed false positives in the `cpp/uninitialized-local` ("Potentially uninitialized local variable") query if there are extraction errors in the function.
+* The `cpp/incorrect-string-type-conversion` query now produces fewer false positives caused by failure to detect byte arrays.
+* The `cpp/incorrect-string-type-conversion` query now produces fewer false positives caused by failure to recognize dynamic checks prior to possible dangerous widening.

cpp/ql/src/codeql-pack.release.yml

@@ -1,2 +1,2 @@
 ---
-lastReleaseVersion: 1.2.4
+lastReleaseVersion: 1.2.5

cpp/ql/src/jsf/4.13 Functions/AV Rule 114.ql

@@ -49,7 +49,7 @@ predicate functionsMissingReturnStmt(Function f, ControlFlowNode blame) {
 predicate functionImperfectlyExtracted(Function f) {
   exists(CompilerError e | f.getBlock().getLocation().subsumes(e.getLocation()))
   or
-  exists(ErrorExpr ee | ee.getEnclosingFunction() = f)
+  f.hasErrors()
   or
   count(f.getType()) > 1
   or

cpp/ql/src/qlpack.yml

@@ -1,5 +1,5 @@
 name: codeql/cpp-queries
-version: 1.2.4
+version: 1.2.6-dev
 groups:
   - cpp
   - queries

cpp/ql/test/examples/expressions/PrintAST.expected

@@ -870,6 +870,8 @@ Throw.cpp:
 #    8|               Type = [BoolType] bool
 #    8|               ValueCategory = prvalue
 #   12|       getChild(1): [Handler] <handler>
+#   12|         getParameter(): [Parameter] e
+#   12|             Type = [PointerType] E *
 #   12|         getBlock(): [CatchBlock] { ... }
 #   13|           getStmt(0): [ExprStmt] ExprStmt
 #   13|             getExpr(): [ReThrowExpr] re-throw exception 

cpp/ql/test/experimental/query-tests/Security/CWE/CWE-125/semmle/tests/test.cpp

@@ -6,7 +6,7 @@ int wprintf (const wchar_t* format, ...);
 int strlen( const char * string );
 int checkErrors();
 
-void goodTest0()
+static void goodTest0()
 {
   char * ptr = "123456789";
   int ret;
@@ -17,7 +17,7 @@ void goodTest0()
     ptr += ret;
   }
 }
-void goodTest1(const char* ptr)
+static void goodTest1(const char* ptr)
 {
   int ret;
   int len;
@@ -27,7 +27,7 @@ void goodTest1(const char* ptr)
     ptr += ret;
   }
 }
-void goodTest2(char* ptr)
+static void goodTest2(char* ptr)
 {
   int ret;
   ptr[10]=0;
@@ -38,7 +38,7 @@ void goodTest2(char* ptr)
   }
 }
 
-void goodTest3(const char* ptr)
+static void goodTest3(const char* ptr)
 {
   int ret;
   int len;
@@ -48,7 +48,7 @@ void goodTest3(const char* ptr)
     ptr += ret;
   }
 }
-void goodTest4(const char* ptr)
+static void goodTest4(const char* ptr)
 {
   int ret;
   int len;
@@ -58,7 +58,7 @@ void goodTest4(const char* ptr)
     ptr += ret;
   }
 }
-void badTest1(const char* ptr)
+static void badTest1(const char* ptr)
 {
   int ret;
   int len;
@@ -68,7 +68,7 @@ void badTest1(const char* ptr)
     ptr += ret;
   }
 }
-void badTest2(const char* ptr)
+static void badTest2(const char* ptr)
 {
   int ret;
   int len;
@@ -79,7 +79,7 @@ void badTest2(const char* ptr)
   }
 }
 
-void goodTest5(const char* ptr,wchar_t *wc,int wc_len)
+static void goodTest5(const char* ptr,wchar_t *wc,int wc_len)
 {
   int ret;
   int len;
@@ -96,7 +96,7 @@ void goodTest5(const char* ptr,wchar_t *wc,int wc_len)
   }
 }
 
-void badTest3(const char* ptr,int wc_len)
+static void badTest3(const char* ptr,int wc_len)
 {
   int ret;
   int len;
@@ -113,7 +113,7 @@ void badTest3(const char* ptr,int wc_len)
     wc++;
   }
 }
-void badTest4(const char* ptr,int wc_len)
+static void badTest4(const char* ptr,int wc_len)
 {
   int ret;
   int len;
@@ -130,7 +130,7 @@ void badTest4(const char* ptr,int wc_len)
     wc++;
   }
 }
-void badTest5(const char* ptr,int wc_len)
+static void badTest5(const char* ptr,int wc_len)
 {
   int ret;
   int len;
@@ -148,7 +148,7 @@ void badTest5(const char* ptr,int wc_len)
   }
 }
 
-void badTest6(const char* ptr,int wc_len)
+static void badTest6(const char* ptr,int wc_len)
 {
   int ret;
   int len;
@@ -171,7 +171,7 @@ void badTest6(const char* ptr,int wc_len)
     ptr+=ret;
   }
 }
-void badTest7(const char* ptr,int wc_len)
+static void badTest7(const char* ptr,int wc_len)
 {
   int ret;
   int len;
@@ -188,7 +188,7 @@ void badTest7(const char* ptr,int wc_len)
     ptr+=ret;
   }
 }
-void badTest8(const char* ptr,wchar_t *wc)
+static void badTest8(const char* ptr,wchar_t *wc)
 {
   int ret;
   int len;

cpp/ql/test/experimental/query-tests/Security/CWE/CWE-125/semmle/tests/test1.cpp

@@ -24,7 +24,7 @@ typedef unsigned int size_t;
 void* calloc (size_t num, size_t size);
 void* malloc (size_t size);
 
-void badTest1(void *src,  int size) {
+static void badTest1(void *src,  int size) {
     WideCharToMultiByte(CP_ACP, 0, (LPCWSTR)src, -1, (LPSTR)src, size, 0, 0); // BAD
     MultiByteToWideChar(CP_ACP, 0, (LPCSTR)src, -1, (LPCWSTR)src, 30); // BAD
 }
@@ -39,43 +39,43 @@ void goodTest2(){
   }
   printf("%s\n", dst);
 }
-void badTest2(){
+static void badTest2(){
   wchar_t src[] = L"0123456789ABCDEF";
   char dst[16];
   WideCharToMultiByte(CP_UTF8, 0, src, -1, dst, 16, NULL, NULL); // BAD
   printf("%s\n", dst);
 }
-void goodTest3(){
+static void goodTest3(){
   char src[] = "0123456789ABCDEF";
   int size = MultiByteToWideChar(CP_UTF8, 0, src,sizeof(src),NULL,0);
   wchar_t * dst = (wchar_t*)calloc(size + 1, sizeof(wchar_t));
   MultiByteToWideChar(CP_UTF8, 0, src, -1, dst, size+1); // GOOD
 }
-void badTest3(){
+static void badTest3(){
   char src[] = "0123456789ABCDEF";
   int size = MultiByteToWideChar(CP_UTF8, 0, src,sizeof(src),NULL,0);
   wchar_t * dst = (wchar_t*)calloc(size + 1, 1);
   MultiByteToWideChar(CP_UTF8, 0, src, -1, dst, size+1); // BAD
 }
-void goodTest4(){
+static void goodTest4(){
   char src[] = "0123456789ABCDEF";
   int size = MultiByteToWideChar(CP_UTF8, 0, src,sizeof(src),NULL,0);
   wchar_t * dst = (wchar_t*)malloc((size + 1)*sizeof(wchar_t));
   MultiByteToWideChar(CP_UTF8, 0, src, -1, dst, size+1); // GOOD
 }
-void badTest4(){
+static void badTest4(){
   char src[] = "0123456789ABCDEF";
   int size = MultiByteToWideChar(CP_UTF8, 0, src,sizeof(src),NULL,0);
   wchar_t * dst = (wchar_t*)malloc(size + 1);
   MultiByteToWideChar(CP_UTF8, 0, src, -1, dst, size+1); // BAD
 }
-int goodTest5(void *src){
+static int goodTest5(void *src){
   return WideCharToMultiByte(CP_ACP, 0, (LPCWSTR)src, -1, 0, 0, 0, 0); // GOOD
 }
-int badTest5 (void *src) {
+static int badTest5 (void *src) {
   return WideCharToMultiByte(CP_ACP, 0, (LPCWSTR)src, -1, 0, 3, 0, 0);  // BAD
 }
-void goodTest6(WCHAR *src)
+static void goodTest6(WCHAR *src)
 {
   int size;
   char dst[5] ="";
@@ -87,7 +87,7 @@ void goodTest6(WCHAR *src)
   WideCharToMultiByte(CP_ACP, 0, src, -1, dst, sizeof(dst), 0, 0); // GOOD
   printf("%s\n", dst);
 }
-void badTest6(WCHAR *src)
+static void badTest6(WCHAR *src)
 {
     char dst[5] ="";
     WideCharToMultiByte(CP_ACP, 0, src, -1, dst, 260, 0, 0); // BAD

cpp/ql/test/experimental/query-tests/Security/CWE/CWE-125/semmle/tests/test2.cpp

@@ -11,14 +11,14 @@ size_t _mbstowcs_l(wchar_t *wcstr,const char *mbstr,size_t count, _locale_t loca
 size_t mbsrtowcs(wchar_t *wcstr,const char *mbstr,size_t count, mbstate_t *mbstate);
 
 
-void badTest1(void *src,  int size) {
+static void badTest1(void *src,  int size) {
     mbstowcs((wchar_t*)src,(char*)src,size); // BAD
     _locale_t locale;
     _mbstowcs_l((wchar_t*)src,(char*)src,size,locale); // BAD
     mbstate_t *mbstate;
     mbsrtowcs((wchar_t*)src,(char*)src,size,mbstate); // BAD
 }
-void goodTest2(){
+static void goodTest2(){
   char src[] = "0123456789ABCDEF";
   wchar_t dst[16];
   int res = mbstowcs(dst, src,16); // GOOD
@@ -29,43 +29,43 @@ void goodTest2(){
   }
   printf("%s\n", dst);
 }
-void badTest2(){
+static void badTest2(){
   char src[] = "0123456789ABCDEF";
   wchar_t dst[16];
   mbstowcs(dst, src,16); // BAD
   printf("%s\n", dst);
 }
-void goodTest3(){
+static void goodTest3(){
   char src[] = "0123456789ABCDEF";
   int size = mbstowcs(NULL, src,NULL);
   wchar_t * dst = (wchar_t*)calloc(size + 1, sizeof(wchar_t));
   mbstowcs(dst, src,size+1); // GOOD
 }
-void badTest3(){
+static void badTest3(){
   char src[] = "0123456789ABCDEF";
   int size = mbstowcs(NULL, src,NULL);
   wchar_t * dst = (wchar_t*)calloc(size + 1, 1);
   mbstowcs(dst, src,size+1); // BAD
 }
-void goodTest4(){
+static void goodTest4(){
   char src[] = "0123456789ABCDEF";
   int size = mbstowcs(NULL, src,NULL);
   wchar_t * dst = (wchar_t*)malloc((size + 1)*sizeof(wchar_t));
   mbstowcs(dst, src,size+1); // GOOD
 }
-void badTest4(){
+static void badTest4(){
   char src[] = "0123456789ABCDEF";
   int size = mbstowcs(NULL, src,NULL);
   wchar_t * dst = (wchar_t*)malloc(size + 1);
   mbstowcs(dst, src,size+1); // BAD
 }
-int goodTest5(void *src){
+static int goodTest5(void *src){
   return mbstowcs(NULL, (char*)src,NULL); // GOOD
 }
-int badTest5 (void *src) {
+static int badTest5 (void *src) {
   return mbstowcs(NULL, (char*)src,3); // BAD
 }
-void goodTest6(void *src){
+static void goodTest6(void *src){
   wchar_t dst[5];
   int size = mbstowcs(NULL, (char*)src,NULL);
   if(size>=sizeof(dst)){
@@ -75,7 +75,7 @@ void goodTest6(void *src){
   mbstowcs(dst, (char*)src,sizeof(dst)); // GOOD
   printf("%s\n", dst);
 }
-void badTest6(void *src){
+static void badTest6(void *src){
   wchar_t dst[5];
   mbstowcs(dst, (char*)src,260); // BAD
   printf("%s\n", dst);

cpp/ql/test/experimental/query-tests/Security/CWE/CWE-125/semmle/tests/test3.cpp

@@ -9,14 +9,14 @@ void goodTest1(unsigned char *src){
   unsigned char dst[50];
   _mbsnbcpy(dst,src,sizeof(dst)); // GOOD
 }
-size_t badTest1(unsigned char *src){
+static size_t badTest1(unsigned char *src){
   int cb = 0;
   unsigned char dst[50];
   while( cb < sizeof(dst) )
     dst[cb++]=*src++; // BAD
   return _mbclen(dst);
 }
-void goodTest2(unsigned char *src){
+static void goodTest2(unsigned char *src){
 
   int cb = 0;
   unsigned char dst[50];
@@ -27,7 +27,7 @@ void goodTest2(unsigned char *src){
     src=_mbsinc(src);
   }
 }
-void badTest2(unsigned char *src){
+static void badTest2(unsigned char *src){
 
   int cb = 0;
   unsigned char dst[50];
@@ -38,11 +38,11 @@ void badTest2(unsigned char *src){
     src=_mbsinc(src);
   }
 }
-void goodTest3(){
+static void goodTest3(){
   wchar_t name[50];
   name[sizeof(name) / sizeof(*name) - 1] = L'\0'; // GOOD
 }
-void badTest3(){
+static void badTest3(){
   wchar_t name[50];
   name[sizeof(name) - 1] = L'\0'; // BAD
 }

cpp/ql/test/library-tests/dataflow/dataflow-tests/flowOut.cpp

@@ -78,7 +78,7 @@ int* deref(int** p) { // $ ast-def=p ir-def=*p ir-def=**p
   return q;
 }
 
-void test1() {
+void flowout_test1() {
   int x = 0;
   int* p = &x;
   deref(&p)[0] = source();
@@ -95,7 +95,7 @@ void addtaint2(int** p) { // $ ast-def=p ir-def=*p ir-def=**p
   addtaint1(q);
 }
 
-void test2() {
+void flowout_test2() {
   int x = 0;
   int* p = &x;
   addtaint2(&p);

cpp/ql/test/library-tests/dataflow/dataflow-tests/self-Iterator.cpp

@@ -15,7 +15,7 @@ template<> struct std::iterator_traits<unsigned long>
 };
 
 
-int test() {
+int iterator_test() {
     unsigned long x = source();
     sink(x); // $ ast ir
 }

cpp/ql/test/library-tests/dataflow/dataflow-tests/true_upon_entry.cpp

@@ -5,7 +5,7 @@ int source();
 void sink(...);
 bool random();
 
-void test1() {
+void on_entry_test1() {
   int x = source();
   for (int i = 0; i < 10; i++) {
     x = 0;
@@ -13,7 +13,7 @@ void test1() {
   sink(x); // $ SPURIOUS: ir
 }
 
-void test2(int iterations) {
+void on_entry_test2(int iterations) {
   int x = source();
   for (int i = 0; i < iterations; i++) {
     x = 0;
@@ -21,7 +21,7 @@ void test2(int iterations) {
   sink(x); // $ ast,ir
 }
 
-void test3() {
+void on_entry_test3() {
   int x = 0;
   for (int i = 0; i < 10; i++) {
     x = source();
@@ -29,7 +29,7 @@ void test3() {
   sink(x); // $ ast,ir
 }
 
-void test4() {
+void on_entry_test4() {
   int x = source();
   for (int i = 0; i < 10; i++) {
     if (random())
@@ -39,7 +39,7 @@ void test4() {
   sink(x); // $ ast,ir
 }
 
-void test5() {
+void on_entry_test5() {
   int x = source();
   for (int i = 0; i < 10; i++) {
     if (random())
@@ -49,7 +49,7 @@ void test5() {
   sink(x); // $ ast,ir
 }
 
-void test6() {
+void on_entry_test6() {
   int y;
   int x = source();
   for (int i = 0; i < 10 && (y = 1); i++) {
@@ -57,7 +57,7 @@ void test6() {
   sink(x); // $ ast,ir
 }
 
-void test7() {
+void on_entry_test7() {
   int y;
   int x = source();
   for (int i = 0; i < 10 && (y = 1); i++) {
@@ -66,7 +66,7 @@ void test7() {
   sink(x); // $ SPURIOUS: ir
 }
 
-void test8() {
+void on_entry_test8() {
   int x = source();
   // It appears to the analysis that the condition can exit after `i < 10`
   // without having assigned to `x`. That is an effect of how the
@@ -78,7 +78,7 @@ void test8() {
   sink(x); // $ SPURIOUS: ast,ir
 }
 
-void test9() {
+void on_entry_test9() {
   int y;
   int x = source();
   for (int i = 0; (y = 1) && i < 10; i++) {
@@ -86,14 +86,14 @@ void test9() {
   sink(x); // $ ast,ir
 }
 
-void test10() {
+void on_entry_test10() {
   int x = source();
   for (int i = 0; (x = 1) && i < 10; i++) {
   }
   sink(x); // no flow
 }
 
-void test10(int b, int d) {
+void on_entry_test10(int b, int d) {
   int i = 0;
   int x = source();
   if (b)

cpp/ql/test/library-tests/dataflow/taint-tests/localTaint.expected

@@ -6584,6 +6584,16 @@ WARNING: module 'TaintTracking' has been deprecated and may be removed in future
 | taint.cpp:767:21:767:24 | ref arg path | taint.cpp:768:8:768:11 | path |  |
 | taint.cpp:768:8:768:11 | path | taint.cpp:768:7:768:11 | * ... |  |
 | taint.cpp:778:37:778:42 | call to source | taint.cpp:779:7:779:9 | obj |  |
+| taint.cpp:785:23:785:28 | source | taint.cpp:785:23:785:28 | source |  |
+| taint.cpp:785:23:785:28 | source | taint.cpp:786:18:786:23 | source |  |
+| taint.cpp:785:23:785:28 | source | taint.cpp:790:15:790:20 | source |  |
+| taint.cpp:786:12:786:16 | call to fopen | taint.cpp:787:7:787:7 | f |  |
+| taint.cpp:786:18:786:23 | source | taint.cpp:786:12:786:16 | call to fopen | TAINT |
+| taint.cpp:789:8:789:9 | f2 | taint.cpp:790:11:790:12 | f2 |  |
+| taint.cpp:789:8:789:9 | f2 | taint.cpp:791:7:791:8 | f2 |  |
+| taint.cpp:790:10:790:12 | ref arg & ... | taint.cpp:790:11:790:12 | f2 [inner post update] |  |
+| taint.cpp:790:10:790:12 | ref arg & ... | taint.cpp:791:7:791:8 | f2 |  |
+| taint.cpp:790:11:790:12 | f2 | taint.cpp:790:10:790:12 | & ... |  |
 | vector.cpp:16:43:16:49 | source1 | vector.cpp:17:26:17:32 | source1 |  |
 | vector.cpp:16:43:16:49 | source1 | vector.cpp:31:38:31:44 | source1 |  |
 | vector.cpp:17:21:17:33 | call to vector | vector.cpp:19:14:19:14 | v |  |

cpp/ql/test/library-tests/dataflow/taint-tests/taint.cpp

@@ -777,4 +777,16 @@ TaintInheritingContentObject source(bool);
 void test_TaintInheritingContent() {
 	TaintInheritingContentObject obj = source(true);
 	sink(obj.flowFromObject); // $ ir MISSING: ast
+}
+
+FILE* fopen(const char*, const char*);
+int fopen_s(FILE** pFile, const char *filename, const char *mode);
+
+void fopen_test(char* source) {
+	FILE* f = fopen(source, "r");
+	sink(f); // $ ast,ir
+
+	FILE* f2;
+	fopen_s(&f2, source, "r");
+	sink(f2); // $ ast,ir
 }

cpp/ql/test/library-tests/files/Files.expected

@@ -1,4 +1,4 @@
-| c.c | library-tests/files/c.c | CFile, MetricFile | C |  |  |
+| c.c | c.c | CFile, MetricFile | C |  |  |
-| files1.cpp | library-tests/files/files1.cpp | CppFile, MetricFile | C++ | swap | t |
+| files1.cpp | files1.cpp | CppFile, MetricFile | C++ | swap | t |
-| files1.h | library-tests/files/files1.h | HeaderFile, MetricFile |  | swap |  |
+| files1.h | files1.h | HeaderFile, MetricFile |  | swap |  |
-| files2.cpp | library-tests/files/files2.cpp | CppFile, MetricFile | C++ | g | x, y |
+| files2.cpp | files2.cpp | CppFile, MetricFile | C++ | g | x, y |

cpp/ql/test/library-tests/ir/ir/PrintAST.expected

@@ -9055,6 +9055,8 @@ ir.cpp:
 #  733|                 Value = [Literal] 7
 #  733|                 ValueCategory = prvalue
 #  735|       getChild(1): [Handler] <handler>
+#  735|         getParameter(): [Parameter] s
+#  735|             Type = [PointerType] const char *
 #  735|         getBlock(): [CatchBlock] { ... }
 #  736|           getStmt(0): [ExprStmt] ExprStmt
 #  736|             getExpr(): [ThrowExpr] throw ...
@@ -9067,6 +9069,8 @@ ir.cpp:
 #  736|                     Type = [PointerType] const char *
 #  736|                     ValueCategory = prvalue(load)
 #  738|       getChild(2): [Handler] <handler>
+#  738|         getParameter(): [Parameter] e
+#  738|             Type = [LValueReferenceType] const String &
 #  738|         getBlock(): [CatchBlock] { ... }
 #  740|       getChild(3): [Handler] <handler>
 #  740|         getBlock(): [CatchAnyBlock] { ... }
@@ -12852,6 +12856,8 @@ ir.cpp:
 # 1200|                 Value = [Literal] 7
 # 1200|                 ValueCategory = prvalue
 # 1202|       getChild(1): [Handler] <handler>
+# 1202|         getParameter(): [Parameter] s
+# 1202|             Type = [PointerType] const char *
 # 1202|         getBlock(): [CatchBlock] { ... }
 # 1203|           getStmt(0): [ExprStmt] ExprStmt
 # 1203|             getExpr(): [ThrowExpr] throw ...
@@ -12864,6 +12870,8 @@ ir.cpp:
 # 1203|                     Type = [PointerType] const char *
 # 1203|                     ValueCategory = prvalue(load)
 # 1205|       getChild(2): [Handler] <handler>
+# 1205|         getParameter(): [Parameter] e
+# 1205|             Type = [LValueReferenceType] const String &
 # 1205|         getBlock(): [CatchBlock] { ... }
 # 1207|     getStmt(1): [ReturnStmt] return ...
 # 1211| [TopLevelFunction] void VectorTypes(int)
@@ -20586,6 +20594,8 @@ ir.cpp:
 # 2281|               Type = [Struct] String
 # 2281|               ValueCategory = lvalue
 # 2282|       getChild(1): [Handler] <handler>
+# 2282|         getParameter(): [Parameter] s
+# 2282|             Type = [PointerType] const char *
 # 2282|         getBlock(): [CatchBlock] { ... }
 # 2283|           getStmt(0): [ExprStmt] ExprStmt
 # 2283|             getExpr(): [ThrowExpr] throw ...
@@ -20598,6 +20608,8 @@ ir.cpp:
 # 2283|                     Type = [PointerType] const char *
 # 2283|                     ValueCategory = prvalue(load)
 # 2285|       getChild(2): [Handler] <handler>
+# 2285|         getParameter(): [Parameter] e
+# 2285|             Type = [LValueReferenceType] const String &
 # 2285|         getBlock(): [CatchBlock] { ... }
 # 2287|       getChild(3): [Handler] <handler>
 # 2287|         getBlock(): [CatchAnyBlock] { ... }
@@ -22845,6 +22857,8 @@ ir.cpp:
 # 2537|                 Value = [Literal] 42
 # 2537|                 ValueCategory = prvalue
 # 2539|       getChild(1): [Handler] <handler>
+# 2539|         getParameter(): [Parameter] (unnamed parameter 0)
+# 2539|             Type = [PlainCharType] char
 # 2539|         getBlock(): [CatchBlock] { ... }
 # 2541|         getImplicitDestructorCall(0): [DestructorCall] call to ~ClassWithDestructor
 # 2541|             Type = [VoidType] void
@@ -23564,6 +23578,42 @@ ir.cpp:
 # 2686|           Value = [CStyleCast] 0
 # 2686|           ValueCategory = prvalue
 # 2687|     getStmt(1): [ReturnStmt] return ...
+# 2691| [TopLevelFunction] int concepts::requires_use()
+# 2691|   <params>: 
+# 2691|   getEntryPoint(): [BlockStmt] { ... }
+# 2692|     getStmt(0): [DeclStmt] declaration
+# 2692|       getDeclarationEntry(0): [VariableDeclarationEntry] definition of y
+# 2692|           Type = [IntType] int
+# 2692|         getVariable().getInitializer(): [Initializer] initializer for y
+#-----|           getExpr(): [RequiresExpr] requires { ... }
+#-----|               Type = [BoolType] bool
+#-----|               Value = [RequiresExpr] 1
+#-----|               ValueCategory = prvalue
+#-----|             <params>: 
+# 2692|             getRequirement(0): [GTExpr,SimpleRequirementExpr] ... > ...
+# 2692|                 Type = [BoolType] bool
+# 2692|                 ValueCategory = prvalue
+# 2692|               getGreaterOperand(): [SizeofTypeOperator] sizeof(int)
+# 2692|                   Type = [LongType] unsigned long
+# 2692|                   Value = [SizeofTypeOperator] 4
+# 2692|                   ValueCategory = prvalue
+# 2692|               getLesserOperand(): [Literal] 0
+# 2692|                   Type = [IntType] int
+# 2692|                   Value = [Literal] 0
+# 2692|                   ValueCategory = prvalue
+# 2692|               getLesserOperand().getFullyConverted(): [CStyleCast] (unsigned long)...
+# 2692|                   Conversion = [IntegralConversion] integral conversion
+# 2692|                   Type = [LongType] unsigned long
+# 2692|                   ValueCategory = prvalue
+#-----|           getExpr().getFullyConverted(): [CStyleCast] (int)...
+#-----|               Conversion = [IntegralConversion] integral conversion
+#-----|               Type = [IntType] int
+#-----|               Value = [CStyleCast] 1
+#-----|               ValueCategory = prvalue
+# 2693|     getStmt(1): [ReturnStmt] return ...
+# 2693|       getExpr(): [VariableAccess] y
+# 2693|           Type = [IntType] int
+# 2693|           ValueCategory = prvalue(load)
 many-defs-per-use.cpp:
 #   34| [TopLevelFunction] void many_defs_per_use()
 #   34|   <params>: 

cpp/ql/test/library-tests/ir/ir/aliased_ir.expected

@@ -19099,6 +19099,24 @@ ir.cpp:
 # 2684|   Block 8
 # 2684|     v2684_14(void) = Unreached : 
 
+# 2691| int concepts::requires_use()
+# 2691|   Block 0
+# 2691|     v2691_1(void)       = EnterFunction            : 
+# 2691|     m2691_2(unknown)    = AliasedDefinition        : 
+# 2691|     m2691_3(unknown)    = InitializeNonLocal       : 
+# 2691|     m2691_4(unknown)    = Chi                      : total:m2691_2, partial:m2691_3
+# 2692|     r2692_1(glval<int>) = VariableAddress[y]       : 
+#-----|     r0_1(int)           = Constant[1]              : 
+#-----|     m0_2(int)           = Store[y]                 : &:r2692_1, r0_1
+# 2693|     r2693_1(glval<int>) = VariableAddress[#return] : 
+# 2693|     r2693_2(glval<int>) = VariableAddress[y]       : 
+# 2693|     r2693_3(int)        = Load[y]                  : &:r2693_2, m0_2
+# 2693|     m2693_4(int)        = Store[#return]           : &:r2693_1, r2693_3
+# 2691|     r2691_5(glval<int>) = VariableAddress[#return] : 
+# 2691|     v2691_6(void)       = ReturnValue              : &:r2691_5, m2693_4
+# 2691|     v2691_7(void)       = AliasedUse               : m2691_3
+# 2691|     v2691_8(void)       = ExitFunction             : 
+
 many-defs-per-use.cpp:
 #   34| void many_defs_per_use()
 #   34|   Block 0

cpp/ql/test/library-tests/ir/ir/ir.cpp

@@ -2686,4 +2686,13 @@ void test(bool b)
   twice_call_use(b ? "" : "");
 }
 
+namespace concepts {
+
+int requires_use() {
+  int y = requires { sizeof(int) > 0; };
+  return y;
+}
+
+}
+
 // semmle-extractor-options: -std=c++20 --clang

cpp/ql/test/library-tests/ir/ir/raw_ir.expected

@@ -17422,6 +17422,23 @@ ir.cpp:
 # 2684|     v2684_7(void) = AliasedUse   : ~m?
 # 2684|     v2684_8(void) = ExitFunction : 
 
+# 2691| int concepts::requires_use()
+# 2691|   Block 0
+# 2691|     v2691_1(void)       = EnterFunction            : 
+# 2691|     mu2691_2(unknown)   = AliasedDefinition        : 
+# 2691|     mu2691_3(unknown)   = InitializeNonLocal       : 
+# 2692|     r2692_1(glval<int>) = VariableAddress[y]       : 
+#-----|     r0_1(int)           = Constant[1]              : 
+#-----|     mu0_2(int)          = Store[y]                 : &:r2692_1, r0_1
+# 2693|     r2693_1(glval<int>) = VariableAddress[#return] : 
+# 2693|     r2693_2(glval<int>) = VariableAddress[y]       : 
+# 2693|     r2693_3(int)        = Load[y]                  : &:r2693_2, ~m?
+# 2693|     mu2693_4(int)       = Store[#return]           : &:r2693_1, r2693_3
+# 2691|     r2691_4(glval<int>) = VariableAddress[#return] : 
+# 2691|     v2691_5(void)       = ReturnValue              : &:r2691_4, ~m?
+# 2691|     v2691_6(void)       = AliasedUse               : ~m?
+# 2691|     v2691_7(void)       = ExitFunction             : 
+
 many-defs-per-use.cpp:
 #   34| void many_defs_per_use()
 #   34|   Block 0

cpp/ql/test/library-tests/macros/inmacroexpansion/inmacroexpansion.expected

@@ -2,25 +2,168 @@
 | file://:0:0:0:0 | (unnamed parameter 0) | false |
 | file://:0:0:0:0 | __super | false |
 | file://:0:0:0:0 | __va_list_tag | false |
+| file://:0:0:0:0 | decltype([...](...){...}) | false |
 | file://:0:0:0:0 | operator= | false |
 | file://:0:0:0:0 | operator= | false |
 | test.cpp:0:0:0:0 | test.cpp | false |
-| test.cpp:2:1:2:61 | #define FOO class S{int i; void f(void) { int j; return; } }; | false |
+| test.cpp:2:1:2:68 | #define CLASS_DECL class S{int i; void f(void) { int j; return; } }; | false |
 | test.cpp:4:1:4:1 | S | false |
 | test.cpp:4:1:4:1 | declaration of S | false |
 | test.cpp:4:1:4:1 | declaration of operator= | false |
 | test.cpp:4:1:4:1 | declaration of operator= | false |
 | test.cpp:4:1:4:1 | operator= | false |
 | test.cpp:4:1:4:1 | operator= | false |
-| test.cpp:4:1:4:3 | FOO | false |
+| test.cpp:4:1:4:10 | CLASS_DECL | false |
-| test.cpp:4:1:4:3 | S | false |
+| test.cpp:4:1:4:10 | S | false |
-| test.cpp:4:1:4:3 | declaration | true |
+| test.cpp:4:1:4:10 | declaration | true |
-| test.cpp:4:1:4:3 | definition of S | true |
+| test.cpp:4:1:4:10 | definition of S | true |
-| test.cpp:4:1:4:3 | definition of f | true |
+| test.cpp:4:1:4:10 | definition of f | true |
-| test.cpp:4:1:4:3 | definition of i | true |
+| test.cpp:4:1:4:10 | definition of i | true |
-| test.cpp:4:1:4:3 | definition of j | true |
+| test.cpp:4:1:4:10 | definition of j | true |
-| test.cpp:4:1:4:3 | f | false |
+| test.cpp:4:1:4:10 | f | false |
-| test.cpp:4:1:4:3 | i | false |
+| test.cpp:4:1:4:10 | i | false |
-| test.cpp:4:1:4:3 | j | true |
+| test.cpp:4:1:4:10 | j | true |
-| test.cpp:4:1:4:3 | return ... | true |
+| test.cpp:4:1:4:10 | return ... | true |
-| test.cpp:4:1:4:3 | { ... } | true |
+| test.cpp:4:1:4:10 | { ... } | true |
+| test.cpp:6:1:6:42 | #define FUNCTION_DECL void f1() { int k; } | false |
+| test.cpp:8:1:8:13 | FUNCTION_DECL | false |
+| test.cpp:8:1:8:13 | declaration | true |
+| test.cpp:8:1:8:13 | definition of f1 | true |
+| test.cpp:8:1:8:13 | definition of k | true |
+| test.cpp:8:1:8:13 | f1 | false |
+| test.cpp:8:1:8:13 | k | true |
+| test.cpp:8:1:8:13 | return ... | true |
+| test.cpp:8:1:8:13 | { ... } | true |
+| test.cpp:10:1:10:33 | #define VARIABLE_DECL int v1 = 1; | false |
+| test.cpp:12:1:12:13 | 1 | true |
+| test.cpp:12:1:12:13 | VARIABLE_DECL | false |
+| test.cpp:12:1:12:13 | definition of v1 | true |
+| test.cpp:12:1:12:13 | initializer for v1 | true |
+| test.cpp:12:1:12:13 | v1 | true |
+| test.cpp:14:1:14:35 | #define TYPE_DECL_1 typedef int t1; | false |
+| test.cpp:16:1:16:11 | TYPE_DECL_1 | false |
+| test.cpp:16:1:16:11 | declaration of t1 | true |
+| test.cpp:16:1:16:11 | t1 | false |
+| test.cpp:18:1:18:35 | #define TYPE_DECL_2 using t2 = int; | false |
+| test.cpp:20:1:20:11 | TYPE_DECL_2 | false |
+| test.cpp:20:1:20:11 | declaration of t2 | true |
+| test.cpp:20:1:20:11 | t2 | false |
+| test.cpp:22:1:22:47 | #define NAMESPACE_DECL namespace ns { int v2; } | false |
+| test.cpp:24:1:24:14 | NAMESPACE_DECL | false |
+| test.cpp:24:1:24:14 | definition of v2 | true |
+| test.cpp:24:1:24:14 | ns | false |
+| test.cpp:24:1:24:14 | ns | false |
+| test.cpp:24:1:24:14 | v2 | true |
+| test.cpp:26:1:26:43 | #define USING_NAMESPACE using namespace ns; | false |
+| test.cpp:28:1:28:34 | #define ENUM_CONSTANT enum_element | false |
+| test.cpp:30:12:30:21 | definition of enum_class | false |
+| test.cpp:30:12:30:21 | enum_class | false |
+| test.cpp:30:25:30:37 | ENUM_CONSTANT | false |
+| test.cpp:30:25:30:37 | enum_element | false |
+| test.cpp:32:1:32:41 | #define USING_ENUM using enum enum_class; | false |
+| test.cpp:34:1:34:10 | USING_ENUM | false |
+| test.cpp:34:1:34:10 | using enum enum_class | false |
+| test.cpp:36:1:36:48 | #define STATIC_ASSERT static_assert(1 == 1, ""); | false |
+| test.cpp:38:1:38:13 | 1 | true |
+| test.cpp:38:1:38:13 | 1 | true |
+| test.cpp:38:1:38:13 | ... == ... | true |
+| test.cpp:38:1:38:13 | STATIC_ASSERT | false |
+| test.cpp:38:1:38:13 | static_assert(..., "") | false |
+| test.cpp:40:1:40:42 | #define ATTRIBUTE [[nodiscard("reason1")]] | false |
+| test.cpp:42:1:42:9 | ATTRIBUTE | false |
+| test.cpp:42:1:42:9 | nodiscard | false |
+| test.cpp:42:1:42:9 | reason1 | false |
+| test.cpp:42:1:42:9 | reason1 | true |
+| test.cpp:43:5:43:6 | declaration of f2 | false |
+| test.cpp:43:5:43:6 | f2 | false |
+| test.cpp:45:1:45:31 | #define ATTRIBUTE_ARG "reason2" | false |
+| test.cpp:47:3:47:11 | nodiscard | false |
+| test.cpp:47:13:47:25 | ATTRIBUTE_ARG | false |
+| test.cpp:47:13:47:25 | reason2 | false |
+| test.cpp:47:13:47:25 | reason2 | true |
+| test.cpp:48:5:48:6 | declaration of f3 | false |
+| test.cpp:48:5:48:6 | f3 | false |
+| test.cpp:50:1:50:16 | #define TYPE int | false |
+| test.cpp:52:1:52:4 | TYPE | false |
+| test.cpp:52:6:52:7 | definition of v3 | true |
+| test.cpp:52:6:52:7 | v3 | true |
+| test.cpp:52:11:52:11 | 1 | false |
+| test.cpp:52:11:52:11 | initializer for v3 | false |
+| test.cpp:54:1:54:29 | #define DERIVATION : public S | false |
+| test.cpp:56:7:56:7 | T | false |
+| test.cpp:56:7:56:7 | T | false |
+| test.cpp:56:7:56:7 | declaration of T | false |
+| test.cpp:56:7:56:7 | declaration of operator= | false |
+| test.cpp:56:7:56:7 | declaration of operator= | false |
+| test.cpp:56:7:56:7 | definition of T | false |
+| test.cpp:56:7:56:7 | operator= | false |
+| test.cpp:56:7:56:7 | operator= | false |
+| test.cpp:56:9:56:18 | DERIVATION | false |
+| test.cpp:56:9:56:18 | derivation | false |
+| test.cpp:58:1:58:31 | #define FRIEND friend int f3(); | false |
+| test.cpp:60:7:60:7 | U | false |
+| test.cpp:60:7:60:7 | declaration of operator= | false |
+| test.cpp:60:7:60:7 | declaration of operator= | false |
+| test.cpp:60:7:60:7 | definition of U | false |
+| test.cpp:60:7:60:7 | operator= | false |
+| test.cpp:60:7:60:7 | operator= | false |
+| test.cpp:61:3:61:8 | FRIEND | false |
+| test.cpp:61:3:61:8 | U's friend | false |
+| test.cpp:64:1:64:24 | #define NAME_QUAL_1 ns:: | false |
+| test.cpp:66:1:66:22 | #define NAME_QUAL_2 ns | false |
+| test.cpp:68:1:68:19 | #define LOCAL_VAR m | false |
+| test.cpp:70:6:70:7 | definition of f4 | false |
+| test.cpp:70:6:70:7 | f4 | false |
+| test.cpp:70:11:76:1 | { ... } | false |
+| test.cpp:71:5:71:8 | ns:: | false |
+| test.cpp:71:5:71:15 | NAME_QUAL_1 | false |
+| test.cpp:71:5:71:18 | v2 | false |
+| test.cpp:71:5:71:19 | ExprStmt | false |
+| test.cpp:72:5:72:8 | ns:: | false |
+| test.cpp:72:5:72:15 | NAME_QUAL_2 | false |
+| test.cpp:72:5:72:21 | v2 | false |
+| test.cpp:72:5:72:22 | ExprStmt | false |
+| test.cpp:73:5:73:23 | declaration | false |
+| test.cpp:73:9:73:17 | LOCAL_VAR | false |
+| test.cpp:73:9:73:17 | definition of m | true |
+| test.cpp:73:9:73:17 | m | true |
+| test.cpp:73:20:73:22 | 42 | false |
+| test.cpp:73:20:73:22 | initializer for m | false |
+| test.cpp:74:5:74:41 | declaration | false |
+| test.cpp:74:10:74:10 | definition of l | false |
+| test.cpp:74:10:74:10 | l | false |
+| test.cpp:74:13:74:40 | [...](...){...} | false |
+| test.cpp:74:13:74:40 | initializer for l | false |
+| test.cpp:74:13:74:40 | {...} | false |
+| test.cpp:74:14:74:14 | (unnamed constructor) | false |
+| test.cpp:74:14:74:14 | (unnamed constructor) | false |
+| test.cpp:74:14:74:14 | (unnamed constructor) | false |
+| test.cpp:74:14:74:14 | declaration of (unnamed constructor) | false |
+| test.cpp:74:14:74:14 | declaration of (unnamed constructor) | false |
+| test.cpp:74:14:74:14 | definition of (unnamed constructor) | false |
+| test.cpp:74:14:74:14 | definition of operator= | false |
+| test.cpp:74:14:74:14 | operator= | false |
+| test.cpp:74:15:74:15 | definition of m | false |
+| test.cpp:74:15:74:15 | m | false |
+| test.cpp:74:15:74:15 | m | false |
+| test.cpp:74:15:74:23 | LOCAL_VAR | false |
+| test.cpp:74:15:74:23 | m | true |
+| test.cpp:74:25:74:25 | definition of operator() | false |
+| test.cpp:74:25:74:25 | operator() | false |
+| test.cpp:74:28:74:40 | { ... } | false |
+| test.cpp:74:30:74:38 | return ... | false |
+| test.cpp:74:37:74:37 | (int)... | false |
+| test.cpp:75:5:75:5 | (const lambda [] type at line 74, col. 14)... | false |
+| test.cpp:75:5:75:5 | l | false |
+| test.cpp:75:5:75:8 | ExprStmt | false |
+| test.cpp:75:6:75:6 | call to operator() | false |
+| test.cpp:76:1:76:1 | return ... | false |
+| test.cpp:78:1:78:15 | #define ID(x) x | false |
+| test.cpp:79:1:79:23 | #define NESTED(x) ID(x) | false |
+| test.cpp:80:5:80:6 | definition of v4 | false |
+| test.cpp:80:5:80:6 | v4 | false |
+| test.cpp:80:10:80:18 | ID(x) | false |
+| test.cpp:80:10:80:18 | NESTED(x) | false |
+| test.cpp:80:17:80:17 | 1 | true |
+| test.cpp:80:17:80:17 | initializer for v4 | true |
+| test.cpp:82:1:82:39 | // semmle-extractor-options: -std=c++20 | false |

cpp/ql/test/library-tests/macros/inmacroexpansion/test.cpp

@@ -1,5 +1,82 @@
 
-#define FOO class S{int i; void f(void) { int j; return; } };
+#define CLASS_DECL class S{int i; void f(void) { int j; return; } };
 
-FOO
+CLASS_DECL
 
+#define FUNCTION_DECL void f1() { int k; }
+
+FUNCTION_DECL
+
+#define VARIABLE_DECL int v1 = 1;
+
+VARIABLE_DECL
+
+#define TYPE_DECL_1 typedef int t1;
+
+TYPE_DECL_1
+
+#define TYPE_DECL_2 using t2 = int;
+
+TYPE_DECL_2
+
+#define NAMESPACE_DECL namespace ns { int v2; }
+
+NAMESPACE_DECL
+
+#define USING_NAMESPACE using namespace ns;
+
+#define ENUM_CONSTANT enum_element
+
+enum class enum_class { ENUM_CONSTANT };
+
+#define USING_ENUM using enum enum_class;
+
+USING_ENUM
+
+#define STATIC_ASSERT static_assert(1 == 1, "");
+
+STATIC_ASSERT
+
+#define ATTRIBUTE [[nodiscard("reason1")]]
+
+ATTRIBUTE
+int f2();
+
+#define ATTRIBUTE_ARG "reason2"
+
+[[nodiscard(ATTRIBUTE_ARG)]]
+int f3();
+
+#define TYPE int
+
+TYPE v3 = 1;
+
+#define DERIVATION : public S
+
+class T DERIVATION {};
+
+#define FRIEND friend int f3();
+
+class U {
+  FRIEND
+};
+
+#define NAME_QUAL_1 ns::
+
+#define NAME_QUAL_2 ns
+
+#define LOCAL_VAR m
+
+void f4() {
+    NAME_QUAL_1 v2;
+    NAME_QUAL_2 :: v2;
+    int LOCAL_VAR = 42;
+    auto l = [LOCAL_VAR]() { return m; };
+    l();
+}
+
+#define ID(x) x
+#define NESTED(x) ID(x)
+int v4 = NESTED(1);
+
+// semmle-extractor-options: -std=c++20

cpp/ql/test/query-tests/Diagnostics/Info.expected

@@ -1,6 +1,6 @@
-| containserror.cpp:0:0:0:0 | containserror.cpp | query-tests/Diagnostics/containserror.cpp | fromSource, normalTermination |
+| containserror.cpp:0:0:0:0 | containserror.cpp | containserror.cpp | fromSource, normalTermination |
-| containswarning.cpp:0:0:0:0 | containswarning.cpp | query-tests/Diagnostics/containswarning.cpp | fromSource, normalTermination |
+| containswarning.cpp:0:0:0:0 | containswarning.cpp | containswarning.cpp | fromSource, normalTermination |
-| doesnotcompile.cpp:0:0:0:0 | doesnotcompile.cpp | query-tests/Diagnostics/doesnotcompile.cpp | ExtractionProblem (severity 1), fromSource, normalTermination |
+| doesnotcompile.cpp:0:0:0:0 | doesnotcompile.cpp | doesnotcompile.cpp | ExtractionProblem (severity 1), fromSource, normalTermination |
 | file://:0:0:0:0 |  |  |  |
-| header.h:0:0:0:0 | header.h | query-tests/Diagnostics/header.h | fromSource |
+| header.h:0:0:0:0 | header.h | header.h | fromSource |
-| successful.cpp:0:0:0:0 | successful.cpp | query-tests/Diagnostics/successful.cpp | fromSource, normalTermination |
+| successful.cpp:0:0:0:0 | successful.cpp | successful.cpp | fromSource, normalTermination |

cpp/ql/test/query-tests/Documentation/DocumentApi/DocumentApi.expected

@@ -1,16 +1,16 @@
-| comment_prototypes.c:29:6:29:11 | proto6 | Functions called from other files should be documented (called from $@). | comment_prototypes_caller1.c:9:5:9:10 | call to proto6 | query-tests/Documentation/DocumentApi/comment_prototypes_caller1.c |
+| comment_prototypes.c:29:6:29:11 | proto6 | Functions called from other files should be documented (called from $@). | comment_prototypes_caller1.c:9:5:9:10 | call to proto6 | comment_prototypes_caller1.c |
-| comment_prototypes.c:29:6:29:11 | proto6 | Functions called from other files should be documented (called from $@). | comment_prototypes_caller2.c:9:5:9:10 | call to proto6 | query-tests/Documentation/DocumentApi/comment_prototypes_caller2.c |
+| comment_prototypes.c:29:6:29:11 | proto6 | Functions called from other files should be documented (called from $@). | comment_prototypes_caller2.c:9:5:9:10 | call to proto6 | comment_prototypes_caller2.c |
-| comment_prototypes.c:34:6:34:11 | proto7 | Functions called from other files should be documented (called from $@). | comment_prototypes_caller1.c:10:5:10:10 | call to proto7 | query-tests/Documentation/DocumentApi/comment_prototypes_caller1.c |
+| comment_prototypes.c:34:6:34:11 | proto7 | Functions called from other files should be documented (called from $@). | comment_prototypes_caller1.c:10:5:10:10 | call to proto7 | comment_prototypes_caller1.c |
-| comment_prototypes.c:34:6:34:11 | proto7 | Functions called from other files should be documented (called from $@). | comment_prototypes_caller2.c:10:5:10:10 | call to proto7 | query-tests/Documentation/DocumentApi/comment_prototypes_caller2.c |
+| comment_prototypes.c:34:6:34:11 | proto7 | Functions called from other files should be documented (called from $@). | comment_prototypes_caller2.c:10:5:10:10 | call to proto7 | comment_prototypes_caller2.c |
-| comment_prototypes.c:45:6:45:11 | proto9 | Functions called from other files should be documented (called from $@). | comment_prototypes_caller1.c:12:5:12:10 | call to proto9 | query-tests/Documentation/DocumentApi/comment_prototypes_caller1.c |
+| comment_prototypes.c:45:6:45:11 | proto9 | Functions called from other files should be documented (called from $@). | comment_prototypes_caller1.c:12:5:12:10 | call to proto9 | comment_prototypes_caller1.c |
-| comment_prototypes.c:45:6:45:11 | proto9 | Functions called from other files should be documented (called from $@). | comment_prototypes_caller2.c:12:5:12:10 | call to proto9 | query-tests/Documentation/DocumentApi/comment_prototypes_caller2.c |
+| comment_prototypes.c:45:6:45:11 | proto9 | Functions called from other files should be documented (called from $@). | comment_prototypes_caller2.c:12:5:12:10 | call to proto9 | comment_prototypes_caller2.c |
-| comment_prototypes.c:50:6:50:12 | proto10 | Functions called from other files should be documented (called from $@). | comment_prototypes_caller1.c:13:5:13:11 | call to proto10 | query-tests/Documentation/DocumentApi/comment_prototypes_caller1.c |
+| comment_prototypes.c:50:6:50:12 | proto10 | Functions called from other files should be documented (called from $@). | comment_prototypes_caller1.c:13:5:13:11 | call to proto10 | comment_prototypes_caller1.c |
-| comment_prototypes.c:50:6:50:12 | proto10 | Functions called from other files should be documented (called from $@). | comment_prototypes_caller2.c:13:5:13:11 | call to proto10 | query-tests/Documentation/DocumentApi/comment_prototypes_caller2.c |
+| comment_prototypes.c:50:6:50:12 | proto10 | Functions called from other files should be documented (called from $@). | comment_prototypes_caller2.c:13:5:13:11 | call to proto10 | comment_prototypes_caller2.c |
-| comment_prototypes.c:55:6:55:12 | proto11 | Functions called from other files should be documented (called from $@). | comment_prototypes_caller1.c:14:5:14:11 | call to proto11 | query-tests/Documentation/DocumentApi/comment_prototypes_caller1.c |
+| comment_prototypes.c:55:6:55:12 | proto11 | Functions called from other files should be documented (called from $@). | comment_prototypes_caller1.c:14:5:14:11 | call to proto11 | comment_prototypes_caller1.c |
-| comment_prototypes.c:55:6:55:12 | proto11 | Functions called from other files should be documented (called from $@). | comment_prototypes_caller2.c:14:5:14:11 | call to proto11 | query-tests/Documentation/DocumentApi/comment_prototypes_caller2.c |
+| comment_prototypes.c:55:6:55:12 | proto11 | Functions called from other files should be documented (called from $@). | comment_prototypes_caller2.c:14:5:14:11 | call to proto11 | comment_prototypes_caller2.c |
-| comment_prototypes.c:66:6:66:12 | proto13 | Functions called from other files should be documented (called from $@). | comment_prototypes_caller1.c:16:5:16:11 | call to proto13 | query-tests/Documentation/DocumentApi/comment_prototypes_caller1.c |
+| comment_prototypes.c:66:6:66:12 | proto13 | Functions called from other files should be documented (called from $@). | comment_prototypes_caller1.c:16:5:16:11 | call to proto13 | comment_prototypes_caller1.c |
-| comment_prototypes.c:66:6:66:12 | proto13 | Functions called from other files should be documented (called from $@). | comment_prototypes_caller2.c:16:5:16:11 | call to proto13 | query-tests/Documentation/DocumentApi/comment_prototypes_caller2.c |
+| comment_prototypes.c:66:6:66:12 | proto13 | Functions called from other files should be documented (called from $@). | comment_prototypes_caller2.c:16:5:16:11 | call to proto13 | comment_prototypes_caller2.c |
-| definition.c:2:6:2:7 | f1 | Functions called from other files should be documented (called from $@). | user1.c:9:5:9:6 | call to f1 | query-tests/Documentation/DocumentApi/user1.c |
+| definition.c:2:6:2:7 | f1 | Functions called from other files should be documented (called from $@). | user1.c:9:5:9:6 | call to f1 | user1.c |
-| definition.c:2:6:2:7 | f1 | Functions called from other files should be documented (called from $@). | user2.c:7:5:7:6 | call to f1 | query-tests/Documentation/DocumentApi/user2.c |
+| definition.c:2:6:2:7 | f1 | Functions called from other files should be documented (called from $@). | user2.c:7:5:7:6 | call to f1 | user2.c |
-| definition.c:32:6:32:7 | f6 | Functions called from other files should be documented (called from $@). | user1.c:14:5:14:6 | call to f6 | query-tests/Documentation/DocumentApi/user1.c |
+| definition.c:32:6:32:7 | f6 | Functions called from other files should be documented (called from $@). | user1.c:14:5:14:6 | call to f6 | user1.c |
-| definition.c:32:6:32:7 | f6 | Functions called from other files should be documented (called from $@). | user2.c:10:5:10:6 | call to f6 | query-tests/Documentation/DocumentApi/user2.c |
+| definition.c:32:6:32:7 | f6 | Functions called from other files should be documented (called from $@). | user2.c:10:5:10:6 | call to f6 | user2.c |

cpp/ql/test/query-tests/Metrics/Dependencies/ExternalDependencies.expected

@@ -1,4 +1,4 @@
-| /query-tests/Metrics/Dependencies/main.cpp<\|>LibC<\|>unknown | 5 |
+| /main.cpp<\|>LibC<\|>unknown | 5 |
-| /query-tests/Metrics/Dependencies/include.h<\|>LibD<\|>unknown | 1 |
+| /include.h<\|>LibD<\|>unknown | 1 |
-| /query-tests/Metrics/Dependencies/main.cpp<\|>LibA<\|>unknown | 1 |
+| /main.cpp<\|>LibA<\|>unknown | 1 |
-| /query-tests/Metrics/Dependencies/main.cpp<\|>LibB<\|>unknown | 1 |
+| /main.cpp<\|>LibB<\|>unknown | 1 |

cpp/ql/test/query-tests/Metrics/Dependencies/ExternalDependenciesSourceLinks.expected

@@ -1,4 +1,4 @@
-| /query-tests/Metrics/Dependencies/include.h<\|>LibD<\|>unknown | include.h:0:0:0:0 | include.h |
+| /include.h<\|>LibD<\|>unknown | include.h:0:0:0:0 | include.h |
-| /query-tests/Metrics/Dependencies/main.cpp<\|>LibA<\|>unknown | main.cpp:0:0:0:0 | main.cpp |
+| /main.cpp<\|>LibA<\|>unknown | main.cpp:0:0:0:0 | main.cpp |
-| /query-tests/Metrics/Dependencies/main.cpp<\|>LibB<\|>unknown | main.cpp:0:0:0:0 | main.cpp |
+| /main.cpp<\|>LibB<\|>unknown | main.cpp:0:0:0:0 | main.cpp |
-| /query-tests/Metrics/Dependencies/main.cpp<\|>LibC<\|>unknown | main.cpp:0:0:0:0 | main.cpp |
+| /main.cpp<\|>LibC<\|>unknown | main.cpp:0:0:0:0 | main.cpp |

cpp/ql/test/query-tests/Security/CWE/CWE-129/semmle/ImproperArrayIndexValidation/ImproperArrayIndexValidation.expected

@@ -2,29 +2,36 @@ edges
 | test1.c:7:26:7:29 | **argv | test1.c:8:11:8:14 | call to atoi | provenance | TaintFunction |
 | test1.c:8:11:8:14 | call to atoi | test1.c:9:9:9:9 | i | provenance |  |
 | test1.c:8:11:8:14 | call to atoi | test1.c:11:9:11:9 | i | provenance |  |
+| test1.c:8:11:8:14 | call to atoi | test1.c:12:9:12:9 | i | provenance |  |
 | test1.c:8:11:8:14 | call to atoi | test1.c:13:9:13:9 | i | provenance |  |
-| test1.c:9:9:9:9 | i | test1.c:16:16:16:16 | i | provenance |  |
+| test1.c:9:9:9:9 | i | test1.c:18:16:18:16 | i | provenance |  |
-| test1.c:11:9:11:9 | i | test1.c:32:16:32:16 | i | provenance |  |
+| test1.c:11:9:11:9 | i | test1.c:34:16:34:16 | i | provenance |  |
-| test1.c:13:9:13:9 | i | test1.c:48:16:48:16 | i | provenance |  |
+| test1.c:12:9:12:9 | i | test1.c:42:16:42:16 | i | provenance |  |
-| test1.c:16:16:16:16 | i | test1.c:18:16:18:16 | i | provenance |  |
+| test1.c:13:9:13:9 | i | test1.c:50:16:50:16 | i | provenance |  |
-| test1.c:32:16:32:16 | i | test1.c:33:11:33:11 | i | provenance |  |
+| test1.c:18:16:18:16 | i | test1.c:20:16:20:16 | i | provenance |  |
-| test1.c:48:16:48:16 | i | test1.c:51:3:51:7 | ... = ... | provenance |  |
+| test1.c:34:16:34:16 | i | test1.c:35:11:35:11 | i | provenance |  |
-| test1.c:51:3:51:7 | ... = ... | test1.c:53:15:53:15 | j | provenance |  |
+| test1.c:42:16:42:16 | i | test1.c:43:11:43:11 | i | provenance |  |
+| test1.c:50:16:50:16 | i | test1.c:53:3:53:7 | ... = ... | provenance |  |
+| test1.c:53:3:53:7 | ... = ... | test1.c:55:15:55:15 | j | provenance |  |
 nodes
 | test1.c:7:26:7:29 | **argv | semmle.label | **argv |
 | test1.c:8:11:8:14 | call to atoi | semmle.label | call to atoi |
 | test1.c:9:9:9:9 | i | semmle.label | i |
 | test1.c:11:9:11:9 | i | semmle.label | i |
+| test1.c:12:9:12:9 | i | semmle.label | i |
 | test1.c:13:9:13:9 | i | semmle.label | i |
-| test1.c:16:16:16:16 | i | semmle.label | i |
 | test1.c:18:16:18:16 | i | semmle.label | i |
-| test1.c:32:16:32:16 | i | semmle.label | i |
+| test1.c:20:16:20:16 | i | semmle.label | i |
-| test1.c:33:11:33:11 | i | semmle.label | i |
+| test1.c:34:16:34:16 | i | semmle.label | i |
-| test1.c:48:16:48:16 | i | semmle.label | i |
+| test1.c:35:11:35:11 | i | semmle.label | i |
-| test1.c:51:3:51:7 | ... = ... | semmle.label | ... = ... |
+| test1.c:42:16:42:16 | i | semmle.label | i |
-| test1.c:53:15:53:15 | j | semmle.label | j |
+| test1.c:43:11:43:11 | i | semmle.label | i |
+| test1.c:50:16:50:16 | i | semmle.label | i |
+| test1.c:53:3:53:7 | ... = ... | semmle.label | ... = ... |
+| test1.c:55:15:55:15 | j | semmle.label | j |
 subpaths
 #select
-| test1.c:18:16:18:16 | i | test1.c:7:26:7:29 | **argv | test1.c:18:16:18:16 | i | An array indexing expression depends on $@ that might be outside the bounds of the array. | test1.c:7:26:7:29 | **argv | a command-line argument |
+| test1.c:20:16:20:16 | i | test1.c:7:26:7:29 | **argv | test1.c:20:16:20:16 | i | An array indexing expression depends on $@ that might be outside the bounds of the array. | test1.c:7:26:7:29 | **argv | a command-line argument |
-| test1.c:33:11:33:11 | i | test1.c:7:26:7:29 | **argv | test1.c:33:11:33:11 | i | An array indexing expression depends on $@ that might be outside the bounds of the array. | test1.c:7:26:7:29 | **argv | a command-line argument |
+| test1.c:35:11:35:11 | i | test1.c:7:26:7:29 | **argv | test1.c:35:11:35:11 | i | An array indexing expression depends on $@ that might be outside the bounds of the array. | test1.c:7:26:7:29 | **argv | a command-line argument |
-| test1.c:53:15:53:15 | j | test1.c:7:26:7:29 | **argv | test1.c:53:15:53:15 | j | An array indexing expression depends on $@ that might be outside the bounds of the array. | test1.c:7:26:7:29 | **argv | a command-line argument |
+| test1.c:43:11:43:11 | i | test1.c:7:26:7:29 | **argv | test1.c:43:11:43:11 | i | An array indexing expression depends on $@ that might be outside the bounds of the array. | test1.c:7:26:7:29 | **argv | a command-line argument |
+| test1.c:55:15:55:15 | j | test1.c:7:26:7:29 | **argv | test1.c:55:15:55:15 | j | An array indexing expression depends on $@ that might be outside the bounds of the array. | test1.c:7:26:7:29 | **argv | a command-line argument |

cpp/ql/test/query-tests/Security/CWE/CWE-129/semmle/ImproperArrayIndexValidation/test1.c

@@ -11,6 +11,8 @@ int main(int argc, char *argv[]) {
   test3(i);
   test4(i);
   test5(i);
+  test6(i);
+  test7(argv[1]);
 }
 
 void test1(int i) {
@@ -38,7 +40,7 @@ void test3(int i) {
 }
 
 void test4(int i) {
-  myArray[i] = 0; // BAD: i has not been validated [NOT REPORTED]
+  myArray[i] = 0; // BAD: i has not been validated
 
   if ((i < 0) || (i >= 10)) return;
 
@@ -52,3 +54,26 @@ void test5(int i) {
 
   j = myArray[j]; // BAD: j has not been validated
 }
+
+extern int myTable[256];
+
+void test6(int i) {
+  unsigned char s = i;
+
+  myTable[s] = 0; // GOOD: Input is small [FALSE POSITIVE]
+}
+
+typedef void FILE;
+#define EOF (-1)
+
+int getc(FILE*);
+
+extern int myMaxCharTable[256];
+
+void test7(FILE* fp) {
+  int ch;
+  while ((ch = getc(fp)) != EOF) {
+    myMaxCharTable[ch] = 0; // GOOD
+  }
+}
+

cpp/ql/test/query-tests/Security/CWE/CWE-131/NoSpaceForZeroTerminator/test.c

@@ -11,14 +11,14 @@ char *strcpy(char *s1, const char *s2);
 
 //// Test code /////
 
-void bad0(char *str) {
+static void bad0(char *str) {
     // BAD -- Not allocating space for '\0' terminator
     char *buffer = malloc(strlen(str));
     strcpy(buffer, str);
     free(buffer);
 }
 
-void good0(char *str) {
+static void good0(char *str) {
     // GOOD -- Allocating extra byte for terminator
     char *buffer = malloc(strlen(str)+1);
     strcpy(buffer, str);
@@ -26,7 +26,7 @@ void good0(char *str) {
 }
 
 
-void bad1(char *str) {
+static void bad1(char *str) {
     int len = strlen(str);
     // BAD -- Not allocating space for '\0' terminator
     char *buffer = malloc(len);
@@ -34,7 +34,7 @@ void bad1(char *str) {
     free(buffer);
 }
 
-void good1(char *str) {
+static void good1(char *str) {
     int len = strlen(str);
     // GOOD -- Allocating extra byte for terminator
     char *buffer = malloc(len+1);
@@ -43,7 +43,7 @@ void good1(char *str) {
 }
 
 
-void bad2(char *str) {
+static void bad2(char *str) {
     int len = strlen(str);
     // BAD -- Not allocating space for '\0' terminator
     char *buffer = malloc(len);
@@ -51,7 +51,7 @@ void bad2(char *str) {
     free(buffer);
 }
 
-void good2(char *str) {
+static void good2(char *str) {
     int len = strlen(str)+1;
     // GOOD -- Allocating extra byte for terminator
     char *buffer = malloc(len);
@@ -59,14 +59,14 @@ void good2(char *str) {
     free(buffer);
 }
 
-void bad3(char *str) {
+static void bad3(char *str) {
     // BAD -- Not allocating space for '\0' terminator
     char *buffer = malloc(strlen(str) * sizeof(char));
     strcpy(buffer, str);
     free(buffer);
 }
 
-void good3(char *str) {
+static void good3(char *str) {
     // GOOD -- Allocating extra byte for terminator
     char *buffer = malloc((strlen(str) + 1) * sizeof(char));
     strcpy(buffer, str);
@@ -75,7 +75,7 @@ void good3(char *str) {
 
 void *memcpy(void *s1, const void *s2, size_t n);
 
-void good4(char *str) {
+static void good4(char *str) {
 	// GOOD -- allocating a non zero-terminated string
 	int len = strlen(str);
 	char *buffer = malloc(len);

cpp/ql/test/query-tests/Security/CWE/CWE-131/NoSpaceForZeroTerminator/test.cpp

@@ -19,28 +19,28 @@ int strcmp(const char *s1, const char *s2);
 
 //// Test code /////
 
-void bad1(wchar_t *wstr) {
+static void bad1(wchar_t *wstr) {
     // BAD -- Not allocating space for '\0' terminator
     wchar_t *wbuffer = (wchar_t *)malloc(wcslen(wstr));
     wcscpy(wbuffer, wstr);
     free(wbuffer);
 }
 
-void bad2(wchar_t *wstr) {
+static void bad2(wchar_t *wstr) {
     // BAD -- Not allocating space for '\0' terminator
     wchar_t *wbuffer = (wchar_t *)malloc(wcslen(wstr) * sizeof(wchar_t));
     wcscpy(wbuffer, wstr);
     free(wbuffer);
 }
 
-void good1(wchar_t *wstr) {
+static void good1(wchar_t *wstr) {
     // GOOD -- Allocating extra character for terminator
     wchar_t *wbuffer = (wchar_t *)malloc((wcslen(wstr) + 1) * sizeof(wchar_t));
     wcscpy(wbuffer, wstr);
     free(wbuffer);
 }
 
-void bad3(char *str) {
+static void bad3(char *str) {
     // BAD -- zero-termination proved by sprintf (as destination)
     char *buffer = (char *)malloc(strlen(str));
     sprintf(buffer, "%s", str);
@@ -50,7 +50,7 @@ void bad3(char *str) {
 void decode(char *dest, char *src);
 void wdecode(wchar_t *dest, wchar_t *src);
 
-void bad4(char *str) {
+static void bad4(char *str) {
     // BAD -- zero-termination proved by wprintf (as parameter)
     char *buffer = (char *)malloc(strlen(str));
     decode(buffer, str);
@@ -58,7 +58,7 @@ void bad4(char *str) {
     free(buffer);
 }
 
-void bad5(char *str) {
+static void bad5(char *str) {
     // BAD -- zero-termination proved by strcat (as destination)
     char *buffer = (char *)malloc(strlen(str));
     buffer[0] = 0;
@@ -66,7 +66,7 @@ void bad5(char *str) {
     free(buffer);
 }
 
-void bad6(char *str, char *dest) {
+static void bad6(char *str, char *dest) {
     // BAD -- zero-termination proved by strcat (as source)
     char *buffer = (char *)malloc(strlen(str));
     decode(buffer, str);
@@ -74,7 +74,7 @@ void bad6(char *str, char *dest) {
     free(buffer);
 }
 
-void bad7(char *str, char *str2) {
+static void bad7(char *str, char *str2) {
     // BAD -- zero-termination proved by strcmp
     char *buffer = (char *)malloc(strlen(str));
     decode(buffer, str);
@@ -84,7 +84,7 @@ void bad7(char *str, char *str2) {
     free(buffer);
 }
 
-void bad8(wchar_t *str) {
+static void bad8(wchar_t *str) {
     // BAD -- zero-termination proved by wcslen
     wchar_t *wbuffer = (wchar_t *)malloc(wcslen(str));
     wdecode(wbuffer, str);
@@ -94,21 +94,21 @@ void bad8(wchar_t *str) {
     free(wbuffer);
 }
 
-void good2(char *str, char *dest) {
+static void good2(char *str, char *dest) {
     // GOOD -- zero-termination not proven
     char *buffer = (char *)malloc(strlen(str));
     decode(buffer, str);
     free(buffer);
 }
 
-void bad9(wchar_t *wstr) {
+static void bad9(wchar_t *wstr) {
     // BAD -- using new
     wchar_t *wbuffer = new wchar_t[wcslen(wstr)];
     wcscpy(wbuffer, wstr);
     delete wbuffer;
 }
 
-void good3(char *str) {
+static void good3(char *str) {
     // GOOD -- zero-termination not required for this printf
     char *buffer = (char *)malloc(strlen(str));
     decode(buffer, str);
@@ -116,7 +116,7 @@ void good3(char *str) {
     free(buffer);
 }
 
-void good4(char *str) {
+static void good4(char *str) {
     // GOOD -- zero-termination not required for this printf
     char *buffer = (char *)malloc(strlen(str));
     decode(buffer, str);

cpp/ql/test/query-tests/Security/CWE/CWE-131/NoSpaceForZeroTerminator/test2.cpp

@@ -38,35 +38,35 @@ namespace std
 
 //// Test code /////
 
-void bad1(char *str) {
+static void bad1(char *str) {
     // BAD -- Not allocating space for '\0' terminator [NOT DETECTED]
     char *buffer = (char *)malloc(strlen(str));
     std::string str2(buffer);
     free(buffer);
 }
 
-void good1(char *str) {
+static void good1(char *str) {
     // GOOD --- copy does not overrun due to size limit
     char *buffer = (char *)malloc(strlen(str));
     std::string str2(buffer, strlen(str));
     free(buffer);
 }
 
-void bad2(wchar_t *str) {
+static void bad2(wchar_t *str) {
     // BAD -- Not allocating space for '\0' terminator [NOT DETECTED]
     wchar_t *buffer = (wchar_t *)calloc(wcslen(str), sizeof(wchar_t));
     wcscpy(buffer, str);
     free(buffer);
 }
 
-void bad3(wchar_t *str) {
+static void bad3(wchar_t *str) {
     // BAD -- Not allocating space for '\0' terminator
     wchar_t *buffer = (wchar_t *)calloc(sizeof(wchar_t), wcslen(str));
     wcscpy(buffer, str);
     free(buffer);
 }
 
-void bad4(char *str) {
+static void bad4(char *str) {
     // BAD -- Not allocating space for '\0' terminator
     char *buffer = (char *)realloc(0, strlen(str));
     strcpy(buffer, str);

cpp/ql/test/query-tests/Security/CWE/CWE-190/semmle/tainted/ArithmeticTainted.expected

@@ -1,4 +1,6 @@
 edges
+| main.cpp:7:27:7:30 | **argv | main.cpp:8:17:8:20 | **argv | provenance |  |
+| main.cpp:8:17:8:20 | **argv | test.c:10:28:10:31 | **argv | provenance |  |
 | test2.cpp:12:21:12:21 | v | test2.cpp:14:11:14:11 | v | provenance |  |
 | test2.cpp:25:22:25:23 | fscanf output argument | test2.cpp:27:13:27:13 | v | provenance |  |
 | test2.cpp:27:13:27:13 | v | test2.cpp:12:21:12:21 | v | provenance |  |
@@ -6,18 +8,15 @@ edges
 | test2.cpp:38:13:38:16 | call to atoi | test2.cpp:39:3:39:18 | ... = ... | provenance |  |
 | test2.cpp:38:13:38:16 | call to atoi | test2.cpp:39:9:39:11 | num | provenance |  |
 | test2.cpp:39:3:39:18 | ... = ... | test2.cpp:40:3:40:5 | num | provenance |  |
-| test3.c:10:27:10:30 | **argv | test.c:11:24:11:27 | call to atoi | provenance | TaintFunction |
-| test3.c:10:27:10:30 | **argv | test.c:41:5:41:24 | ... = ... | provenance | TaintFunction |
-| test3.c:10:27:10:30 | **argv | test.c:51:5:51:24 | ... = ... | provenance | TaintFunction |
 | test5.cpp:5:5:5:17 | *getTaintedInt | test5.cpp:17:6:17:18 | call to getTaintedInt | provenance |  |
 | test5.cpp:5:5:5:17 | *getTaintedInt | test5.cpp:18:6:18:18 | call to getTaintedInt | provenance |  |
 | test5.cpp:9:7:9:9 | gets output argument | test5.cpp:10:9:10:27 | call to strtoul | provenance | TaintFunction |
 | test5.cpp:10:9:10:27 | call to strtoul | test5.cpp:5:5:5:17 | *getTaintedInt | provenance |  |
 | test5.cpp:18:2:18:20 | ... = ... | test5.cpp:19:6:19:6 | y | provenance |  |
 | test5.cpp:18:6:18:18 | call to getTaintedInt | test5.cpp:18:2:18:20 | ... = ... | provenance |  |
-| test.c:10:27:10:30 | **argv | test.c:11:24:11:27 | call to atoi | provenance | TaintFunction |
+| test.c:10:28:10:31 | **argv | test.c:11:24:11:27 | call to atoi | provenance | TaintFunction |
-| test.c:10:27:10:30 | **argv | test.c:41:5:41:24 | ... = ... | provenance | TaintFunction |
+| test.c:10:28:10:31 | **argv | test.c:41:5:41:24 | ... = ... | provenance | TaintFunction |
-| test.c:10:27:10:30 | **argv | test.c:51:5:51:24 | ... = ... | provenance | TaintFunction |
+| test.c:10:28:10:31 | **argv | test.c:51:5:51:24 | ... = ... | provenance | TaintFunction |
 | test.c:11:24:11:27 | call to atoi | test.c:14:15:14:28 | maxConnections | provenance |  |
 | test.c:41:5:41:24 | ... = ... | test.c:44:7:44:10 | len2 | provenance |  |
 | test.c:41:5:41:24 | ... = ... | test.c:44:7:44:12 | ... -- | provenance |  |
@@ -26,6 +25,8 @@ edges
 | test.c:51:5:51:24 | ... = ... | test.c:54:7:54:12 | ... -- | provenance |  |
 | test.c:54:7:54:12 | ... -- | test.c:54:7:54:10 | len3 | provenance |  |
 nodes
+| main.cpp:7:27:7:30 | **argv | semmle.label | **argv |
+| main.cpp:8:17:8:20 | **argv | semmle.label | **argv |
 | test2.cpp:12:21:12:21 | v | semmle.label | v |
 | test2.cpp:14:11:14:11 | v | semmle.label | v |
 | test2.cpp:25:22:25:23 | fscanf output argument | semmle.label | fscanf output argument |
@@ -35,7 +36,6 @@ nodes
 | test2.cpp:39:3:39:18 | ... = ... | semmle.label | ... = ... |
 | test2.cpp:39:9:39:11 | num | semmle.label | num |
 | test2.cpp:40:3:40:5 | num | semmle.label | num |
-| test3.c:10:27:10:30 | **argv | semmle.label | **argv |
 | test5.cpp:5:5:5:17 | *getTaintedInt | semmle.label | *getTaintedInt |
 | test5.cpp:9:7:9:9 | gets output argument | semmle.label | gets output argument |
 | test5.cpp:10:9:10:27 | call to strtoul | semmle.label | call to strtoul |
@@ -43,7 +43,7 @@ nodes
 | test5.cpp:18:2:18:20 | ... = ... | semmle.label | ... = ... |
 | test5.cpp:18:6:18:18 | call to getTaintedInt | semmle.label | call to getTaintedInt |
 | test5.cpp:19:6:19:6 | y | semmle.label | y |
-| test.c:10:27:10:30 | **argv | semmle.label | **argv |
+| test.c:10:28:10:31 | **argv | semmle.label | **argv |
 | test.c:11:24:11:27 | call to atoi | semmle.label | call to atoi |
 | test.c:14:15:14:28 | maxConnections | semmle.label | maxConnections |
 | test.c:41:5:41:24 | ... = ... | semmle.label | ... = ... |
@@ -61,19 +61,7 @@ subpaths
 | test5.cpp:17:6:17:18 | call to getTaintedInt | test5.cpp:9:7:9:9 | gets output argument | test5.cpp:17:6:17:18 | call to getTaintedInt | $@ flows to an operand of an arithmetic expression, potentially causing an overflow. | test5.cpp:9:7:9:9 | gets output argument | string read by gets |
 | test5.cpp:19:6:19:6 | y | test5.cpp:9:7:9:9 | gets output argument | test5.cpp:19:6:19:6 | y | $@ flows to an operand of an arithmetic expression, potentially causing an overflow. | test5.cpp:9:7:9:9 | gets output argument | string read by gets |
 | test5.cpp:19:6:19:6 | y | test5.cpp:9:7:9:9 | gets output argument | test5.cpp:19:6:19:6 | y | $@ flows to an operand of an arithmetic expression, potentially causing an underflow. | test5.cpp:9:7:9:9 | gets output argument | string read by gets |
-| test.c:14:15:14:28 | maxConnections | test3.c:10:27:10:30 | **argv | test.c:14:15:14:28 | maxConnections | $@ flows to an operand of an arithmetic expression, potentially causing an overflow. | test3.c:10:27:10:30 | **argv | a command-line argument |
+| test.c:14:15:14:28 | maxConnections | main.cpp:7:27:7:30 | **argv | test.c:14:15:14:28 | maxConnections | $@ flows to an operand of an arithmetic expression, potentially causing an overflow. | main.cpp:7:27:7:30 | **argv | a command-line argument |
-| test.c:14:15:14:28 | maxConnections | test3.c:10:27:10:30 | **argv | test.c:14:15:14:28 | maxConnections | $@ flows to an operand of an arithmetic expression, potentially causing an overflow. | test.c:10:27:10:30 | **argv | a command-line argument |
+| test.c:14:15:14:28 | maxConnections | main.cpp:7:27:7:30 | **argv | test.c:14:15:14:28 | maxConnections | $@ flows to an operand of an arithmetic expression, potentially causing an underflow. | main.cpp:7:27:7:30 | **argv | a command-line argument |
-| test.c:14:15:14:28 | maxConnections | test3.c:10:27:10:30 | **argv | test.c:14:15:14:28 | maxConnections | $@ flows to an operand of an arithmetic expression, potentially causing an underflow. | test3.c:10:27:10:30 | **argv | a command-line argument |
+| test.c:44:7:44:10 | len2 | main.cpp:7:27:7:30 | **argv | test.c:44:7:44:10 | len2 | $@ flows to an operand of an arithmetic expression, potentially causing an underflow. | main.cpp:7:27:7:30 | **argv | a command-line argument |
-| test.c:14:15:14:28 | maxConnections | test3.c:10:27:10:30 | **argv | test.c:14:15:14:28 | maxConnections | $@ flows to an operand of an arithmetic expression, potentially causing an underflow. | test.c:10:27:10:30 | **argv | a command-line argument |
+| test.c:54:7:54:10 | len3 | main.cpp:7:27:7:30 | **argv | test.c:54:7:54:10 | len3 | $@ flows to an operand of an arithmetic expression, potentially causing an underflow. | main.cpp:7:27:7:30 | **argv | a command-line argument |
-| test.c:14:15:14:28 | maxConnections | test.c:10:27:10:30 | **argv | test.c:14:15:14:28 | maxConnections | $@ flows to an operand of an arithmetic expression, potentially causing an overflow. | test3.c:10:27:10:30 | **argv | a command-line argument |
-| test.c:14:15:14:28 | maxConnections | test.c:10:27:10:30 | **argv | test.c:14:15:14:28 | maxConnections | $@ flows to an operand of an arithmetic expression, potentially causing an overflow. | test.c:10:27:10:30 | **argv | a command-line argument |
-| test.c:14:15:14:28 | maxConnections | test.c:10:27:10:30 | **argv | test.c:14:15:14:28 | maxConnections | $@ flows to an operand of an arithmetic expression, potentially causing an underflow. | test3.c:10:27:10:30 | **argv | a command-line argument |
-| test.c:14:15:14:28 | maxConnections | test.c:10:27:10:30 | **argv | test.c:14:15:14:28 | maxConnections | $@ flows to an operand of an arithmetic expression, potentially causing an underflow. | test.c:10:27:10:30 | **argv | a command-line argument |
-| test.c:44:7:44:10 | len2 | test3.c:10:27:10:30 | **argv | test.c:44:7:44:10 | len2 | $@ flows to an operand of an arithmetic expression, potentially causing an underflow. | test3.c:10:27:10:30 | **argv | a command-line argument |
-| test.c:44:7:44:10 | len2 | test3.c:10:27:10:30 | **argv | test.c:44:7:44:10 | len2 | $@ flows to an operand of an arithmetic expression, potentially causing an underflow. | test.c:10:27:10:30 | **argv | a command-line argument |
-| test.c:44:7:44:10 | len2 | test.c:10:27:10:30 | **argv | test.c:44:7:44:10 | len2 | $@ flows to an operand of an arithmetic expression, potentially causing an underflow. | test3.c:10:27:10:30 | **argv | a command-line argument |
-| test.c:44:7:44:10 | len2 | test.c:10:27:10:30 | **argv | test.c:44:7:44:10 | len2 | $@ flows to an operand of an arithmetic expression, potentially causing an underflow. | test.c:10:27:10:30 | **argv | a command-line argument |
-| test.c:54:7:54:10 | len3 | test3.c:10:27:10:30 | **argv | test.c:54:7:54:10 | len3 | $@ flows to an operand of an arithmetic expression, potentially causing an underflow. | test3.c:10:27:10:30 | **argv | a command-line argument |
-| test.c:54:7:54:10 | len3 | test3.c:10:27:10:30 | **argv | test.c:54:7:54:10 | len3 | $@ flows to an operand of an arithmetic expression, potentially causing an underflow. | test.c:10:27:10:30 | **argv | a command-line argument |
-| test.c:54:7:54:10 | len3 | test.c:10:27:10:30 | **argv | test.c:54:7:54:10 | len3 | $@ flows to an operand of an arithmetic expression, potentially causing an underflow. | test3.c:10:27:10:30 | **argv | a command-line argument |
-| test.c:54:7:54:10 | len3 | test.c:10:27:10:30 | **argv | test.c:54:7:54:10 | len3 | $@ flows to an operand of an arithmetic expression, potentially causing an underflow. | test.c:10:27:10:30 | **argv | a command-line argument |

cpp/ql/test/query-tests/Security/CWE/CWE-190/semmle/tainted/IntegerOverflowTainted.expected

@@ -4,20 +4,15 @@
 | test2.cpp:17:11:17:22 | ... * ... | $@ flows an expression which might overflow. | test2.cpp:25:22:25:23 | fscanf output argument | value read by fscanf |
 | test2.cpp:39:9:39:18 | ... + ... | $@ flows an expression which might overflow. | test2.cpp:36:9:36:14 | fgets output argument | string read by fgets |
 | test2.cpp:40:3:40:13 | ... += ... | $@ flows an expression which might overflow. | test2.cpp:36:9:36:14 | fgets output argument | string read by fgets |
-| test3.c:12:11:12:34 | * ... | $@ flows an expression which might overflow negatively. | test3.c:10:27:10:30 | **argv | a command-line argument |
+| test3.c:12:11:12:34 | * ... | $@ flows an expression which might overflow negatively. | main.cpp:7:27:7:30 | **argv | a command-line argument |
-| test3.c:12:11:12:34 | * ... | $@ flows an expression which might overflow negatively. | test.c:10:27:10:30 | **argv | a command-line argument |
+| test3.c:13:11:13:20 | * ... | $@ flows an expression which might overflow negatively. | main.cpp:7:27:7:30 | **argv | a command-line argument |
-| test3.c:13:11:13:20 | * ... | $@ flows an expression which might overflow negatively. | test3.c:10:27:10:30 | **argv | a command-line argument |
+| test4.cpp:13:7:13:20 | access to array | $@ flows an expression which might overflow negatively. | main.cpp:7:27:7:30 | **argv | a command-line argument |
-| test3.c:13:11:13:20 | * ... | $@ flows an expression which might overflow negatively. | test.c:10:27:10:30 | **argv | a command-line argument |
-| test4.cpp:13:7:13:20 | access to array | $@ flows an expression which might overflow negatively. | test4.cpp:8:27:8:30 | **argv | a command-line argument |
 | test5.cpp:10:9:10:27 | call to strtoul | $@ flows an expression which might overflow. | test5.cpp:9:7:9:9 | gets output argument | string read by gets |
 | test5.cpp:17:6:17:27 | ... * ... | $@ flows an expression which might overflow. | test5.cpp:9:7:9:9 | gets output argument | string read by gets |
 | test5.cpp:19:6:19:13 | ... * ... | $@ flows an expression which might overflow. | test5.cpp:9:7:9:9 | gets output argument | string read by gets |
 | test6.cpp:11:10:11:15 | s | $@ flows an expression which might overflow. | test6.cpp:39:23:39:24 | fscanf output argument | value read by fscanf |
 | test6.cpp:16:10:16:15 | s | $@ flows an expression which might overflow. | test6.cpp:39:23:39:24 | fscanf output argument | value read by fscanf |
 | test6.cpp:30:11:30:16 | s | $@ flows an expression which might overflow. | test6.cpp:39:23:39:24 | fscanf output argument | value read by fscanf |
-| test.c:14:15:14:35 | ... * ... | $@ flows an expression which might overflow. | test3.c:10:27:10:30 | **argv | a command-line argument |
+| test.c:14:15:14:35 | ... * ... | $@ flows an expression which might overflow. | main.cpp:7:27:7:30 | **argv | a command-line argument |
-| test.c:14:15:14:35 | ... * ... | $@ flows an expression which might overflow. | test.c:10:27:10:30 | **argv | a command-line argument |
+| test.c:44:7:44:12 | ... -- | $@ flows an expression which might overflow negatively. | main.cpp:7:27:7:30 | **argv | a command-line argument |
-| test.c:44:7:44:12 | ... -- | $@ flows an expression which might overflow negatively. | test3.c:10:27:10:30 | **argv | a command-line argument |
+| test.c:54:7:54:12 | ... -- | $@ flows an expression which might overflow negatively. | main.cpp:7:27:7:30 | **argv | a command-line argument |
-| test.c:44:7:44:12 | ... -- | $@ flows an expression which might overflow negatively. | test.c:10:27:10:30 | **argv | a command-line argument |
-| test.c:54:7:54:12 | ... -- | $@ flows an expression which might overflow negatively. | test3.c:10:27:10:30 | **argv | a command-line argument |
-| test.c:54:7:54:12 | ... -- | $@ flows an expression which might overflow negatively. | test.c:10:27:10:30 | **argv | a command-line argument |

cpp/ql/test/query-tests/Security/CWE/CWE-190/semmle/tainted/main.cpp

@@ -0,0 +1,12 @@
+extern "C" {
+int main1(int argc, char** argv);
+int main3(int argc, char** argv);
+}
+int main4(int argc, char** argv);
+
+int main(int argc, char** argv) {
+    main1(argc, argv);
+    main3(argc, argv);
+    main4(argc, argv);
+    return 0;
+}

cpp/ql/test/query-tests/Security/CWE/CWE-190/semmle/tainted/test.c

@@ -7,7 +7,7 @@ void startServer(int heapSize);
 typedef unsigned long size_t;
 size_t strlen(const char *s);
 
-int main(int argc, char** argv) {
+int main1(int argc, char** argv) {
   int maxConnections = atoi(argv[1]);
 
   // BAD: arithmetic on a user input without any validation

cpp/ql/test/query-tests/Security/CWE/CWE-190/semmle/tainted/test3.c

@@ -7,7 +7,7 @@
 // Regression test for ODASA-6054: IntegerOverflowTainted should
 // not report a result if the overflow happens in a macro expansion
 // from a macro that is defined in a system header.
-int main(int argc, char **argv) {
+int main3(int argc, char **argv) {
   char *cmd = argv[0];
   int x = (int)(unsigned char)*cmd;  // BAD: overflow
   int y = CAST(*cmd);  // BAD: overflow in macro expansion (macro is not from a system header)

cpp/ql/test/query-tests/Security/CWE/CWE-190/semmle/tainted/test4.cpp

@@ -5,7 +5,7 @@
 // by comparing the value to 0. This means that the cast cannot overflow,
 // regardless of what the input type is.
 
-int main(int argc, char **argv) {
+int main4(int argc, char **argv) {
   char *p = argv[0];
   if (!p[0]) {  // GOOD: cast to bool.
     return 1;

cpp/ql/test/query-tests/Security/CWE/CWE-457/semmle/tests/UninitializedLocal.expected

@@ -1,5 +1,6 @@
 edges
 nodes
+| errors.cpp:13:7:13:7 | definition of x | semmle.label | definition of x |
 | test.cpp:11:6:11:8 | definition of foo | semmle.label | definition of foo |
 | test.cpp:111:6:111:8 | definition of foo | semmle.label | definition of foo |
 | test.cpp:226:7:226:7 | definition of x | semmle.label | definition of x |
@@ -14,6 +15,7 @@ nodes
 | test.cpp:472:6:472:6 | definition of x | semmle.label | definition of x |
 | test.cpp:479:6:479:6 | definition of x | semmle.label | definition of x |
 #select
+| errors.cpp:14:18:14:18 | x | errors.cpp:13:7:13:7 | definition of x | errors.cpp:13:7:13:7 | definition of x | The variable $@ may not be initialized at this access. | errors.cpp:13:7:13:7 | x | x |
 | test.cpp:12:6:12:8 | foo | test.cpp:11:6:11:8 | definition of foo | test.cpp:11:6:11:8 | definition of foo | The variable $@ may not be initialized at this access. | test.cpp:11:6:11:8 | foo | foo |
 | test.cpp:113:6:113:8 | foo | test.cpp:111:6:111:8 | definition of foo | test.cpp:111:6:111:8 | definition of foo | The variable $@ may not be initialized at this access. | test.cpp:111:6:111:8 | foo | foo |
 | test.cpp:227:3:227:3 | x | test.cpp:226:7:226:7 | definition of x | test.cpp:226:7:226:7 | definition of x | The variable $@ may not be initialized at this access. | test.cpp:226:7:226:7 | x | x |

cpp/ql/test/query-tests/Security/CWE/CWE-457/semmle/tests/errors.cpp

@@ -0,0 +1,15 @@
+// semmle-extractor-options: --expect_errors
+
+int f1() {
+  int x;
+  initialize(&x); // error expression - initialize() is not defined
+  return x;       // GOOD - assume x is initialized
+}
+
+void * operator new(unsigned long, bool);
+void operator delete(void*, bool);
+
+int f2() {
+  int x;
+  new(true) int (x);  // BAD, ignore implicit error expression
+}

cpp/ql/test/query-tests/Security/CWE/CWE-497/semmle/tests/tests2.cpp

@@ -50,7 +50,7 @@ int val();
 const char *global1 = mysql_get_client_info();
 const char *global2 = "abc";
 
-void test1()
+void test7()
 {
 	int sock = socket(val(), val(), val());
 

cpp/ql/test/query-tests/Security/CWE/CWE-704/WcharCharConversion.cpp

@@ -53,4 +53,59 @@ void NonStringFalsePositiveTest2(unsigned char* buffer)
 {
 	wchar_t *lpWchar = NULL;
 	lpWchar = (LPWSTR)buffer; // Possible False Positive
-}
+}
+
+typedef unsigned char BYTE;
+using FOO = BYTE*;
+
+void NonStringFalsePositiveTest3(FOO buffer)
+{
+	wchar_t *lpWchar = NULL;
+	lpWchar = (LPWSTR)buffer; // GOOD
+}
+
+#define UNICODE 0x8
+
+// assume EMPTY_MACRO is tied to if UNICODE is enabled
+#ifdef EMPTY_MACRO
+typedef WCHAR* LPTSTR;
+#else
+typedef char* LPTSTR;
+#endif
+
+void CheckedConversionFalsePositiveTest3(unsigned short flags, LPTSTR buffer)
+{
+	wchar_t *lpWchar = NULL;
+	if(flags & UNICODE)
+		lpWchar = (LPWSTR)buffer; // GOOD
+	else
+		lpWchar = (LPWSTR)buffer; // BUG
+
+	if((flags & UNICODE) == 0x8)
+		lpWchar = (LPWSTR)buffer; // GOOD
+	else
+		lpWchar = (LPWSTR)buffer; // BUG
+
+	if((flags & UNICODE) != 0x8)
+		lpWchar = (LPWSTR)buffer; // BUG
+	else
+		lpWchar = (LPWSTR)buffer; // GOOD
+
+	// Bad operator precedence
+	if(flags & UNICODE == 0x8)
+		lpWchar = (LPWSTR)buffer; // BUG
+	else
+		lpWchar = (LPWSTR)buffer; // BUG
+
+	if((flags & UNICODE) != 0)
+		lpWchar = (LPWSTR)buffer; // GOOD
+	else
+		lpWchar = (LPWSTR)buffer; // BUG
+
+	if((flags & UNICODE) == 0)
+		lpWchar = (LPWSTR)buffer; // BUG
+	else
+		lpWchar = (LPWSTR)buffer; // GOOD
+
+	lpWchar = (LPWSTR)buffer; // BUG
+}

cpp/ql/test/query-tests/Security/CWE/CWE-704/WcharCharConversion.expected

@@ -3,3 +3,11 @@
 | WcharCharConversion.cpp:24:22:24:27 | lpChar | Conversion from char * to wchar_t *. Use of invalid string can lead to undefined behavior. |
 | WcharCharConversion.cpp:26:23:26:28 | lpChar | Conversion from char * to LPCWSTR. Use of invalid string can lead to undefined behavior. |
 | WcharCharConversion.cpp:27:17:27:22 | lpChar | Conversion from char * to LPWSTR. Use of invalid string can lead to undefined behavior. |
+| WcharCharConversion.cpp:82:21:82:26 | buffer | Conversion from LPTSTR to LPWSTR. Use of invalid string can lead to undefined behavior. |
+| WcharCharConversion.cpp:87:21:87:26 | buffer | Conversion from LPTSTR to LPWSTR. Use of invalid string can lead to undefined behavior. |
+| WcharCharConversion.cpp:90:21:90:26 | buffer | Conversion from LPTSTR to LPWSTR. Use of invalid string can lead to undefined behavior. |
+| WcharCharConversion.cpp:96:21:96:26 | buffer | Conversion from LPTSTR to LPWSTR. Use of invalid string can lead to undefined behavior. |
+| WcharCharConversion.cpp:98:21:98:26 | buffer | Conversion from LPTSTR to LPWSTR. Use of invalid string can lead to undefined behavior. |
+| WcharCharConversion.cpp:103:21:103:26 | buffer | Conversion from LPTSTR to LPWSTR. Use of invalid string can lead to undefined behavior. |
+| WcharCharConversion.cpp:106:21:106:26 | buffer | Conversion from LPTSTR to LPWSTR. Use of invalid string can lead to undefined behavior. |
+| WcharCharConversion.cpp:110:20:110:25 | buffer | Conversion from LPTSTR to LPWSTR. Use of invalid string can lead to undefined behavior. |

csharp/ql/campaigns/Solorigate/lib/CHANGELOG.md

@@ -1,3 +1,7 @@
+## 1.7.27
+
+No user-facing changes.
+
 ## 1.7.26
 
 No user-facing changes.

csharp/ql/campaigns/Solorigate/lib/change-notes/released/1.7.27.md

@@ -0,0 +1,3 @@
+## 1.7.27
+
+No user-facing changes.

csharp/ql/campaigns/Solorigate/lib/codeql-pack.release.yml

@@ -1,2 +1,2 @@
 ---
-lastReleaseVersion: 1.7.26
+lastReleaseVersion: 1.7.27

csharp/ql/campaigns/Solorigate/lib/qlpack.yml

@@ -1,5 +1,5 @@
 name: codeql/csharp-solorigate-all
-version: 1.7.26
+version: 1.7.28-dev
 groups:
   - csharp
   - solorigate

csharp/ql/campaigns/Solorigate/src/CHANGELOG.md

@@ -1,3 +1,7 @@
+## 1.7.27
+
+No user-facing changes.
+
 ## 1.7.26
 
 No user-facing changes.

csharp/ql/campaigns/Solorigate/src/change-notes/released/1.7.27.md

@@ -0,0 +1,3 @@
+## 1.7.27
+
+No user-facing changes.

csharp/ql/campaigns/Solorigate/src/codeql-pack.release.yml

@@ -1,2 +1,2 @@
 ---
-lastReleaseVersion: 1.7.26
+lastReleaseVersion: 1.7.27

csharp/ql/campaigns/Solorigate/src/qlpack.yml

@@ -1,5 +1,5 @@
 name: codeql/csharp-solorigate-queries
-version: 1.7.26
+version: 1.7.28-dev
 groups:
   - csharp
   - solorigate

csharp/ql/lib/CHANGELOG.md

@@ -1,3 +1,7 @@
+## 3.0.1
+
+No user-facing changes.
+
 ## 3.0.0
 
 ### Breaking Changes

csharp/ql/lib/change-notes/released/3.0.1.md

@@ -0,0 +1,3 @@
+## 3.0.1
+
+No user-facing changes.

csharp/ql/lib/codeql-pack.release.yml

@@ -1,2 +1,2 @@
 ---
-lastReleaseVersion: 3.0.0
+lastReleaseVersion: 3.0.1

csharp/ql/lib/qlpack.yml

@@ -1,5 +1,5 @@
 name: codeql/csharp-all
-version: 3.0.0
+version: 3.0.2-dev
 groups: csharp
 dbscheme: semmlecode.csharp.dbscheme
 extractor: csharp

csharp/ql/lib/semmle/code/csharp/controlflow/internal/ControlFlowGraphImpl.qll

@@ -53,7 +53,6 @@ private predicate idOf(AstNode x, int y) = equivalenceRelation(id/2)(x, y)
 private module CfgInput implements CfgShared::InputSig<Location> {
   private import ControlFlowGraphImpl as Impl
   private import Completion as Comp
-  private import Splitting as Splitting
   private import SuccessorType as ST
   private import semmle.code.csharp.Caching
 
@@ -80,10 +79,6 @@ private module CfgInput implements CfgShared::InputSig<Location> {
     Impl::scopeLast(scope, last, c)
   }
 
-  class SplitKindBase = Splitting::TSplitKind;
-
-  class Split = Splitting::Split;
-
   class SuccessorType = ST::SuccessorType;
 
   SuccessorType getAMatchingSuccessorType(Completion c) { result = c.getAMatchingSuccessorType() }
@@ -102,7 +97,21 @@ private module CfgInput implements CfgShared::InputSig<Location> {
   }
 }
 
-import CfgShared::Make<Location, CfgInput>
+private module CfgSplittingInput implements CfgShared::SplittingInputSig<Location, CfgInput> {
+  private import Splitting as S
+
+  class SplitKindBase = S::TSplitKind;
+
+  class Split = S::Split;
+}
+
+private module ConditionalCompletionSplittingInput implements
+  CfgShared::ConditionalCompletionSplittingInputSig<Location, CfgInput, CfgSplittingInput>
+{
+  import Splitting::ConditionalCompletionSplitting::ConditionalCompletionSplittingInput
+}
+
+import CfgShared::MakeWithSplitting<Location, CfgInput, CfgSplittingInput, ConditionalCompletionSplittingInput>
 
 /**
  * A compilation.

csharp/ql/lib/semmle/code/csharp/controlflow/internal/Splitting.qll

@@ -5,7 +5,8 @@
  */
 
 import csharp
-private import Completion
+private import Completion as Comp
+private import Comp
 private import ControlFlowGraphImpl
 private import semmle.code.csharp.controlflow.ControlFlowGraph::ControlFlow as Cfg
 private import semmle.code.csharp.controlflow.internal.PreSsa
@@ -260,10 +261,12 @@ module ConditionalCompletionSplitting {
 
     ConditionalCompletionSplit() { this = TConditionalCompletionSplit(completion) }
 
+    ConditionalCompletion getCompletion() { result = completion }
+
     override string toString() { result = completion.toString() }
   }
 
-  private class ConditionalCompletionSplitKind extends SplitKind, TConditionalCompletionSplitKind {
+  private class ConditionalCompletionSplitKind_ extends SplitKind, TConditionalCompletionSplitKind {
     override int getListOrder() { result = InitializerSplitting::getNextListOrder() }
 
     override predicate isEnabled(AstNode cfe) { this.appliesTo(cfe) }
@@ -271,89 +274,64 @@ module ConditionalCompletionSplitting {
     override string toString() { result = "ConditionalCompletion" }
   }
 
-  int getNextListOrder() { result = InitializerSplitting::getNextListOrder() + 1 }
+  module ConditionalCompletionSplittingInput {
+    private import Completion as Comp
 
-  private class ConditionalCompletionSplitImpl extends SplitImpl instanceof ConditionalCompletionSplit
+    class ConditionalCompletion = Comp::ConditionalCompletion;
-  {
-    ConditionalCompletion completion;
 
-    ConditionalCompletionSplitImpl() { this = TConditionalCompletionSplit(completion) }
+    class ConditionalCompletionSplitKind extends ConditionalCompletionSplitKind_, TSplitKind { }
 
-    override ConditionalCompletionSplitKind getKind() { any() }
+    class ConditionalCompletionSplit = ConditionalCompletionSplitting::ConditionalCompletionSplit;
 
-    override predicate hasEntry(AstNode pred, AstNode succ, Completion c) {
+    bindingset[parent, parentCompletion]
-      succ(pred, succ, c) and
+    predicate condPropagateExpr(
-      last(succ, _, completion) and
+      AstNode parent, ConditionalCompletion parentCompletion, AstNode child,
+      ConditionalCompletion childCompletion
+    ) {
+      child = parent.(LogicalNotExpr).getOperand() and
+      childCompletion.getDual() = parentCompletion
+      or
+      childCompletion = parentCompletion and
       (
-        last(succ.(LogicalNotExpr).getOperand(), pred, c) and
+        child = parent.(LogicalAndExpr).getAnOperand()
-        completion.(BooleanCompletion).getDual() = c
         or
-        last(succ.(LogicalAndExpr).getAnOperand(), pred, c) and
+        child = parent.(LogicalOrExpr).getAnOperand()
-        completion = c
         or
-        last(succ.(LogicalOrExpr).getAnOperand(), pred, c) and
+        parent = any(ConditionalExpr ce | child = [ce.getThen(), ce.getElse()])
-        completion = c
         or
-        succ =
+        child = parent.(SwitchExpr).getACase()
-          any(ConditionalExpr ce |
-            last([ce.getThen(), ce.getElse()], pred, c) and
-            completion = c
-          )
         or
-        succ =
+        child = parent.(SwitchCaseExpr).getBody()
+        or
+        parent =
           any(NullCoalescingExpr nce |
-            exists(Expr operand |
+            if childCompletion instanceof NullnessCompletion
-              last(operand, pred, c) and
+            then child = nce.getRightOperand()
-              completion = c
+            else child = nce.getAnOperand()
-            |
-              if c instanceof NullnessCompletion
-              then operand = nce.getRightOperand()
-              else operand = nce.getAnOperand()
-            )
           )
+      )
+      or
+      child = parent.(NotPatternExpr).getPattern() and
+      childCompletion.getDual() = parentCompletion
+      or
+      child = parent.(IsExpr).getPattern() and
+      parentCompletion.(BooleanCompletion).getValue() =
+        childCompletion.(MatchingCompletion).getValue()
+      or
+      childCompletion = parentCompletion and
+      (
+        child = parent.(AndPatternExpr).getAnOperand()
         or
-        last(succ.(SwitchExpr).getACase(), pred, c) and
+        child = parent.(OrPatternExpr).getAnOperand()
-        completion = c
         or
-        last(succ.(SwitchCaseExpr).getBody(), pred, c) and
+        child = parent.(RecursivePatternExpr).getAChildExpr()
-        completion = c
         or
-        last(succ.(NotPatternExpr).getPattern(), pred, c) and
+        child = parent.(PropertyPatternExpr).getPattern(_)
-        completion.(MatchingCompletion).getDual() = c
-        or
-        last(succ.(IsExpr).getPattern(), pred, c) and
-        completion.(BooleanCompletion).getValue() = c.(MatchingCompletion).getValue()
-        or
-        last(succ.(AndPatternExpr).getAnOperand(), pred, c) and
-        completion = c
-        or
-        last(succ.(OrPatternExpr).getAnOperand(), pred, c) and
-        completion = c
-        or
-        last(succ.(RecursivePatternExpr).getAChildExpr(), pred, c) and
-        completion = c
-        or
-        last(succ.(PropertyPatternExpr).getPattern(_), pred, c) and
-        completion = c
       )
     }
-
-    override predicate hasEntryScope(CfgScope scope, AstNode first) { none() }
-
-    override predicate hasExit(AstNode pred, AstNode succ, Completion c) {
-      this.appliesTo(pred) and
-      succ(pred, succ, c) and
-      if c instanceof ConditionalCompletion then completion = c else any()
-    }
-
-    override predicate hasExitScope(CfgScope scope, AstNode last, Completion c) {
-      this.appliesTo(last) and
-      scopeLast(scope, last, c) and
-      if c instanceof ConditionalCompletion then completion = c else any()
-    }
-
-    override predicate hasSuccessor(AstNode pred, AstNode succ, Completion c) { none() }
   }
+
+  int getNextListOrder() { result = InitializerSplitting::getNextListOrder() + 1 }
 }
 
 module AssertionSplitting {