Add diagnostics rule for arbitrary build messages

Python: Add test of keyword argument with same name as positional-only parameter

.github/actions/cache-query-compilation/action.yml

@@ -9,7 +9,7 @@ inputs:
 outputs:
   cache-dir:
     description: "The directory where the cache was stored"
-    value: ${{ steps.fill-compilation-dir.outputs.compdir }}
+    value: ${{ steps.output-compilation-dir.outputs.compdir }}
 
 runs:
   using: composite
@@ -27,7 +27,9 @@ runs:
       if: ${{ github.event_name == 'pull_request' }}
       uses: actions/cache/restore@v3
       with:
-        path: '**/.cache'
+        path: |
+          **/.cache
+          ~/.codeql/compile-cache
         key: codeql-compile-${{ inputs.key }}-pr-${{ github.sha }}
         restore-keys: |
           codeql-compile-${{ inputs.key }}-${{ github.base_ref }}-${{ env.merge_base }}
@@ -37,18 +39,111 @@ runs:
       if: ${{ github.event_name != 'pull_request' }}
       uses: actions/cache@v3
       with:
-        path: '**/.cache'
+        path: |
+          **/.cache
+          ~/.codeql/compile-cache
         key: codeql-compile-${{ inputs.key }}-${{ github.ref_name }}-${{ github.sha }} # just fill on main
         restore-keys: | # restore the latest cache if the exact cache is unavailable, to speed up compilation.
           codeql-compile-${{ inputs.key }}-${{ github.ref_name }}-
           codeql-compile-${{ inputs.key }}-main-
-    - name: Fill compilation cache directory
-      id: fill-compilation-dir
+    - name: Output-compilationdir
+      id: output-compilation-dir
       shell: bash
       run: |
-        # Move all the existing cache into another folder, so we only preserve the cache for the current queries.
-        node $GITHUB_WORKSPACE/.github/actions/cache-query-compilation/move-caches.js ${COMBINED_CACHE_DIR}
-
         echo "compdir=${COMBINED_CACHE_DIR}" >> $GITHUB_OUTPUT
       env:
         COMBINED_CACHE_DIR: ${{ runner.temp }}/compilation-dir
+    - name: Fill compilation cache directory
+      id: fill-compilation-dir
+      uses: actions/github-script@v6
+      env: 
+        COMBINED_CACHE_DIR: ${{ runner.temp }}/compilation-dir
+      with:
+        script: |
+          // # Move all the existing cache into another folder, so we only preserve the cache for the current queries.
+          // mkdir -p ${COMBINED_CACHE_DIR}
+          // rm -f **/.cache/{lock,size} # -f to avoid errors if the cache is empty.
+          // # copy the contents of the .cache folders into the combined cache folder.
+          // cp -r **/.cache/* ${COMBINED_CACHE_DIR}/ || : # ignore missing files
+          // # clean up the .cache folders
+          // rm -rf **/.cache/*
+
+          const fs = require("fs");
+          const path = require("path");
+          const os = require("os");
+
+          // the first argv is the cache folder to create.
+          const COMBINED_CACHE_DIR = process.env.COMBINED_CACHE_DIR;
+
+          function* walkCaches(dir) {
+            const files = fs.readdirSync(dir, { withFileTypes: true });
+            for (const file of files) {
+              if (file.isDirectory()) {
+                const filePath = path.join(dir, file.name);
+                yield* walkCaches(filePath);
+                if (file.name === ".cache") {
+                  yield filePath;
+                }
+              }
+            }
+          }
+
+          async function copyDir(src, dest) {
+            for await (const file of await fs.promises.readdir(src, { withFileTypes: true })) {
+              const srcPath = path.join(src, file.name);
+              const destPath = path.join(dest, file.name);
+              if (file.isDirectory()) {
+                if (!fs.existsSync(destPath)) {
+                  fs.mkdirSync(destPath);
+                }
+                await copyDir(srcPath, destPath);
+              } else {
+                await fs.promises.copyFile(srcPath, destPath);
+              }
+            }
+          }
+
+          async function main() {
+            const cacheDirs = [...walkCaches(".")];
+
+            for (const dir of cacheDirs) {
+              console.log(`Found .cache dir at ${dir}`);
+            }
+
+            const globalCacheDir = path.join(os.homedir(), ".codeql", "compile-cache");
+            if (fs.existsSync(globalCacheDir)) {
+              console.log("Found global home dir: " + globalCacheDir);
+              cacheDirs.push(globalCacheDir);
+            }
+
+            if (cacheDirs.length === 0) {
+              console.log("No cache dirs found");
+              return;
+            }
+
+            // mkdir -p ${COMBINED_CACHE_DIR}
+            fs.mkdirSync(COMBINED_CACHE_DIR, { recursive: true });
+
+            // rm -f **/.cache/{lock,size} # -f to avoid errors if the cache is empty.
+            await Promise.all(
+              cacheDirs.map((cacheDir) =>
+                (async function () {
+                  await fs.promises.rm(path.join(cacheDir, "lock"), { force: true });
+                  await fs.promises.rm(path.join(cacheDir, "size"), { force: true });
+                })()
+              )
+            );
+
+            // # copy the contents of the .cache folders into the combined cache folder.
+            // cp -r **/.cache/* ${COMBINED_CACHE_DIR}/ || : # ignore missing files
+            await Promise.all(
+              cacheDirs.map((cacheDir) => copyDir(cacheDir, COMBINED_CACHE_DIR))
+            );
+
+            // # clean up the .cache folders
+            // rm -rf **/.cache/*
+            await Promise.all(
+              cacheDirs.map((cacheDir) => fs.promises.rm(cacheDir, { recursive: true }))
+            );
+          }
+          main();

.github/actions/cache-query-compilation/move-caches.js

@@ -1,75 +0,0 @@
-// # Move all the existing cache into another folder, so we only preserve the cache for the current queries.
-// mkdir -p ${COMBINED_CACHE_DIR}
-// rm -f **/.cache/{lock,size} # -f to avoid errors if the cache is empty.
-// # copy the contents of the .cache folders into the combined cache folder.
-// cp -r **/.cache/* ${COMBINED_CACHE_DIR}/ || : # ignore missing files
-// # clean up the .cache folders
-// rm -rf **/.cache/*
-
-const fs = require("fs");
-const path = require("path");
-
-// the first argv is the cache folder to create.
-const COMBINED_CACHE_DIR = process.argv[2];
-
-function* walkCaches(dir) {
-  const files = fs.readdirSync(dir, { withFileTypes: true });
-  for (const file of files) {
-    if (file.isDirectory()) {
-      const filePath = path.join(dir, file.name);
-      yield* walkCaches(filePath);
-      if (file.name === ".cache") {
-        yield filePath;
-      }
-    }
-  }
-}
-
-async function copyDir(src, dest) {
-  for await (const file of await fs.promises.readdir(src, { withFileTypes: true })) {
-    const srcPath = path.join(src, file.name);
-    const destPath = path.join(dest, file.name);
-    if (file.isDirectory()) {
-      if (!fs.existsSync(destPath)) {
-        fs.mkdirSync(destPath);
-      }
-      await copyDir(srcPath, destPath);
-    } else {
-      await fs.promises.copyFile(srcPath, destPath);
-    }
-  }
-}
-
-async function main() {
-  const cacheDirs = [...walkCaches(".")];
-
-  for (const dir of cacheDirs) {
-    console.log(`Found .cache dir at ${dir}`);
-  }
-
-  // mkdir -p ${COMBINED_CACHE_DIR}
-  fs.mkdirSync(COMBINED_CACHE_DIR, { recursive: true });
-
-  // rm -f **/.cache/{lock,size} # -f to avoid errors if the cache is empty.
-  await Promise.all(
-    cacheDirs.map((cacheDir) =>
-      (async function () {
-        await fs.promises.rm(path.join(cacheDir, "lock"), { force: true });
-        await fs.promises.rm(path.join(cacheDir, "size"), { force: true });
-      })()
-    )
-  );
-
-  // # copy the contents of the .cache folders into the combined cache folder.
-  // cp -r **/.cache/* ${COMBINED_CACHE_DIR}/ || : # ignore missing files
-  await Promise.all(
-    cacheDirs.map((cacheDir) => copyDir(cacheDir, COMBINED_CACHE_DIR))
-  );
-
-  // # clean up the .cache folders
-  // rm -rf **/.cache/*
-  await Promise.all(
-    cacheDirs.map((cacheDir) => fs.promises.rm(cacheDir, { recursive: true }))
-  );
-}
-main();

.github/actions/find-latest-bundle/action.yml

@@ -1,26 +0,0 @@
-name: Find Latest CodeQL Bundle
-description: Finds the URL of the latest released version of the CodeQL bundle.
-outputs:
-  url:
-    description: The download URL of the latest CodeQL bundle release
-    value: ${{ steps.find-latest.outputs.url }}
-runs:
-  using: composite
-  steps:
-    - name: Find Latest Release
-      id: find-latest
-      shell: pwsh
-      run: |
-        $Latest = gh release list --repo github/codeql-action --exclude-drafts --limit 1000 |
-          ForEach-Object { $C = $_ -split "`t"; return @{ type = $C[1]; tag = $C[2]; } } |
-          Where-Object { $_.type -eq 'Latest' }
-
-        $Tag = $Latest.tag
-        if ($Tag -eq '') {
-          throw 'Failed to find latest bundle release.'
-        }
-
-        Write-Output "Latest bundle tag is '${Tag}'."
-        "url=https://github.com/github/codeql-action/releases/download/${Tag}/codeql-bundle-linux64.tar.gz" >> $env:GITHUB_OUTPUT
-      env:
-        GITHUB_TOKEN: ${{ github.token }}

.github/dependabot.yml

@@ -1,19 +1,12 @@
 version: 2
 updates:
   - package-ecosystem: "cargo"
-    directory: "ruby/node-types"
+    directory: "ruby"
     schedule:
       interval: "daily"
+
   - package-ecosystem: "cargo"
-    directory: "ruby/generator"
-    schedule:
-      interval: "daily"
-  - package-ecosystem: "cargo"
-    directory: "ruby/extractor"
-    schedule:
-      interval: "daily"
-  - package-ecosystem: "cargo"
-    directory: "ruby/autobuilder"
+    directory: "ql"
     schedule:
       interval: "daily"
 

.github/workflows/check-qldoc.yml

@@ -26,9 +26,8 @@ jobs:
         shell: bash
         run: |
           EXIT_CODE=0
-          # TODO: remove the swift exception from the regex when we fix generated QLdoc
           # TODO: remove the shared exception from the regex when coverage of qlpacks without dbschemes is supported
-          changed_lib_packs="$(git diff --name-only --diff-filter=ACMRT HEAD^ HEAD | { grep -Po '^(?!(swift|shared))[a-z]*/ql/lib' || true; } | sort -u)"
+          changed_lib_packs="$(git diff --name-only --diff-filter=ACMRT HEAD^ HEAD | { grep -Po '^(?!(shared))[a-z]*/ql/lib' || true; } | sort -u)"
           for pack_dir in ${changed_lib_packs}; do
             lang="${pack_dir%/ql/lib}"
             codeql generate library-doc-coverage --output="${RUNNER_TEMP}/${lang}-current.txt" --dir="${pack_dir}"

.github/workflows/codeql-analysis.yml

@@ -28,9 +28,9 @@ jobs:
 
     steps:
     - name: Setup dotnet
-      uses: actions/setup-dotnet@v2
+      uses: actions/setup-dotnet@v3
       with:
-        dotnet-version: 6.0.202
+        dotnet-version: 7.0.102
 
     - name: Checkout repository
       uses: actions/checkout@v3

.github/workflows/compile-queries.yml

@@ -24,14 +24,14 @@ jobs:
         with: 
           key: all-queries
       - name: check formatting
-        run: find */ql -type f \( -name "*.qll" -o -name "*.ql" \) -print0 | xargs -0 codeql query format --check-only
+        run: find */ql -type f \( -name "*.qll" -o -name "*.ql" \) -print0 | xargs -0 -n 3000 -P 10 codeql query format -q --check-only
       - name: compile queries - check-only
         # run with --check-only if running in a PR (github.sha != main)
         if : ${{ github.event_name == 'pull_request' }}
         shell: bash
-        run: codeql query compile -j0 */ql/{src,examples} --keep-going --warnings=error --check-only --compilation-cache "${{ steps.query-cache.outputs.cache-dir }}"
+        run: codeql query compile -q -j0 */ql/{src,examples} --keep-going --warnings=error --check-only --compilation-cache "${{ steps.query-cache.outputs.cache-dir }}"
       - name: compile queries - full
         # do full compile if running on main - this populates the cache
         if : ${{ github.event_name != 'pull_request' }}
         shell: bash
-        run: codeql query compile -j0 */ql/{src,examples} --keep-going --warnings=error --compilation-cache "${{ steps.query-cache.outputs.cache-dir }}"
+        run: codeql query compile -q -j0 */ql/{src,examples} --keep-going --warnings=error --compilation-cache "${{ steps.query-cache.outputs.cache-dir }}"

.github/workflows/csharp-qltest.yml

@@ -77,10 +77,10 @@ jobs:
       - name: Setup dotnet
         uses: actions/setup-dotnet@v3
         with:
-          dotnet-version: 6.0.202
+          dotnet-version: 7.0.102
       - name: Extractor unit tests
         run: |
-          dotnet test -p:RuntimeFrameworkVersion=6.0.4 "${{ github.workspace }}/csharp/extractor/Semmle.Util.Tests"
-          dotnet test -p:RuntimeFrameworkVersion=6.0.4 "${{ github.workspace }}/csharp/extractor/Semmle.Extraction.Tests"
-          dotnet test -p:RuntimeFrameworkVersion=6.0.4 "${{ github.workspace }}/csharp/autobuilder/Semmle.Autobuild.CSharp.Tests"
-          dotnet test -p:RuntimeFrameworkVersion=6.0.4 "${{ github.workspace }}/cpp/autobuilder/Semmle.Autobuild.Cpp.Tests"
+          dotnet test -p:RuntimeFrameworkVersion=7.0.2 "${{ github.workspace }}/csharp/extractor/Semmle.Util.Tests"
+          dotnet test -p:RuntimeFrameworkVersion=7.0.2 "${{ github.workspace }}/csharp/extractor/Semmle.Extraction.Tests"
+          dotnet test -p:RuntimeFrameworkVersion=7.0.2 "${{ github.workspace }}/csharp/autobuilder/Semmle.Autobuild.CSharp.Tests"
+          dotnet test -p:RuntimeFrameworkVersion=7.0.2 "${{ github.workspace }}/cpp/autobuilder/Semmle.Autobuild.Cpp.Tests"

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

@@ -12,10 +12,10 @@ jobs:
     name: Test MacOS
     runs-on: macos-latest
     steps:
-      - name: Set up Go 1.19
+      - name: Set up Go 1.20
         uses: actions/setup-go@v3
         with:
-          go-version: 1.19
+          go-version: 1.20.0
         id: go
 
       - name: Check out code
@@ -47,10 +47,10 @@ jobs:
     name: Test Windows
     runs-on: windows-latest-xl
     steps:
-      - name: Set up Go 1.19
+      - name: Set up Go 1.20
         uses: actions/setup-go@v3
         with:
-          go-version: 1.19
+          go-version: 1.20.0
         id: go
 
       - name: Check out code

.github/workflows/go-tests.yml

@@ -20,10 +20,10 @@ jobs:
     name: Test Linux (Ubuntu)
     runs-on: ubuntu-latest-xl
     steps:
-      - name: Set up Go 1.19
+      - name: Set up Go 1.20
         uses: actions/setup-go@v3
         with:
-          go-version: 1.19
+          go-version: 1.20.0
         id: go
 
       - name: Check out code

.github/workflows/ql-for-ql-build.yml

@@ -5,13 +5,6 @@ on:
     branches: [main]
   pull_request:
     branches: [main]
-    paths:
-      - "ql/**"
-      - "**.qll"
-      - "**.ql"
-      - "**.dbscheme"
-      - "**/qlpack.yml"
-      - ".github/workflows/ql-for-ql-build.yml"
 
 env:
   CARGO_TERM_COLOR: always
@@ -22,66 +15,26 @@ jobs:
     steps:
       ### Build the queries ###
       - uses: actions/checkout@v3
-      - name: Find latest bundle
-        id: find-latest-bundle
-        uses: ./.github/actions/find-latest-bundle
+        with:
+          fetch-depth: 0
       - name: Find codeql
         id: find-codeql
-        uses: github/codeql-action/init@45955cb1830b640e2c1603ad72ad542a49d47b96
+        uses: github/codeql-action/init@v2
         with:
           languages: javascript # does not matter
-          tools: ${{ steps.find-latest-bundle.outputs.url }}
-      - name: Get CodeQL version
-        id: get-codeql-version
-        run: |
-          echo "version=$("${CODEQL}" --version | head -n 1 | rev | cut -d " " -f 1 | rev)" >> $GITHUB_OUTPUT
-        shell: bash
-        env:
-          CODEQL: ${{ steps.find-codeql.outputs.codeql-path }}
       - uses: ./.github/actions/os-version
         id: os_version
-      - name: Cache entire pack
-        id: cache-pack
-        uses: actions/cache@v3
-        with:
-          path: ${{ runner.temp }}/pack
-          key: ${{ runner.os }}-${{ steps.os_version.outputs.version }}-pack-${{ hashFiles('ql/**/Cargo.lock') }}-${{ hashFiles('ql/**/*.rs') }}-${{ hashFiles('ql/**/*.ql*') }}-${{ hashFiles('ql/**/qlpack.yml') }}-${{ hashFiles('ql/ql/src/ql.dbscheme*') }}-${{ steps.get-codeql-version.outputs.version }}--${{ hashFiles('.github/workflows/ql-for-ql-build.yml') }}
-      - name: Cache queries
-        if: steps.cache-pack.outputs.cache-hit != 'true'
-        id: cache-queries
-        uses: actions/cache@v3
-        with:
-          path: ${{ runner.temp }}/queries
-          key: queries-${{ hashFiles('ql/**/*.ql*') }}-${{ hashFiles('ql/**/qlpack.yml') }}-${{ hashFiles('ql/ql/src/ql.dbscheme*') }}-${{ steps.get-codeql-version.outputs.version }}--${{ hashFiles('.github/workflows/ql-for-ql-build.yml') }}
-      - name: Build query pack
-        if: steps.cache-queries.outputs.cache-hit != 'true' && steps.cache-pack.outputs.cache-hit != 'true'
-        run: |
-          cd ql/ql/src
-          "${CODEQL}" pack create -j 16
-          mv .codeql/pack/codeql/ql/0.0.0 ${{ runner.temp }}/queries
-        env:
-          CODEQL: ${{ steps.find-codeql.outputs.codeql-path }}
-      - name: Move cache queries to pack
-        if: steps.cache-pack.outputs.cache-hit != 'true'
-        run: |
-          cp -r ${{ runner.temp }}/queries ${{ runner.temp }}/pack
-        env:
-          CODEQL: ${{ steps.find-codeql.outputs.codeql-path }}
-
       ### Build the extractor ###
       - name: Cache entire extractor
-        if: steps.cache-pack.outputs.cache-hit != 'true'
         id: cache-extractor
         uses: actions/cache@v3
         with:
           path: |
-            ql/target/release/ql-autobuilder
-            ql/target/release/ql-autobuilder.exe
-            ql/target/release/ql-extractor
-            ql/target/release/ql-extractor.exe
+            ql/extractor-pack/
+            ql/target/release/buramu
           key: ${{ runner.os }}-${{ steps.os_version.outputs.version }}-extractor-${{ hashFiles('ql/**/Cargo.lock') }}-${{ hashFiles('ql/**/*.rs') }}
       - name: Cache cargo
-        if: steps.cache-extractor.outputs.cache-hit != 'true' && steps.cache-pack.outputs.cache-hit != 'true'
+        if: steps.cache-extractor.outputs.cache-hit != 'true'
         uses: actions/cache@v3
         with:
           path: |
@@ -89,77 +42,33 @@ jobs:
             ~/.cargo/git
             ql/target
           key: ${{ runner.os }}-${{ steps.os_version.outputs.version }}-rust-cargo-${{ hashFiles('ql/**/Cargo.lock') }}
-      - name: Check formatting
-        if: steps.cache-extractor.outputs.cache-hit != 'true' && steps.cache-pack.outputs.cache-hit != 'true'
-        run: cd ql; cargo fmt --all -- --check
-      - name: Build
-        if: steps.cache-extractor.outputs.cache-hit != 'true' && steps.cache-pack.outputs.cache-hit != 'true'
-        run: cd ql; cargo build --verbose
-      - name: Run tests
-        if: steps.cache-extractor.outputs.cache-hit != 'true' && steps.cache-pack.outputs.cache-hit != 'true'
-        run: cd ql; cargo test --verbose
       - name: Release build
-        if: steps.cache-extractor.outputs.cache-hit != 'true' && steps.cache-pack.outputs.cache-hit != 'true'
-        run: cd ql; cargo build --release
-      - name: Generate dbscheme
-        if: steps.cache-extractor.outputs.cache-hit != 'true' && steps.cache-pack.outputs.cache-hit != 'true'
-        run: ql/target/release/ql-generator --dbscheme ql/ql/src/ql.dbscheme --library ql/ql/src/codeql_ql/ast/internal/TreeSitter.qll
-
-      ### Package the queries and extractor ###
-      - name: Package pack
-        if: steps.cache-pack.outputs.cache-hit != 'true'
-        run: |
-          cp -r ql/codeql-extractor.yml ql/tools ql/ql/src/ql.dbscheme.stats ${PACK}/
-          mkdir -p ${PACK}/tools/linux64
-          cp ql/target/release/ql-autobuilder  ${PACK}/tools/linux64/autobuilder
-          cp ql/target/release/ql-extractor ${PACK}/tools/linux64/extractor
-          chmod +x ${PACK}/tools/linux64/autobuilder
-          chmod +x ${PACK}/tools/linux64/extractor
+        if: steps.cache-extractor.outputs.cache-hit != 'true'
+        run: cd ql; ./scripts/create-extractor-pack.sh       
         env:
-          PACK: ${{ runner.temp }}/pack
-
-      ### Run the analysis ###
-      - name: Hack codeql-action options
+          GH_TOKEN: ${{ github.token }}   
+      - name: Cache compilation cache
+        id: query-cache
+        uses: ./.github/actions/cache-query-compilation
+        with: 
+          key: run-ql-for-ql
+      - name: Make database and analyze
         run: |
-          JSON=$(jq -nc --arg pack "${PACK}" '.database."run-queries"=["--search-path", $pack] | .resolve.queries=["--search-path", $pack] | .resolve.extractor=["--search-path", $pack] | .resolve.languages=["--search-path", $pack] | .database.init=["--search-path", $pack]')
-          echo "CODEQL_ACTION_EXTRA_OPTIONS=${JSON}" >> ${GITHUB_ENV}
-        env:
-          PACK: ${{ runner.temp }}/pack
-
-      - name: Create CodeQL config file
-        run: |
-          echo "paths-ignore:" >> ${CONF}
-          echo "  - ql/ql/test" >> ${CONF}
-          echo "  - \"*/ql/lib/upgrades/\"" >> ${CONF}
-          echo "disable-default-queries: true" >> ${CONF}
-          echo "queries:" >> ${CONF}
-          echo "  - uses: ./ql/ql/src/codeql-suites/ql-code-scanning.qls" >> ${CONF}
-          echo "Config file: "
-          cat ${CONF}
-        env:
-          CONF: ./ql-for-ql-config.yml
-      - name: Initialize CodeQL
-        uses: github/codeql-action/init@45955cb1830b640e2c1603ad72ad542a49d47b96
+          ./ql/target/release/buramu | tee deprecated.blame # Add a blame file for the extractor to parse.
+          ${CODEQL} database create -l=ql --search-path ql/extractor-pack ${DB}
+          ${CODEQL} database analyze -j0 --format=sarif-latest --output=ql-for-ql.sarif ${DB} ql/ql/src/codeql-suites/ql-code-scanning.qls  --compilation-cache "${{ steps.query-cache.outputs.cache-dir }}"
+        env: 
+          CODEQL: ${{ steps.find-codeql.outputs.codeql-path }}
+          DB: ${{ runner.temp }}/DB
+          LGTM_INDEX_FILTERS: |
+            exclude:ql/ql/test
+            exclude:*/ql/lib/upgrades/
+            exclude:java/ql/integration-tests
+      - name: Upload sarif to code-scanning
+        uses: github/codeql-action/upload-sarif@v2
         with:
-          languages: ql
-          db-location: ${{ runner.temp }}/db
-          config-file: ./ql-for-ql-config.yml
-          tools: ${{ steps.find-latest-bundle.outputs.url }}
-      - name: Move pack queries
-        run: |
-          cp -r ${PACK}/queries ql/ql/src
-        env:
-          PACK: ${{ runner.temp }}/pack
-
-      - name: Perform CodeQL Analysis
-        uses: github/codeql-action/analyze@45955cb1830b640e2c1603ad72ad542a49d47b96
-        with:
-          category: "ql-for-ql"
-      - name: Copy sarif file to CWD
-        run: cp ../results/ql.sarif ./ql-for-ql.sarif
-      - name: Fixup the $scema in sarif  # Until https://github.com/microsoft/sarif-vscode-extension/pull/436/ is part in a stable release
-        run: |
-          sed -i 's/\$schema.*/\$schema": "https:\/\/raw.githubusercontent.com\/oasis-tcs\/sarif-spec\/master\/Schemata\/sarif-schema-2.1.0",/' ql-for-ql.sarif
+          sarif_file: ql-for-ql.sarif
+          category: ql-for-ql
       - name: Sarif as artifact
         uses: actions/upload-artifact@v3
         with:

.github/workflows/ql-for-ql-dataset_measure.yml

@@ -25,7 +25,7 @@ jobs:
 
       - name: Find codeql
         id: find-codeql
-        uses: github/codeql-action/init@45955cb1830b640e2c1603ad72ad542a49d47b96
+        uses: github/codeql-action/init@v2
         with:
           languages: javascript # does not matter
       - uses: ./.github/actions/os-version

.github/workflows/ql-for-ql-tests.yml

@@ -6,11 +6,13 @@ on:
     paths:
       - "ql/**"
       - codeql-workspace.yml
+      - .github/workflows/ql-for-ql-tests.yml
   pull_request:
     branches: [main]
     paths:
       - "ql/**"
       - codeql-workspace.yml
+      - .github/workflows/ql-for-ql-tests.yml
 
 env:
   CARGO_TERM_COLOR: always
@@ -22,7 +24,7 @@ jobs:
       - uses: actions/checkout@v3
       - name: Find codeql
         id: find-codeql
-        uses: github/codeql-action/init@45955cb1830b640e2c1603ad72ad542a49d47b96
+        uses: github/codeql-action/init@v2
         with:
           languages: javascript # does not matter
       - uses: ./.github/actions/os-version
@@ -34,6 +36,8 @@ jobs:
             ~/.cargo/git
             ql/target
           key: ${{ runner.os }}-${{ steps.os_version.outputs.version }}-qltest-cargo-${{ hashFiles('ql/rust-toolchain.toml', 'ql/**/Cargo.lock') }}
+      - name: Check formatting
+        run: cd ql; cargo fmt --all -- --check
       - name: Build extractor
         run: |
           cd ql;
@@ -65,7 +69,7 @@ jobs:
           echo "/usr/local/opt/gnu-tar/libexec/gnubin" >> $GITHUB_PATH
       - name: Find codeql
         id: find-codeql
-        uses: github/codeql-action/init@77a8d2d10c0b403a8b4aadbd223dc489ecd22683
+        uses: github/codeql-action/init@v2
         with:
           languages: javascript # does not matter
       - uses: ./.github/actions/os-version

.github/workflows/swift.yml

@@ -5,6 +5,7 @@ on:
     paths:
       - "swift/**"
       - "misc/bazel/**"
+      - "misc/codegen/**"
       - "*.bazel*"
       - .github/workflows/swift.yml
       - .github/actions/**
@@ -19,6 +20,7 @@ on:
     paths:
       - "swift/**"
       - "misc/bazel/**"
+      - "misc/codegen/**"
       - "*.bazel*"
       - .github/workflows/swift.yml
       - .github/actions/**

.pre-commit-config.yaml

@@ -53,5 +53,5 @@ repos:
         name: Run Swift code generation unit tests
         files: ^swift/codegen/.*\.py$
         language: system
-        entry: bazel test //swift/codegen/test
+        entry: bazel test //misc/codegen/test
         pass_filenames: false

CODEOWNERS

@@ -2,10 +2,11 @@
 /csharp/ @github/codeql-csharp
 /go/ @github/codeql-go
 /java/ @github/codeql-java
-/javascript/ @github/codeql-javascript
-/python/ @github/codeql-python
-/ruby/ @github/codeql-ruby
+/javascript/ @github/codeql-dynamic
+/python/ @github/codeql-dynamic
+/ruby/ @github/codeql-dynamic
 /swift/ @github/codeql-swift
+/misc/codegen/ @github/codeql-swift
 /java/kotlin-extractor/ @github/codeql-kotlin
 /java/kotlin-explorer/ @github/codeql-kotlin
 

README.md

@@ -10,6 +10,8 @@ There is [extensive documentation](https://codeql.github.com/docs/) on getting s
 
 We welcome contributions to our standard library and standard checks. Do you have an idea for a new check, or how to improve an existing query? Then please go ahead and open a pull request! Before you do, though, please take the time to read our [contributing guidelines](CONTRIBUTING.md). You can also consult our [style guides](https://github.com/github/codeql/tree/main/docs) to learn how to format your code for consistency and clarity, how to write query metadata, and how to write query help documentation for your query.
 
+For information on contributing to CodeQL documentation, see the "[contributing guide](docs/codeql/CONTRIBUTING.md)" for docs.
+
 ## License
 
 The code in this repository is licensed under the [MIT License](LICENSE) by [GitHub](https://github.com).

config/identical-files.json

@@ -1,6 +1,28 @@
 {
   "DataFlow Java/C++/C#/Go/Python/Ruby/Swift": [
+    "java/ql/lib/semmle/code/java/dataflow/internal/DataFlow.qll",
+    "cpp/ql/lib/semmle/code/cpp/dataflow/internal/DataFlow.qll",
+    "cpp/ql/lib/semmle/code/cpp/ir/dataflow/internal/DataFlow.qll",
+    "cpp/ql/lib/experimental/semmle/code/cpp/ir/dataflow/internal/DataFlow.qll",
+    "csharp/ql/lib/semmle/code/csharp/dataflow/internal/DataFlow.qll",
+    "go/ql/lib/semmle/go/dataflow/internal/DataFlow.qll",
+    "python/ql/lib/semmle/python/dataflow/new/internal/DataFlow.qll",
+    "ruby/ql/lib/codeql/ruby/dataflow/internal/DataFlow.qll",
+    "swift/ql/lib/codeql/swift/dataflow/internal/DataFlow.qll"
+  ],
+  "DataFlowImpl Java/C++/C#/Go/Python/Ruby/Swift": [
     "java/ql/lib/semmle/code/java/dataflow/internal/DataFlowImpl.qll",
+    "cpp/ql/lib/semmle/code/cpp/dataflow/internal/DataFlowImpl.qll",
+    "cpp/ql/lib/semmle/code/cpp/ir/dataflow/internal/DataFlowImpl.qll",
+    "cpp/ql/lib/experimental/semmle/code/cpp/ir/dataflow/internal/DataFlowImpl.qll",
+    "csharp/ql/lib/semmle/code/csharp/dataflow/internal/DataFlowImpl.qll",
+    "go/ql/lib/semmle/go/dataflow/internal/DataFlowImpl.qll",
+    "python/ql/lib/semmle/python/dataflow/new/internal/DataFlowImpl.qll",
+    "ruby/ql/lib/codeql/ruby/dataflow/internal/DataFlowImpl.qll",
+    "swift/ql/lib/codeql/swift/dataflow/internal/DataFlowImpl.qll"
+  ],
+  "DataFlow Java/C++/C#/Go/Python/Ruby/Swift Legacy Configuration": [
+    "java/ql/lib/semmle/code/java/dataflow/internal/DataFlowImpl1.qll",
     "java/ql/lib/semmle/code/java/dataflow/internal/DataFlowImpl2.qll",
     "java/ql/lib/semmle/code/java/dataflow/internal/DataFlowImpl3.qll",
     "java/ql/lib/semmle/code/java/dataflow/internal/DataFlowImpl4.qll",
@@ -8,38 +30,38 @@
     "java/ql/lib/semmle/code/java/dataflow/internal/DataFlowImpl6.qll",
     "java/ql/lib/semmle/code/java/dataflow/internal/DataFlowImplForSerializability.qll",
     "java/ql/lib/semmle/code/java/dataflow/internal/DataFlowImplForOnActivityResult.qll",
-    "cpp/ql/lib/semmle/code/cpp/dataflow/internal/DataFlowImpl.qll",
+    "cpp/ql/lib/semmle/code/cpp/dataflow/internal/DataFlowImpl1.qll",
     "cpp/ql/lib/semmle/code/cpp/dataflow/internal/DataFlowImpl2.qll",
     "cpp/ql/lib/semmle/code/cpp/dataflow/internal/DataFlowImpl3.qll",
     "cpp/ql/lib/semmle/code/cpp/dataflow/internal/DataFlowImpl4.qll",
     "cpp/ql/lib/semmle/code/cpp/dataflow/internal/DataFlowImplLocal.qll",
-    "cpp/ql/lib/semmle/code/cpp/ir/dataflow/internal/DataFlowImpl.qll",
+    "cpp/ql/lib/semmle/code/cpp/ir/dataflow/internal/DataFlowImpl1.qll",
     "cpp/ql/lib/semmle/code/cpp/ir/dataflow/internal/DataFlowImpl2.qll",
     "cpp/ql/lib/semmle/code/cpp/ir/dataflow/internal/DataFlowImpl3.qll",
     "cpp/ql/lib/semmle/code/cpp/ir/dataflow/internal/DataFlowImpl4.qll",
-    "cpp/ql/lib/experimental/semmle/code/cpp/ir/dataflow/internal/DataFlowImpl.qll",
+    "cpp/ql/lib/experimental/semmle/code/cpp/ir/dataflow/internal/DataFlowImpl1.qll",
     "cpp/ql/lib/experimental/semmle/code/cpp/ir/dataflow/internal/DataFlowImpl2.qll",
     "cpp/ql/lib/experimental/semmle/code/cpp/ir/dataflow/internal/DataFlowImpl3.qll",
     "cpp/ql/lib/experimental/semmle/code/cpp/ir/dataflow/internal/DataFlowImpl4.qll",
-    "csharp/ql/lib/semmle/code/csharp/dataflow/internal/DataFlowImpl.qll",
+    "csharp/ql/lib/semmle/code/csharp/dataflow/internal/DataFlowImpl1.qll",
     "csharp/ql/lib/semmle/code/csharp/dataflow/internal/DataFlowImpl2.qll",
     "csharp/ql/lib/semmle/code/csharp/dataflow/internal/DataFlowImpl3.qll",
     "csharp/ql/lib/semmle/code/csharp/dataflow/internal/DataFlowImpl4.qll",
     "csharp/ql/lib/semmle/code/csharp/dataflow/internal/DataFlowImpl5.qll",
     "csharp/ql/lib/semmle/code/csharp/dataflow/internal/DataFlowImplForContentDataFlow.qll",
-    "go/ql/lib/semmle/go/dataflow/internal/DataFlowImpl.qll",
+    "go/ql/lib/semmle/go/dataflow/internal/DataFlowImpl1.qll",
     "go/ql/lib/semmle/go/dataflow/internal/DataFlowImpl2.qll",
     "go/ql/lib/semmle/go/dataflow/internal/DataFlowImplForStringsNewReplacer.qll",
-    "python/ql/lib/semmle/python/dataflow/new/internal/DataFlowImpl.qll",
+    "python/ql/lib/semmle/python/dataflow/new/internal/DataFlowImpl1.qll",
     "python/ql/lib/semmle/python/dataflow/new/internal/DataFlowImpl2.qll",
     "python/ql/lib/semmle/python/dataflow/new/internal/DataFlowImpl3.qll",
     "python/ql/lib/semmle/python/dataflow/new/internal/DataFlowImpl4.qll",
-    "ruby/ql/lib/codeql/ruby/dataflow/internal/DataFlowImpl.qll",
+    "python/ql/lib/semmle/python/dataflow/new/internal/DataFlowImplForRegExp.qll",
+    "ruby/ql/lib/codeql/ruby/dataflow/internal/DataFlowImpl1.qll",
     "ruby/ql/lib/codeql/ruby/dataflow/internal/DataFlowImpl2.qll",
-    "ruby/ql/lib/codeql/ruby/dataflow/internal/DataFlowImplForRegExp.qll",
     "ruby/ql/lib/codeql/ruby/dataflow/internal/DataFlowImplForHttpClientLibraries.qll",
     "ruby/ql/lib/codeql/ruby/dataflow/internal/DataFlowImplForPathname.qll",
-    "swift/ql/lib/codeql/swift/dataflow/internal/DataFlowImpl.qll"
+    "swift/ql/lib/codeql/swift/dataflow/internal/DataFlowImpl1.qll"
   ],
   "DataFlow Java/C++/C#/Go/Python/Ruby/Swift Common": [
     "java/ql/lib/semmle/code/java/dataflow/internal/DataFlowImplCommon.qll",
@@ -52,7 +74,18 @@
     "ruby/ql/lib/codeql/ruby/dataflow/internal/DataFlowImplCommon.qll",
     "swift/ql/lib/codeql/swift/dataflow/internal/DataFlowImplCommon.qll"
   ],
-  "TaintTracking::Configuration Java/C++/C#/Go/Python/Ruby/Swift": [
+  "TaintTracking Java/C++/C#/Go/Python/Ruby/Swift": [
+    "cpp/ql/lib/semmle/code/cpp/dataflow/internal/tainttracking1/TaintTracking.qll",
+    "cpp/ql/lib/semmle/code/cpp/ir/dataflow/internal/tainttracking1/TaintTracking.qll",
+    "cpp/ql/lib/experimental/semmle/code/cpp/ir/dataflow/internal/tainttracking1/TaintTracking.qll",
+    "csharp/ql/lib/semmle/code/csharp/dataflow/internal/tainttracking1/TaintTracking.qll",
+    "go/ql/lib/semmle/go/dataflow/internal/tainttracking1/TaintTracking.qll",
+    "java/ql/lib/semmle/code/java/dataflow/internal/tainttracking1/TaintTracking.qll",
+    "python/ql/lib/semmle/python/dataflow/new/internal/tainttracking1/TaintTracking.qll",
+    "ruby/ql/lib/codeql/ruby/dataflow/internal/tainttracking1/TaintTracking.qll",
+    "swift/ql/lib/codeql/swift/dataflow/internal/tainttracking1/TaintTracking.qll"
+  ],
+  "TaintTracking Legacy Configuration Java/C++/C#/Go/Python/Ruby/Swift": [
     "cpp/ql/lib/semmle/code/cpp/dataflow/internal/tainttracking1/TaintTrackingImpl.qll",
     "cpp/ql/lib/semmle/code/cpp/dataflow/internal/tainttracking2/TaintTrackingImpl.qll",
     "cpp/ql/lib/semmle/code/cpp/ir/dataflow/internal/tainttracking1/TaintTrackingImpl.qll",

cpp/autobuilder/Semmle.Autobuild.Cpp.Tests/BuildScripts.cs

@@ -1,5 +1,6 @@
 using Xunit;
 using Semmle.Autobuild.Shared;
+using Semmle.Util;
 using System.Collections.Generic;
 using System;
 using System.Linq;
@@ -75,6 +76,15 @@ namespace Semmle.Autobuild.Cpp.Tests
             throw new ArgumentException("Missing RunProcess " + pattern);
         }
 
+        int IBuildActions.RunProcess(string cmd, string args, string? workingDirectory, IDictionary<string, string>? env, BuildOutputHandler onOutput, BuildOutputHandler onError)
+        {
+            var ret = (this as IBuildActions).RunProcess(cmd, args, workingDirectory, env, out var stdout);
+
+            stdout.ForEach(line => onOutput(line));
+
+            return ret;
+        }
+
         public IList<string> DirectoryDeleteIn = new List<string>();
 
         void IBuildActions.DirectoryDelete(string dir, bool recursive)
@@ -131,6 +141,14 @@ namespace Semmle.Autobuild.Cpp.Tests
 
         bool IBuildActions.IsWindows() => IsWindows;
 
+        public bool IsMacOs { get; set; }
+
+        bool IBuildActions.IsMacOs() => IsMacOs;
+
+        public bool IsArm { get; set; }
+
+        bool IBuildActions.IsArm() => IsArm;
+
         string IBuildActions.PathCombine(params string[] parts)
         {
             return string.Join(IsWindows ? '\\' : '/', parts.Where(p => !string.IsNullOrWhiteSpace(p)));
@@ -176,6 +194,15 @@ namespace Semmle.Autobuild.Cpp.Tests
             if (!DownloadFiles.Contains((address, fileName)))
                 throw new ArgumentException($"Missing DownloadFile, {address}, {fileName}");
         }
+
+        public IDiagnosticsWriter CreateDiagnosticsWriter(string filename) => new TestDiagnosticWriter();
+    }
+
+    internal class TestDiagnosticWriter : IDiagnosticsWriter
+    {
+        public IList<DiagnosticMessage> Diagnostics { get; } = new List<DiagnosticMessage>();
+
+        public void AddEntry(DiagnosticMessage message) => this.Diagnostics.Add(message);
     }
 
     /// <summary>
@@ -235,6 +262,7 @@ namespace Semmle.Autobuild.Cpp.Tests
             Actions.GetEnvironmentVariable[$"CODEQL_EXTRACTOR_{codeqlUpperLanguage}_TRAP_DIR"] = "";
             Actions.GetEnvironmentVariable[$"CODEQL_EXTRACTOR_{codeqlUpperLanguage}_SOURCE_ARCHIVE_DIR"] = "";
             Actions.GetEnvironmentVariable[$"CODEQL_EXTRACTOR_{codeqlUpperLanguage}_ROOT"] = $@"C:\codeql\{codeqlUpperLanguage.ToLowerInvariant()}";
+            Actions.GetEnvironmentVariable[$"CODEQL_EXTRACTOR_{codeqlUpperLanguage}_DIAGNOSTIC_DIR"] = "";
             Actions.GetEnvironmentVariable["CODEQL_JAVA_HOME"] = @"C:\codeql\tools\java";
             Actions.GetEnvironmentVariable["CODEQL_PLATFORM"] = "win64";
             Actions.GetEnvironmentVariable["SEMMLE_DIST"] = @"C:\odasa";

cpp/autobuilder/Semmle.Autobuild.Cpp.Tests/Semmle.Autobuild.Cpp.Tests.csproj

@@ -2,7 +2,7 @@
 
   <PropertyGroup>
     <OutputType>Exe</OutputType>
-    <TargetFramework>net6.0</TargetFramework>
+    <TargetFramework>net7.0</TargetFramework>
     <GenerateAssemblyInfo>false</GenerateAssemblyInfo>
     <RuntimeIdentifiers>win-x64;linux-x64;osx-x64</RuntimeIdentifiers>
     <Nullable>enable</Nullable>

cpp/autobuilder/Semmle.Autobuild.Cpp/CppAutobuilder.cs

@@ -1,4 +1,5 @@
 using Semmle.Autobuild.Shared;
+using Semmle.Util;
 
 namespace Semmle.Autobuild.Cpp
 {
@@ -21,7 +22,7 @@ namespace Semmle.Autobuild.Cpp
 
     public class CppAutobuilder : Autobuilder<CppAutobuildOptions>
     {
-        public CppAutobuilder(IBuildActions actions, CppAutobuildOptions options) : base(actions, options) { }
+        public CppAutobuilder(IBuildActions actions, CppAutobuildOptions options) : base(actions, options, new DiagnosticClassifier()) { }
 
         public override BuildScript GetBuildScript()
         {

cpp/autobuilder/Semmle.Autobuild.Cpp/Semmle.Autobuild.Cpp.csproj

@@ -1,7 +1,7 @@
 <Project Sdk="Microsoft.NET.Sdk">
 
   <PropertyGroup>
-    <TargetFramework>net6.0</TargetFramework>
+    <TargetFramework>net7.0</TargetFramework>
     <AssemblyName>Semmle.Autobuild.Cpp</AssemblyName>
     <RootNamespace>Semmle.Autobuild.Cpp</RootNamespace>
     <ApplicationIcon />

cpp/ql/lib/CHANGELOG.md

@@ -1,3 +1,11 @@
+## 0.5.3
+
+No user-facing changes.
+
+## 0.5.2
+
+No user-facing changes.
+
 ## 0.5.1
 
 No user-facing changes.

cpp/ql/lib/change-notes/2023-03-02-dataflow-conf-module.md

@@ -0,0 +1,9 @@
+---
+category: majorAnalysis
+---
+* The main data flow and taint tracking APIs have been changed. The old APIs
+  remain in place for now and translate to the new through a
+  backwards-compatible wrapper. If multiple configurations are in scope
+  simultaneously, then this may affect results slightly. The new API is quite
+  similar to the old, but makes use of a configuration module instead of a
+  configuration class.

cpp/ql/lib/change-notes/released/0.5.2.md

@@ -0,0 +1,3 @@
+## 0.5.2
+
+No user-facing changes.

cpp/ql/lib/change-notes/released/0.5.3.md

@@ -0,0 +1,3 @@
+## 0.5.3
+
+No user-facing changes.

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

@@ -1,2 +1,2 @@
 ---
-lastReleaseVersion: 0.5.1
+lastReleaseVersion: 0.5.3

cpp/ql/lib/experimental/semmle/code/cpp/ir/dataflow/DataFlow.qll

@@ -22,5 +22,6 @@
 import cpp
 
 module DataFlow {
-  import experimental.semmle.code.cpp.ir.dataflow.internal.DataFlowImpl
+  import experimental.semmle.code.cpp.ir.dataflow.internal.DataFlow
+  import experimental.semmle.code.cpp.ir.dataflow.internal.DataFlowImpl1
 }

cpp/ql/lib/experimental/semmle/code/cpp/ir/dataflow/TaintTracking.qll

@@ -19,5 +19,6 @@ import semmle.code.cpp.ir.dataflow.DataFlow
 import semmle.code.cpp.ir.dataflow.DataFlow2
 
 module TaintTracking {
+  import experimental.semmle.code.cpp.ir.dataflow.internal.tainttracking1.TaintTracking
   import experimental.semmle.code.cpp.ir.dataflow.internal.tainttracking1.TaintTrackingImpl
 }

cpp/ql/lib/experimental/semmle/code/cpp/ir/dataflow/internal/DataFlow.qll

@@ -0,0 +1,245 @@
+/**
+ * Provides an implementation of global (interprocedural) data flow. This file
+ * re-exports the local (intraprocedural) data flow analysis from
+ * `DataFlowImplSpecific::Public` and adds a global analysis, mainly exposed
+ * through the `Make` and `MakeWithState` modules.
+ */
+
+private import DataFlowImplCommon
+private import DataFlowImplSpecific::Private
+import DataFlowImplSpecific::Public
+import DataFlowImplCommonPublic
+private import DataFlowImpl
+
+/** An input configuration for data flow. */
+signature module ConfigSig {
+  /**
+   * Holds if `source` is a relevant data flow source.
+   */
+  predicate isSource(Node source);
+
+  /**
+   * Holds if `sink` is a relevant data flow sink.
+   */
+  predicate isSink(Node sink);
+
+  /**
+   * Holds if data flow through `node` is prohibited. This completely removes
+   * `node` from the data flow graph.
+   */
+  default predicate isBarrier(Node node) { none() }
+
+  /** Holds if data flow into `node` is prohibited. */
+  default predicate isBarrierIn(Node node) { none() }
+
+  /** Holds if data flow out of `node` is prohibited. */
+  default predicate isBarrierOut(Node node) { none() }
+
+  /**
+   * Holds if data may flow from `node1` to `node2` in addition to the normal data-flow steps.
+   */
+  default predicate isAdditionalFlowStep(Node node1, Node node2) { none() }
+
+  /**
+   * Holds if an arbitrary number of implicit read steps of content `c` may be
+   * taken at `node`.
+   */
+  default predicate allowImplicitRead(Node node, ContentSet c) { none() }
+
+  /**
+   * Gets the virtual dispatch branching limit when calculating field flow.
+   * This can be overridden to a smaller value to improve performance (a
+   * value of 0 disables field flow), or a larger value to get more results.
+   */
+  default int fieldFlowBranchLimit() { result = 2 }
+
+  /**
+   * Gets a data flow configuration feature to add restrictions to the set of
+   * valid flow paths.
+   *
+   * - `FeatureHasSourceCallContext`:
+   *    Assume that sources have some existing call context to disallow
+   *    conflicting return-flow directly following the source.
+   * - `FeatureHasSinkCallContext`:
+   *    Assume that sinks have some existing call context to disallow
+   *    conflicting argument-to-parameter flow directly preceding the sink.
+   * - `FeatureEqualSourceSinkCallContext`:
+   *    Implies both of the above and additionally ensures that the entire flow
+   *    path preserves the call context.
+   *
+   * These features are generally not relevant for typical end-to-end data flow
+   * queries, but should only be used for constructing paths that need to
+   * somehow be pluggable in another path context.
+   */
+  default FlowFeature getAFeature() { none() }
+
+  /** Holds if sources should be grouped in the result of `hasFlowPath`. */
+  default predicate sourceGrouping(Node source, string sourceGroup) { none() }
+
+  /** Holds if sinks should be grouped in the result of `hasFlowPath`. */
+  default predicate sinkGrouping(Node sink, string sinkGroup) { none() }
+
+  /**
+   * Holds if hidden nodes should be included in the data flow graph.
+   *
+   * This feature should only be used for debugging or when the data flow graph
+   * is not visualized (as it is in a `path-problem` query).
+   */
+  default predicate includeHiddenNodes() { none() }
+}
+
+/** An input configuration for data flow using flow state. */
+signature module StateConfigSig {
+  bindingset[this]
+  class FlowState;
+
+  /**
+   * Holds if `source` is a relevant data flow source with the given initial
+   * `state`.
+   */
+  predicate isSource(Node source, FlowState state);
+
+  /**
+   * Holds if `sink` is a relevant data flow sink accepting `state`.
+   */
+  predicate isSink(Node sink, FlowState state);
+
+  /**
+   * Holds if data flow through `node` is prohibited. This completely removes
+   * `node` from the data flow graph.
+   */
+  default predicate isBarrier(Node node) { none() }
+
+  /**
+   * Holds if data flow through `node` is prohibited when the flow state is
+   * `state`.
+   */
+  predicate isBarrier(Node node, FlowState state);
+
+  /** Holds if data flow into `node` is prohibited. */
+  default predicate isBarrierIn(Node node) { none() }
+
+  /** Holds if data flow out of `node` is prohibited. */
+  default predicate isBarrierOut(Node node) { none() }
+
+  /**
+   * Holds if data may flow from `node1` to `node2` in addition to the normal data-flow steps.
+   */
+  default predicate isAdditionalFlowStep(Node node1, Node node2) { none() }
+
+  /**
+   * Holds if data may flow from `node1` to `node2` in addition to the normal data-flow steps.
+   * This step is only applicable in `state1` and updates the flow state to `state2`.
+   */
+  predicate isAdditionalFlowStep(Node node1, FlowState state1, Node node2, FlowState state2);
+
+  /**
+   * Holds if an arbitrary number of implicit read steps of content `c` may be
+   * taken at `node`.
+   */
+  default predicate allowImplicitRead(Node node, ContentSet c) { none() }
+
+  /**
+   * Gets the virtual dispatch branching limit when calculating field flow.
+   * This can be overridden to a smaller value to improve performance (a
+   * value of 0 disables field flow), or a larger value to get more results.
+   */
+  default int fieldFlowBranchLimit() { result = 2 }
+
+  /**
+   * Gets a data flow configuration feature to add restrictions to the set of
+   * valid flow paths.
+   *
+   * - `FeatureHasSourceCallContext`:
+   *    Assume that sources have some existing call context to disallow
+   *    conflicting return-flow directly following the source.
+   * - `FeatureHasSinkCallContext`:
+   *    Assume that sinks have some existing call context to disallow
+   *    conflicting argument-to-parameter flow directly preceding the sink.
+   * - `FeatureEqualSourceSinkCallContext`:
+   *    Implies both of the above and additionally ensures that the entire flow
+   *    path preserves the call context.
+   *
+   * These features are generally not relevant for typical end-to-end data flow
+   * queries, but should only be used for constructing paths that need to
+   * somehow be pluggable in another path context.
+   */
+  default FlowFeature getAFeature() { none() }
+
+  /** Holds if sources should be grouped in the result of `hasFlowPath`. */
+  default predicate sourceGrouping(Node source, string sourceGroup) { none() }
+
+  /** Holds if sinks should be grouped in the result of `hasFlowPath`. */
+  default predicate sinkGrouping(Node sink, string sinkGroup) { none() }
+
+  /**
+   * Holds if hidden nodes should be included in the data flow graph.
+   *
+   * This feature should only be used for debugging or when the data flow graph
+   * is not visualized (as it is in a `path-problem` query).
+   */
+  default predicate includeHiddenNodes() { none() }
+}
+
+/**
+ * Gets the exploration limit for `hasPartialFlow` and `hasPartialFlowRev`
+ * measured in approximate number of interprocedural steps.
+ */
+signature int explorationLimitSig();
+
+/**
+ * The output of a data flow computation.
+ */
+signature module DataFlowSig {
+  /**
+   * A `Node` augmented with a call context (except for sinks) and an access path.
+   * Only those `PathNode`s that are reachable from a source, and which can reach a sink, are generated.
+   */
+  class PathNode;
+
+  /**
+   * Holds if data can flow from `source` to `sink`.
+   *
+   * The corresponding paths are generated from the end-points and the graph
+   * included in the module `PathGraph`.
+   */
+  predicate hasFlowPath(PathNode source, PathNode sink);
+
+  /**
+   * Holds if data can flow from `source` to `sink`.
+   */
+  predicate hasFlow(Node source, Node sink);
+
+  /**
+   * Holds if data can flow from some source to `sink`.
+   */
+  predicate hasFlowTo(Node sink);
+
+  /**
+   * Holds if data can flow from some source to `sink`.
+   */
+  predicate hasFlowToExpr(DataFlowExpr sink);
+}
+
+/**
+ * Constructs a standard data flow computation.
+ */
+module Make<ConfigSig Config> implements DataFlowSig {
+  private module C implements FullStateConfigSig {
+    import DefaultState<Config>
+    import Config
+  }
+
+  import Impl<C>
+}
+
+/**
+ * Constructs a data flow computation using flow state.
+ */
+module MakeWithState<StateConfigSig Config> implements DataFlowSig {
+  private module C implements FullStateConfigSig {
+    import Config
+  }
+
+  import Impl<C>
+}

cpp/ql/lib/experimental/semmle/code/cpp/ir/dataflow/internal/DataFlowImpl1.qll

@@ -0,0 +1,396 @@
+/**
+ * DEPRECATED: Use `Make` and `MakeWithState` instead.
+ *
+ * Provides a `Configuration` class backwards-compatible interface to the data
+ * flow library.
+ */
+
+private import DataFlowImplCommon
+private import DataFlowImplSpecific::Private
+import DataFlowImplSpecific::Public
+private import DataFlowImpl
+import DataFlowImplCommonPublic
+import FlowStateString
+
+/**
+ * A configuration of interprocedural data flow analysis. This defines
+ * sources, sinks, and any other configurable aspect of the analysis. Each
+ * use of the global data flow library must define its own unique extension
+ * of this abstract class. To create a configuration, extend this class with
+ * a subclass whose characteristic predicate is a unique singleton string.
+ * For example, write
+ *
+ * ```ql
+ * class MyAnalysisConfiguration extends DataFlow::Configuration {
+ *   MyAnalysisConfiguration() { this = "MyAnalysisConfiguration" }
+ *   // Override `isSource` and `isSink`.
+ *   // Optionally override `isBarrier`.
+ *   // Optionally override `isAdditionalFlowStep`.
+ * }
+ * ```
+ * Conceptually, this defines a graph where the nodes are `DataFlow::Node`s and
+ * the edges are those data-flow steps that preserve the value of the node
+ * along with any additional edges defined by `isAdditionalFlowStep`.
+ * Specifying nodes in `isBarrier` will remove those nodes from the graph, and
+ * specifying nodes in `isBarrierIn` and/or `isBarrierOut` will remove in-going
+ * and/or out-going edges from those nodes, respectively.
+ *
+ * Then, to query whether there is flow between some `source` and `sink`,
+ * write
+ *
+ * ```ql
+ * exists(MyAnalysisConfiguration cfg | cfg.hasFlow(source, sink))
+ * ```
+ *
+ * Multiple configurations can coexist, but two classes extending
+ * `DataFlow::Configuration` should never depend on each other. One of them
+ * should instead depend on a `DataFlow2::Configuration`, a
+ * `DataFlow3::Configuration`, or a `DataFlow4::Configuration`.
+ */
+abstract class Configuration extends string {
+  bindingset[this]
+  Configuration() { any() }
+
+  /**
+   * Holds if `source` is a relevant data flow source.
+   */
+  predicate isSource(Node source) { none() }
+
+  /**
+   * Holds if `source` is a relevant data flow source with the given initial
+   * `state`.
+   */
+  predicate isSource(Node source, FlowState state) { none() }
+
+  /**
+   * Holds if `sink` is a relevant data flow sink.
+   */
+  predicate isSink(Node sink) { none() }
+
+  /**
+   * Holds if `sink` is a relevant data flow sink accepting `state`.
+   */
+  predicate isSink(Node sink, FlowState state) { none() }
+
+  /**
+   * Holds if data flow through `node` is prohibited. This completely removes
+   * `node` from the data flow graph.
+   */
+  predicate isBarrier(Node node) { none() }
+
+  /**
+   * Holds if data flow through `node` is prohibited when the flow state is
+   * `state`.
+   */
+  predicate isBarrier(Node node, FlowState state) { none() }
+
+  /** Holds if data flow into `node` is prohibited. */
+  predicate isBarrierIn(Node node) { none() }
+
+  /** Holds if data flow out of `node` is prohibited. */
+  predicate isBarrierOut(Node node) { none() }
+
+  /**
+   * DEPRECATED: Use `isBarrier` and `BarrierGuard` module instead.
+   *
+   * Holds if data flow through nodes guarded by `guard` is prohibited.
+   */
+  deprecated predicate isBarrierGuard(BarrierGuard guard) { none() }
+
+  /**
+   * DEPRECATED: Use `isBarrier` and `BarrierGuard` module instead.
+   *
+   * Holds if data flow through nodes guarded by `guard` is prohibited when
+   * the flow state is `state`
+   */
+  deprecated predicate isBarrierGuard(BarrierGuard guard, FlowState state) { none() }
+
+  /**
+   * Holds if data may flow from `node1` to `node2` in addition to the normal data-flow steps.
+   */
+  predicate isAdditionalFlowStep(Node node1, Node node2) { none() }
+
+  /**
+   * Holds if data may flow from `node1` to `node2` in addition to the normal data-flow steps.
+   * This step is only applicable in `state1` and updates the flow state to `state2`.
+   */
+  predicate isAdditionalFlowStep(Node node1, FlowState state1, Node node2, FlowState state2) {
+    none()
+  }
+
+  /**
+   * Holds if an arbitrary number of implicit read steps of content `c` may be
+   * taken at `node`.
+   */
+  predicate allowImplicitRead(Node node, ContentSet c) { none() }
+
+  /**
+   * Gets the virtual dispatch branching limit when calculating field flow.
+   * This can be overridden to a smaller value to improve performance (a
+   * value of 0 disables field flow), or a larger value to get more results.
+   */
+  int fieldFlowBranchLimit() { result = 2 }
+
+  /**
+   * Gets a data flow configuration feature to add restrictions to the set of
+   * valid flow paths.
+   *
+   * - `FeatureHasSourceCallContext`:
+   *    Assume that sources have some existing call context to disallow
+   *    conflicting return-flow directly following the source.
+   * - `FeatureHasSinkCallContext`:
+   *    Assume that sinks have some existing call context to disallow
+   *    conflicting argument-to-parameter flow directly preceding the sink.
+   * - `FeatureEqualSourceSinkCallContext`:
+   *    Implies both of the above and additionally ensures that the entire flow
+   *    path preserves the call context.
+   *
+   * These features are generally not relevant for typical end-to-end data flow
+   * queries, but should only be used for constructing paths that need to
+   * somehow be pluggable in another path context.
+   */
+  FlowFeature getAFeature() { none() }
+
+  /** Holds if sources should be grouped in the result of `hasFlowPath`. */
+  predicate sourceGrouping(Node source, string sourceGroup) { none() }
+
+  /** Holds if sinks should be grouped in the result of `hasFlowPath`. */
+  predicate sinkGrouping(Node sink, string sinkGroup) { none() }
+
+  /**
+   * Holds if data may flow from `source` to `sink` for this configuration.
+   */
+  predicate hasFlow(Node source, Node sink) { hasFlow(source, sink, this) }
+
+  /**
+   * Holds if data may flow from `source` to `sink` for this configuration.
+   *
+   * The corresponding paths are generated from the end-points and the graph
+   * included in the module `PathGraph`.
+   */
+  predicate hasFlowPath(PathNode source, PathNode sink) { hasFlowPath(source, sink, this) }
+
+  /**
+   * Holds if data may flow from some source to `sink` for this configuration.
+   */
+  predicate hasFlowTo(Node sink) { hasFlowTo(sink, this) }
+
+  /**
+   * Holds if data may flow from some source to `sink` for this configuration.
+   */
+  predicate hasFlowToExpr(DataFlowExpr sink) { this.hasFlowTo(exprNode(sink)) }
+
+  /**
+   * DEPRECATED: Use `FlowExploration<explorationLimit>` instead.
+   *
+   * Gets the exploration limit for `hasPartialFlow` and `hasPartialFlowRev`
+   * measured in approximate number of interprocedural steps.
+   */
+  deprecated int explorationLimit() { none() }
+
+  /**
+   * Holds if hidden nodes should be included in the data flow graph.
+   *
+   * This feature should only be used for debugging or when the data flow graph
+   * is not visualized (for example in a `path-problem` query).
+   */
+  predicate includeHiddenNodes() { none() }
+}
+
+/**
+ * This class exists to prevent mutual recursion between the user-overridden
+ * member predicates of `Configuration` and the rest of the data-flow library.
+ * Good performance cannot be guaranteed in the presence of such recursion, so
+ * it should be replaced by using more than one copy of the data flow library.
+ */
+abstract private class ConfigurationRecursionPrevention extends Configuration {
+  bindingset[this]
+  ConfigurationRecursionPrevention() { any() }
+
+  override predicate hasFlow(Node source, Node sink) {
+    strictcount(Node n | this.isSource(n)) < 0
+    or
+    strictcount(Node n | this.isSource(n, _)) < 0
+    or
+    strictcount(Node n | this.isSink(n)) < 0
+    or
+    strictcount(Node n | this.isSink(n, _)) < 0
+    or
+    strictcount(Node n1, Node n2 | this.isAdditionalFlowStep(n1, n2)) < 0
+    or
+    strictcount(Node n1, Node n2 | this.isAdditionalFlowStep(n1, _, n2, _)) < 0
+    or
+    super.hasFlow(source, sink)
+  }
+}
+
+/** A bridge class to access the deprecated `isBarrierGuard`. */
+private class BarrierGuardGuardedNodeBridge extends Unit {
+  abstract predicate guardedNode(Node n, Configuration config);
+
+  abstract predicate guardedNode(Node n, FlowState state, Configuration config);
+}
+
+private class BarrierGuardGuardedNode extends BarrierGuardGuardedNodeBridge {
+  deprecated override predicate guardedNode(Node n, Configuration config) {
+    exists(BarrierGuard g |
+      config.isBarrierGuard(g) and
+      n = g.getAGuardedNode()
+    )
+  }
+
+  deprecated override predicate guardedNode(Node n, FlowState state, Configuration config) {
+    exists(BarrierGuard g |
+      config.isBarrierGuard(g, state) and
+      n = g.getAGuardedNode()
+    )
+  }
+}
+
+private FlowState relevantState(Configuration config) {
+  config.isSource(_, result) or
+  config.isSink(_, result) or
+  config.isBarrier(_, result) or
+  config.isAdditionalFlowStep(_, result, _, _) or
+  config.isAdditionalFlowStep(_, _, _, result)
+}
+
+private newtype TConfigState =
+  TMkConfigState(Configuration config, FlowState state) {
+    state = relevantState(config) or state instanceof FlowStateEmpty
+  }
+
+private Configuration getConfig(TConfigState state) { state = TMkConfigState(result, _) }
+
+private FlowState getState(TConfigState state) { state = TMkConfigState(_, result) }
+
+private predicate singleConfiguration() { 1 = strictcount(Configuration c) }
+
+private module Config implements FullStateConfigSig {
+  class FlowState = TConfigState;
+
+  predicate isSource(Node source, FlowState state) {
+    getConfig(state).isSource(source, getState(state))
+    or
+    getConfig(state).isSource(source) and getState(state) instanceof FlowStateEmpty
+  }
+
+  predicate isSink(Node sink, FlowState state) {
+    getConfig(state).isSink(sink, getState(state))
+    or
+    getConfig(state).isSink(sink) and getState(state) instanceof FlowStateEmpty
+  }
+
+  predicate isBarrier(Node node) { none() }
+
+  predicate isBarrier(Node node, FlowState state) {
+    getConfig(state).isBarrier(node, getState(state)) or
+    getConfig(state).isBarrier(node) or
+    any(BarrierGuardGuardedNodeBridge b).guardedNode(node, getState(state), getConfig(state)) or
+    any(BarrierGuardGuardedNodeBridge b).guardedNode(node, getConfig(state))
+  }
+
+  predicate isBarrierIn(Node node) { any(Configuration config).isBarrierIn(node) }
+
+  predicate isBarrierOut(Node node) { any(Configuration config).isBarrierOut(node) }
+
+  predicate isAdditionalFlowStep(Node node1, Node node2) {
+    singleConfiguration() and
+    any(Configuration config).isAdditionalFlowStep(node1, node2)
+  }
+
+  predicate isAdditionalFlowStep(Node node1, FlowState state1, Node node2, FlowState state2) {
+    getConfig(state1).isAdditionalFlowStep(node1, getState(state1), node2, getState(state2)) and
+    getConfig(state2) = getConfig(state1)
+    or
+    not singleConfiguration() and
+    getConfig(state1).isAdditionalFlowStep(node1, node2) and
+    state2 = state1
+  }
+
+  predicate allowImplicitRead(Node node, ContentSet c) {
+    any(Configuration config).allowImplicitRead(node, c)
+  }
+
+  int fieldFlowBranchLimit() { result = min(any(Configuration config).fieldFlowBranchLimit()) }
+
+  FlowFeature getAFeature() { result = any(Configuration config).getAFeature() }
+
+  predicate sourceGrouping(Node source, string sourceGroup) {
+    any(Configuration config).sourceGrouping(source, sourceGroup)
+  }
+
+  predicate sinkGrouping(Node sink, string sinkGroup) {
+    any(Configuration config).sinkGrouping(sink, sinkGroup)
+  }
+
+  predicate includeHiddenNodes() { any(Configuration config).includeHiddenNodes() }
+}
+
+private import Impl<Config> as I
+import I
+
+/**
+ * A `Node` augmented with a call context (except for sinks), an access path, and a configuration.
+ * Only those `PathNode`s that are reachable from a source, and which can reach a sink, are generated.
+ */
+class PathNode instanceof I::PathNode {
+  /** Gets a textual representation of this element. */
+  final string toString() { result = super.toString() }
+
+  /**
+   * Gets a textual representation of this element, including a textual
+   * representation of the call context.
+   */
+  final string toStringWithContext() { result = super.toStringWithContext() }
+
+  /**
+   * Holds if this element is at the specified location.
+   * The location spans column `startcolumn` of line `startline` to
+   * column `endcolumn` of line `endline` in file `filepath`.
+   * For more information, see
+   * [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
+   */
+  final predicate hasLocationInfo(
+    string filepath, int startline, int startcolumn, int endline, int endcolumn
+  ) {
+    super.hasLocationInfo(filepath, startline, startcolumn, endline, endcolumn)
+  }
+
+  /** Gets the underlying `Node`. */
+  final Node getNode() { result = super.getNode() }
+
+  /** Gets the `FlowState` of this node. */
+  final FlowState getState() { result = getState(super.getState()) }
+
+  /** Gets the associated configuration. */
+  final Configuration getConfiguration() { result = getConfig(super.getState()) }
+
+  /** Gets a successor of this node, if any. */
+  final PathNode getASuccessor() { result = super.getASuccessor() }
+
+  /** Holds if this node is a source. */
+  final predicate isSource() { super.isSource() }
+
+  /** Holds if this node is a grouping of source nodes. */
+  final predicate isSourceGroup(string group) { super.isSourceGroup(group) }
+
+  /** Holds if this node is a grouping of sink nodes. */
+  final predicate isSinkGroup(string group) { super.isSinkGroup(group) }
+}
+
+private predicate hasFlow(Node source, Node sink, Configuration config) {
+  exists(PathNode source0, PathNode sink0 |
+    hasFlowPath(source0, sink0, config) and
+    source0.getNode() = source and
+    sink0.getNode() = sink
+  )
+}
+
+private predicate hasFlowPath(PathNode source, PathNode sink, Configuration config) {
+  hasFlowPath(source, sink) and source.getConfiguration() = config
+}
+
+private predicate hasFlowTo(Node sink, Configuration config) { hasFlow(_, sink, config) }
+
+predicate flowsTo = hasFlow/3;

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

@@ -3,15 +3,18 @@ private import DataFlowImplSpecific::Public
 import Cached
 
 module DataFlowImplCommonPublic {
-  /** A state value to track during data flow. */
-  class FlowState = string;
+  /** Provides `FlowState = string`. */
+  module FlowStateString {
+    /** A state value to track during data flow. */
+    class FlowState = string;
 
-  /**
-   * The default state, which is used when the state is unspecified for a source
-   * or a sink.
-   */
-  class FlowStateEmpty extends FlowState {
-    FlowStateEmpty() { this = "" }
+    /**
+     * The default state, which is used when the state is unspecified for a source
+     * or a sink.
+     */
+    class FlowStateEmpty extends FlowState {
+      FlowStateEmpty() { this = "" }
+    }
   }
 
   private newtype TFlowFeature =
@@ -707,8 +710,8 @@ private module Cached {
      * Gets a viable dispatch target of `call` in the context `ctx`. This is
      * restricted to those `call`s for which a context might make a difference.
      */
-    pragma[nomagic]
-    private DataFlowCallable viableImplInCallContextExt(DataFlowCall call, DataFlowCall ctx) {
+    cached
+    DataFlowCallable viableImplInCallContextExt(DataFlowCall call, DataFlowCall ctx) {
       result = viableImplInCallContext(call, ctx) and
       result = viableCallable(call)
       or
@@ -1391,6 +1394,9 @@ class TypedContentApprox extends MkTypedContentApprox {
   /** Gets a typed content approximated by this value. */
   TypedContent getATypedContent() { result = getATypedContent(this) }
 
+  /** Gets the content. */
+  ContentApprox getContent() { result = c }
+
   /** Gets the container type. */
   DataFlowType getContainerType() { result = t }
 
@@ -1408,6 +1414,8 @@ abstract class ApproxAccessPathFront extends TApproxAccessPathFront {
 
   abstract boolean toBoolNonEmpty();
 
+  TypedContentApprox getHead() { this = TApproxFrontHead(result) }
+
   pragma[nomagic]
   TypedContent getAHead() {
     exists(TypedContentApprox cont |

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

@@ -565,3 +565,12 @@ private class MyConsistencyConfiguration extends Consistency::ConsistencyConfigu
     any()
   }
 }
+
+/**
+ * Gets an additional term that is added to the `join` and `branch` computations to reflect
+ * an additional forward or backwards branching factor that is not taken into account
+ * when calculating the (virtual) dispatch cost.
+ *
+ * Argument `arg` is part of a path from a source to a sink, and `p` is the target parameter.
+ */
+int getAdditionalFlowIntoCallNodeTerm(ArgumentNode arg, ParameterNode p) { none() }

cpp/ql/lib/experimental/semmle/code/cpp/ir/dataflow/internal/PrintIRLocalFlow.qll

@@ -97,23 +97,23 @@ private string getNodeProperty(DataFlow::Node node, string key) {
     |
       kind, ", "
     )
-  or
-  // Is there partial flow from a source to this node?
-  // This property will only be emitted if partial flow is enabled by overriding
-  // `DataFlow::Configuration::explorationLimit()`.
-  key = "pflow" and
-  result =
-    strictconcat(DataFlow::PartialPathNode sourceNode, DataFlow::PartialPathNode destNode, int dist,
-      int order1, int order2 |
-      any(DataFlow::Configuration cfg).hasPartialFlow(sourceNode, destNode, dist) and
-      destNode.getNode() = node and
-      // Only print flow from a source in the same function.
-      sourceNode.getNode().getEnclosingCallable() = node.getEnclosingCallable()
-    |
-      nodeId(sourceNode.getNode(), order1, order2) + "+" + dist.toString(), ", "
-      order by
-        order1, order2, dist desc
-    )
+  // or
+  // // Is there partial flow from a source to this node?
+  // // This property will only be emitted if partial flow is enabled by overriding
+  // // `DataFlow::Configuration::explorationLimit()`.
+  // key = "pflow" and
+  // result =
+  //   strictconcat(DataFlow::PartialPathNode sourceNode, DataFlow::PartialPathNode destNode, int dist,
+  //     int order1, int order2 |
+  //     any(DataFlow::Configuration cfg).hasPartialFlow(sourceNode, destNode, dist) and
+  //     destNode.getNode() = node and
+  //     // Only print flow from a source in the same function.
+  //     sourceNode.getNode().getEnclosingCallable() = node.getEnclosingCallable()
+  //   |
+  //     nodeId(sourceNode.getNode(), order1, order2) + "+" + dist.toString(), ", "
+  //     order by
+  //       order1, order2, dist desc
+  //   )
 }
 
 /**

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

@@ -0,0 +1,63 @@
+/**
+ * Provides classes for performing local (intra-procedural) and
+ * global (inter-procedural) taint-tracking analyses.
+ */
+
+import TaintTrackingParameter::Public
+private import TaintTrackingParameter::Private
+
+private module AddTaintDefaults<DataFlowInternal::FullStateConfigSig Config> implements
+DataFlowInternal::FullStateConfigSig {
+  import Config
+
+  predicate isBarrier(DataFlow::Node node) {
+    Config::isBarrier(node) or defaultTaintSanitizer(node)
+  }
+
+  predicate isAdditionalFlowStep(DataFlow::Node node1, DataFlow::Node node2) {
+    Config::isAdditionalFlowStep(node1, node2) or
+    defaultAdditionalTaintStep(node1, node2)
+  }
+
+  predicate allowImplicitRead(DataFlow::Node node, DataFlow::ContentSet c) {
+    Config::allowImplicitRead(node, c)
+    or
+    (
+      Config::isSink(node, _) or
+      Config::isAdditionalFlowStep(node, _) or
+      Config::isAdditionalFlowStep(node, _, _, _)
+    ) and
+    defaultImplicitTaintRead(node, c)
+  }
+}
+
+/**
+ * Constructs a standard taint tracking computation.
+ */
+module Make<DataFlow::ConfigSig Config> implements DataFlow::DataFlowSig {
+  private module Config0 implements DataFlowInternal::FullStateConfigSig {
+    import DataFlowInternal::DefaultState<Config>
+    import Config
+  }
+
+  private module C implements DataFlowInternal::FullStateConfigSig {
+    import AddTaintDefaults<Config0>
+  }
+
+  import DataFlowInternal::Impl<C>
+}
+
+/**
+ * Constructs a taint tracking computation using flow state.
+ */
+module MakeWithState<DataFlow::StateConfigSig Config> implements DataFlow::DataFlowSig {
+  private module Config0 implements DataFlowInternal::FullStateConfigSig {
+    import Config
+  }
+
+  private module C implements DataFlowInternal::FullStateConfigSig {
+    import AddTaintDefaults<Config0>
+  }
+
+  import DataFlowInternal::Impl<C>
+}

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

@@ -2,4 +2,5 @@ import experimental.semmle.code.cpp.ir.dataflow.internal.TaintTrackingUtil as Pu
 
 module Private {
   import experimental.semmle.code.cpp.ir.dataflow.DataFlow::DataFlow as DataFlow
+  import experimental.semmle.code.cpp.ir.dataflow.internal.DataFlowImpl as DataFlowInternal
 }

cpp/ql/lib/experimental/semmle/code/cpp/rangeanalysis/extensions/ConstantShiftExprRange.qll

@@ -0,0 +1,263 @@
+private import cpp
+private import experimental.semmle.code.cpp.models.interfaces.SimpleRangeAnalysisExpr
+private import semmle.code.cpp.rangeanalysis.RangeAnalysisUtils
+
+float evaluateConstantExpr(Expr e) {
+  result = e.getValue().toFloat()
+  or
+  // This handles when a constant value is put into a variable
+  // and the variable is used later
+  exists(SsaDefinition defn, StackVariable sv |
+    defn.getAUse(sv) = e and
+    result = defn.getDefiningValue(sv).getValue().toFloat()
+  )
+}
+
+// If the constant right operand is negative or is greater than or equal to the number of
+// bits in the left operands type, then the result is undefined (except on the IA-32
+// architecture where the shift value is masked with 0b00011111, but we can't
+// assume the architecture).
+bindingset[val]
+private predicate isValidShiftExprShift(float val, Expr l) {
+  val >= 0 and
+  // We use getFullyConverted because the spec says to use the *promoted* left operand
+  val < (l.getFullyConverted().getUnderlyingType().getSize() * 8)
+}
+
+bindingset[val, shift, max_val]
+private predicate canLShiftOverflow(int val, int shift, int max_val) {
+  // val << shift = val * 2^shift > max_val => val > max_val/2^shift = max_val >> b
+  val > max_val.bitShiftRight(shift)
+}
+
+/**
+ * A range analysis expression consisting of the `>>` or `>>=` operator when at least
+ * one operand is a constant (and if the right operand is a constant, it must be "valid"
+ * (see `isValidShiftExprShift`)). When handling any undefined behavior, it leaves the
+ * values unconstrained. From the C++ standard: "The behavior is undefined if the right
+ * operand is negative, or greater than or equal to the length in bits of the promoted
+ * left operand. The value of E1 >> E2 is E1 right-shifted E2 bit positions. If E1 has an
+ * unsigned type or if E1 has a signed type and a non-negative value, the value of the
+ * result is the integral part of the quotient of E1/2^E2. If E1 has a signed type and a
+ * negative value, the resulting value is implementation-defined."
+ */
+class ConstantRShiftExprRange extends SimpleRangeAnalysisExpr {
+  /**
+   * Holds for `a >> b` or `a >>= b` in one of the following two cases:
+   * 1. `a` is a constant and `b` is not
+   * 2. `b` is constant
+   *
+   * We don't handle the case where `a` and `b` are both non-constant values.
+   */
+  ConstantRShiftExprRange() {
+    getUnspecifiedType() instanceof IntegralType and
+    exists(Expr l, Expr r |
+      l = this.(RShiftExpr).getLeftOperand() and
+      r = this.(RShiftExpr).getRightOperand()
+      or
+      l = this.(AssignRShiftExpr).getLValue() and
+      r = this.(AssignRShiftExpr).getRValue()
+    |
+      l.getUnspecifiedType() instanceof IntegralType and
+      r.getUnspecifiedType() instanceof IntegralType and
+      (
+        // If the left operand is a constant, verify that the right operand is not a constant
+        exists(evaluateConstantExpr(l)) and not exists(evaluateConstantExpr(r))
+        or
+        // If the right operand is a constant, check if it is a valid shift expression
+        exists(float constROp |
+          constROp = evaluateConstantExpr(r) and isValidShiftExprShift(constROp, l)
+        )
+      )
+    )
+  }
+
+  Expr getLeftOperand() {
+    result = this.(RShiftExpr).getLeftOperand() or
+    result = this.(AssignRShiftExpr).getLValue()
+  }
+
+  Expr getRightOperand() {
+    result = this.(RShiftExpr).getRightOperand() or
+    result = this.(AssignRShiftExpr).getRValue()
+  }
+
+  override float getLowerBounds() {
+    exists(int lLower, int lUpper, int rLower, int rUpper |
+      lLower = getFullyConvertedLowerBounds(getLeftOperand()) and
+      lUpper = getFullyConvertedUpperBounds(getLeftOperand()) and
+      rLower = getFullyConvertedLowerBounds(getRightOperand()) and
+      rUpper = getFullyConvertedUpperBounds(getRightOperand()) and
+      lLower <= lUpper and
+      rLower <= rUpper
+    |
+      if
+        lLower < 0
+        or
+        not (
+          isValidShiftExprShift(rLower, getLeftOperand()) and
+          isValidShiftExprShift(rUpper, getLeftOperand())
+        )
+      then
+        // We don't want to deal with shifting negative numbers at the moment,
+        // and a negative shift is implementation defined, so we set the result
+        // to the minimum value
+        result = exprMinVal(this)
+      else
+        // We can get the smallest value by shifting the smallest bound by the largest bound
+        result = lLower.bitShiftRight(rUpper)
+    )
+  }
+
+  override float getUpperBounds() {
+    exists(int lLower, int lUpper, int rLower, int rUpper |
+      lLower = getFullyConvertedLowerBounds(getLeftOperand()) and
+      lUpper = getFullyConvertedUpperBounds(getLeftOperand()) and
+      rLower = getFullyConvertedLowerBounds(getRightOperand()) and
+      rUpper = getFullyConvertedUpperBounds(getRightOperand()) and
+      lLower <= lUpper and
+      rLower <= rUpper
+    |
+      if
+        lLower < 0
+        or
+        not (
+          isValidShiftExprShift(rLower, getLeftOperand()) and
+          isValidShiftExprShift(rUpper, getLeftOperand())
+        )
+      then
+        // We don't want to deal with shifting negative numbers at the moment,
+        // and a negative shift is implementation defined, so we set the result
+        // to the maximum value
+        result = exprMaxVal(this)
+      else
+        // We can get the largest value by shifting the largest bound by the smallest bound
+        result = lUpper.bitShiftRight(rLower)
+    )
+  }
+
+  override predicate dependsOnChild(Expr child) {
+    child = getLeftOperand() or child = getRightOperand()
+  }
+}
+
+/**
+ * A range analysis expression consisting of the `<<` or `<<=` operator when at least
+ * one operand is a constant (and if the right operand is a constant, it must be "valid"
+ * (see `isValidShiftExprShift`)). When handling any undefined behavior, it leaves the
+ * values unconstrained. From the C++ standard: "The behavior is undefined if the right
+ * operand is negative, or greater than or equal to the length in bits of the promoted left operand.
+ * The value of E1 << E2 is E1 left-shifted E2 bit positions; vacated bits are zero-filled. If E1
+ * has an unsigned type, the value of the result is E1 x 2 E2, reduced modulo one more than the
+ * maximum value representable in the result type. Otherwise, if E1 has a signed type and
+ * non-negative value, and E1 x 2 E2 is representable in the corresponding unsigned type of the
+ * result type, then that value, converted to the result type, is the resulting value; otherwise,
+ * the behavior is undefined."
+ */
+class ConstantLShiftExprRange extends SimpleRangeAnalysisExpr {
+  /**
+   * Holds for `a << b` or `a <<= b` in one of the following two cases:
+   * 1. `a` is a constant and `b` is not
+   * 2. `b` is constant
+   *
+   * We don't handle the case where `a` and `b` are both non-constant values.
+   */
+  ConstantLShiftExprRange() {
+    getUnspecifiedType() instanceof IntegralType and
+    exists(Expr l, Expr r |
+      l = this.(LShiftExpr).getLeftOperand() and
+      r = this.(LShiftExpr).getRightOperand()
+      or
+      l = this.(AssignLShiftExpr).getLValue() and
+      r = this.(AssignLShiftExpr).getRValue()
+    |
+      l.getUnspecifiedType() instanceof IntegralType and
+      r.getUnspecifiedType() instanceof IntegralType and
+      (
+        // If the left operand is a constant, verify that the right operand is not a constant
+        exists(evaluateConstantExpr(l)) and not exists(evaluateConstantExpr(r))
+        or
+        // If the right operand is a constant, check if it is a valid shift expression
+        exists(float constROp |
+          constROp = evaluateConstantExpr(r) and isValidShiftExprShift(constROp, l)
+        )
+      )
+    )
+  }
+
+  Expr getLeftOperand() {
+    result = this.(LShiftExpr).getLeftOperand() or
+    result = this.(AssignLShiftExpr).getLValue()
+  }
+
+  Expr getRightOperand() {
+    result = this.(LShiftExpr).getRightOperand() or
+    result = this.(AssignLShiftExpr).getRValue()
+  }
+
+  override float getLowerBounds() {
+    exists(int lLower, int lUpper, int rLower, int rUpper |
+      lLower = getFullyConvertedLowerBounds(getLeftOperand()) and
+      lUpper = getFullyConvertedUpperBounds(getLeftOperand()) and
+      rLower = getFullyConvertedLowerBounds(getRightOperand()) and
+      rUpper = getFullyConvertedUpperBounds(getRightOperand()) and
+      lLower <= lUpper and
+      rLower <= rUpper
+    |
+      if
+        lLower < 0
+        or
+        not (
+          isValidShiftExprShift(rLower, getLeftOperand()) and
+          isValidShiftExprShift(rUpper, getLeftOperand())
+        )
+      then
+        // We don't want to deal with shifting negative numbers at the moment,
+        // and a negative shift is undefined, so we set to the minimum value
+        result = exprMinVal(this)
+      else
+        // If we have `0b01010000 << [0, 2]`, the max value for 8 bits is 0b10100000
+        // (a shift of 1) but doing a shift by the upper bound would give 0b01000000.
+        // So if the left shift operation causes an overflow, we just assume the max value
+        // If necessary, we may be able to improve this bound in the future
+        if canLShiftOverflow(lUpper, rUpper, exprMaxVal(this))
+        then result = exprMinVal(this)
+        else result = lLower.bitShiftLeft(rLower)
+    )
+  }
+
+  override float getUpperBounds() {
+    exists(int lLower, int lUpper, int rLower, int rUpper |
+      lLower = getFullyConvertedLowerBounds(getLeftOperand()) and
+      lUpper = getFullyConvertedUpperBounds(getLeftOperand()) and
+      rLower = getFullyConvertedLowerBounds(getRightOperand()) and
+      rUpper = getFullyConvertedUpperBounds(getRightOperand()) and
+      lLower <= lUpper and
+      rLower <= rUpper
+    |
+      if
+        lLower < 0
+        or
+        not (
+          isValidShiftExprShift(rLower, getLeftOperand()) and
+          isValidShiftExprShift(rUpper, getLeftOperand())
+        )
+      then
+        // We don't want to deal with shifting negative numbers at the moment,
+        // and a negative shift is undefined, so we set it to the maximum value
+        result = exprMaxVal(this)
+      else
+        // If we have `0b01010000 << [0, 2]`, the max value for 8 bits is 0b10100000
+        // (a shift of 1) but doing a shift by the upper bound would give 0b01000000.
+        // So if the left shift operation causes an overflow, we just assume the max value
+        // If necessary, we may be able to improve this bound in the future
+        if canLShiftOverflow(lUpper, rUpper, exprMaxVal(this))
+        then result = exprMaxVal(this)
+        else result = lUpper.bitShiftLeft(rUpper)
+    )
+  }
+
+  override predicate dependsOnChild(Expr child) {
+    child = getLeftOperand() or child = getRightOperand()
+  }
+}

cpp/ql/lib/qlpack.yml

@@ -1,5 +1,5 @@
 name: codeql/cpp-all
-version: 0.5.1
+version: 0.5.4-dev
 groups: cpp
 dbscheme: semmlecode.cpp.dbscheme
 extractor: cpp

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

@@ -68,7 +68,9 @@ class Declaration extends Locatable, @declaration {
    * Holds if this declaration has the fully-qualified name `qualifiedName`.
    * See `getQualifiedName`.
    */
-  predicate hasQualifiedName(string qualifiedName) { this.getQualifiedName() = qualifiedName }
+  deprecated predicate hasQualifiedName(string qualifiedName) {
+    this.getQualifiedName() = qualifiedName
+  }
 
   /**
    * Holds if this declaration has a fully-qualified name with a name-space
@@ -184,7 +186,7 @@ class Declaration extends Locatable, @declaration {
   predicate hasDefinition() { exists(this.getDefinition()) }
 
   /** DEPRECATED: Use `hasDefinition` instead. */
-  predicate isDefined() { this.hasDefinition() }
+  deprecated predicate isDefined() { this.hasDefinition() }
 
   /** Gets the preferred location of this declaration, if any. */
   override Location getLocation() { none() }

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

@@ -41,7 +41,7 @@ class Function extends Declaration, ControlFlowNode, AccessHolder, @function {
    * `min<int>(int, int) -> int`, and the full signature of the uninstantiated
    * template on the first line would be `min<T>(T, T) -> T`.
    */
-  string getFullSignature() {
+  deprecated string getFullSignature() {
     exists(string name, string templateArgs, string args |
       result = name + templateArgs + args + " -> " + this.getType().toString() and
       name = this.getQualifiedName() and

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

@@ -12,7 +12,9 @@ predicate freeFunction(Function f, int argNum) { argNum = f.(DeallocationFunctio
  *
  * DEPRECATED: Use `DeallocationExpr` instead (this also includes `delete` expressions).
  */
-predicate freeCall(FunctionCall fc, Expr arg) { arg = fc.(DeallocationExpr).getFreedExpr() }
+deprecated predicate freeCall(FunctionCall fc, Expr arg) {
+  arg = fc.(DeallocationExpr).getFreedExpr()
+}
 
 /**
  * Is e some kind of allocation or deallocation (`new`, `alloc`, `realloc`, `delete`, `free` etc)?

cpp/ql/lib/semmle/code/cpp/dataflow/DataFlow.qll

@@ -20,5 +20,6 @@
 import cpp
 
 module DataFlow {
-  import semmle.code.cpp.dataflow.internal.DataFlowImpl
+  import semmle.code.cpp.dataflow.internal.DataFlow
+  import semmle.code.cpp.dataflow.internal.DataFlowImpl1
 }

cpp/ql/lib/semmle/code/cpp/dataflow/TaintTracking.qll

@@ -19,5 +19,6 @@ import semmle.code.cpp.dataflow.DataFlow
 import semmle.code.cpp.dataflow.DataFlow2
 
 module TaintTracking {
+  import semmle.code.cpp.dataflow.internal.tainttracking1.TaintTracking
   import semmle.code.cpp.dataflow.internal.tainttracking1.TaintTrackingImpl
 }

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

@@ -0,0 +1,245 @@
+/**
+ * Provides an implementation of global (interprocedural) data flow. This file
+ * re-exports the local (intraprocedural) data flow analysis from
+ * `DataFlowImplSpecific::Public` and adds a global analysis, mainly exposed
+ * through the `Make` and `MakeWithState` modules.
+ */
+
+private import DataFlowImplCommon
+private import DataFlowImplSpecific::Private
+import DataFlowImplSpecific::Public
+import DataFlowImplCommonPublic
+private import DataFlowImpl
+
+/** An input configuration for data flow. */
+signature module ConfigSig {
+  /**
+   * Holds if `source` is a relevant data flow source.
+   */
+  predicate isSource(Node source);
+
+  /**
+   * Holds if `sink` is a relevant data flow sink.
+   */
+  predicate isSink(Node sink);
+
+  /**
+   * Holds if data flow through `node` is prohibited. This completely removes
+   * `node` from the data flow graph.
+   */
+  default predicate isBarrier(Node node) { none() }
+
+  /** Holds if data flow into `node` is prohibited. */
+  default predicate isBarrierIn(Node node) { none() }
+
+  /** Holds if data flow out of `node` is prohibited. */
+  default predicate isBarrierOut(Node node) { none() }
+
+  /**
+   * Holds if data may flow from `node1` to `node2` in addition to the normal data-flow steps.
+   */
+  default predicate isAdditionalFlowStep(Node node1, Node node2) { none() }
+
+  /**
+   * Holds if an arbitrary number of implicit read steps of content `c` may be
+   * taken at `node`.
+   */
+  default predicate allowImplicitRead(Node node, ContentSet c) { none() }
+
+  /**
+   * Gets the virtual dispatch branching limit when calculating field flow.
+   * This can be overridden to a smaller value to improve performance (a
+   * value of 0 disables field flow), or a larger value to get more results.
+   */
+  default int fieldFlowBranchLimit() { result = 2 }
+
+  /**
+   * Gets a data flow configuration feature to add restrictions to the set of
+   * valid flow paths.
+   *
+   * - `FeatureHasSourceCallContext`:
+   *    Assume that sources have some existing call context to disallow
+   *    conflicting return-flow directly following the source.
+   * - `FeatureHasSinkCallContext`:
+   *    Assume that sinks have some existing call context to disallow
+   *    conflicting argument-to-parameter flow directly preceding the sink.
+   * - `FeatureEqualSourceSinkCallContext`:
+   *    Implies both of the above and additionally ensures that the entire flow
+   *    path preserves the call context.
+   *
+   * These features are generally not relevant for typical end-to-end data flow
+   * queries, but should only be used for constructing paths that need to
+   * somehow be pluggable in another path context.
+   */
+  default FlowFeature getAFeature() { none() }
+
+  /** Holds if sources should be grouped in the result of `hasFlowPath`. */
+  default predicate sourceGrouping(Node source, string sourceGroup) { none() }
+
+  /** Holds if sinks should be grouped in the result of `hasFlowPath`. */
+  default predicate sinkGrouping(Node sink, string sinkGroup) { none() }
+
+  /**
+   * Holds if hidden nodes should be included in the data flow graph.
+   *
+   * This feature should only be used for debugging or when the data flow graph
+   * is not visualized (as it is in a `path-problem` query).
+   */
+  default predicate includeHiddenNodes() { none() }
+}
+
+/** An input configuration for data flow using flow state. */
+signature module StateConfigSig {
+  bindingset[this]
+  class FlowState;
+
+  /**
+   * Holds if `source` is a relevant data flow source with the given initial
+   * `state`.
+   */
+  predicate isSource(Node source, FlowState state);
+
+  /**
+   * Holds if `sink` is a relevant data flow sink accepting `state`.
+   */
+  predicate isSink(Node sink, FlowState state);
+
+  /**
+   * Holds if data flow through `node` is prohibited. This completely removes
+   * `node` from the data flow graph.
+   */
+  default predicate isBarrier(Node node) { none() }
+
+  /**
+   * Holds if data flow through `node` is prohibited when the flow state is
+   * `state`.
+   */
+  predicate isBarrier(Node node, FlowState state);
+
+  /** Holds if data flow into `node` is prohibited. */
+  default predicate isBarrierIn(Node node) { none() }
+
+  /** Holds if data flow out of `node` is prohibited. */
+  default predicate isBarrierOut(Node node) { none() }
+
+  /**
+   * Holds if data may flow from `node1` to `node2` in addition to the normal data-flow steps.
+   */
+  default predicate isAdditionalFlowStep(Node node1, Node node2) { none() }
+
+  /**
+   * Holds if data may flow from `node1` to `node2` in addition to the normal data-flow steps.
+   * This step is only applicable in `state1` and updates the flow state to `state2`.
+   */
+  predicate isAdditionalFlowStep(Node node1, FlowState state1, Node node2, FlowState state2);
+
+  /**
+   * Holds if an arbitrary number of implicit read steps of content `c` may be
+   * taken at `node`.
+   */
+  default predicate allowImplicitRead(Node node, ContentSet c) { none() }
+
+  /**
+   * Gets the virtual dispatch branching limit when calculating field flow.
+   * This can be overridden to a smaller value to improve performance (a
+   * value of 0 disables field flow), or a larger value to get more results.
+   */
+  default int fieldFlowBranchLimit() { result = 2 }
+
+  /**
+   * Gets a data flow configuration feature to add restrictions to the set of
+   * valid flow paths.
+   *
+   * - `FeatureHasSourceCallContext`:
+   *    Assume that sources have some existing call context to disallow
+   *    conflicting return-flow directly following the source.
+   * - `FeatureHasSinkCallContext`:
+   *    Assume that sinks have some existing call context to disallow
+   *    conflicting argument-to-parameter flow directly preceding the sink.
+   * - `FeatureEqualSourceSinkCallContext`:
+   *    Implies both of the above and additionally ensures that the entire flow
+   *    path preserves the call context.
+   *
+   * These features are generally not relevant for typical end-to-end data flow
+   * queries, but should only be used for constructing paths that need to
+   * somehow be pluggable in another path context.
+   */
+  default FlowFeature getAFeature() { none() }
+
+  /** Holds if sources should be grouped in the result of `hasFlowPath`. */
+  default predicate sourceGrouping(Node source, string sourceGroup) { none() }
+
+  /** Holds if sinks should be grouped in the result of `hasFlowPath`. */
+  default predicate sinkGrouping(Node sink, string sinkGroup) { none() }
+
+  /**
+   * Holds if hidden nodes should be included in the data flow graph.
+   *
+   * This feature should only be used for debugging or when the data flow graph
+   * is not visualized (as it is in a `path-problem` query).
+   */
+  default predicate includeHiddenNodes() { none() }
+}
+
+/**
+ * Gets the exploration limit for `hasPartialFlow` and `hasPartialFlowRev`
+ * measured in approximate number of interprocedural steps.
+ */
+signature int explorationLimitSig();
+
+/**
+ * The output of a data flow computation.
+ */
+signature module DataFlowSig {
+  /**
+   * A `Node` augmented with a call context (except for sinks) and an access path.
+   * Only those `PathNode`s that are reachable from a source, and which can reach a sink, are generated.
+   */
+  class PathNode;
+
+  /**
+   * Holds if data can flow from `source` to `sink`.
+   *
+   * The corresponding paths are generated from the end-points and the graph
+   * included in the module `PathGraph`.
+   */
+  predicate hasFlowPath(PathNode source, PathNode sink);
+
+  /**
+   * Holds if data can flow from `source` to `sink`.
+   */
+  predicate hasFlow(Node source, Node sink);
+
+  /**
+   * Holds if data can flow from some source to `sink`.
+   */
+  predicate hasFlowTo(Node sink);
+
+  /**
+   * Holds if data can flow from some source to `sink`.
+   */
+  predicate hasFlowToExpr(DataFlowExpr sink);
+}
+
+/**
+ * Constructs a standard data flow computation.
+ */
+module Make<ConfigSig Config> implements DataFlowSig {
+  private module C implements FullStateConfigSig {
+    import DefaultState<Config>
+    import Config
+  }
+
+  import Impl<C>
+}
+
+/**
+ * Constructs a data flow computation using flow state.
+ */
+module MakeWithState<StateConfigSig Config> implements DataFlowSig {
+  private module C implements FullStateConfigSig {
+    import Config
+  }
+
+  import Impl<C>
+}

cpp/ql/lib/semmle/code/cpp/dataflow/internal/DataFlowImpl1.qll

@@ -0,0 +1,396 @@
+/**
+ * DEPRECATED: Use `Make` and `MakeWithState` instead.
+ *
+ * Provides a `Configuration` class backwards-compatible interface to the data
+ * flow library.
+ */
+
+private import DataFlowImplCommon
+private import DataFlowImplSpecific::Private
+import DataFlowImplSpecific::Public
+private import DataFlowImpl
+import DataFlowImplCommonPublic
+import FlowStateString
+
+/**
+ * A configuration of interprocedural data flow analysis. This defines
+ * sources, sinks, and any other configurable aspect of the analysis. Each
+ * use of the global data flow library must define its own unique extension
+ * of this abstract class. To create a configuration, extend this class with
+ * a subclass whose characteristic predicate is a unique singleton string.
+ * For example, write
+ *
+ * ```ql
+ * class MyAnalysisConfiguration extends DataFlow::Configuration {
+ *   MyAnalysisConfiguration() { this = "MyAnalysisConfiguration" }
+ *   // Override `isSource` and `isSink`.
+ *   // Optionally override `isBarrier`.
+ *   // Optionally override `isAdditionalFlowStep`.
+ * }
+ * ```
+ * Conceptually, this defines a graph where the nodes are `DataFlow::Node`s and
+ * the edges are those data-flow steps that preserve the value of the node
+ * along with any additional edges defined by `isAdditionalFlowStep`.
+ * Specifying nodes in `isBarrier` will remove those nodes from the graph, and
+ * specifying nodes in `isBarrierIn` and/or `isBarrierOut` will remove in-going
+ * and/or out-going edges from those nodes, respectively.
+ *
+ * Then, to query whether there is flow between some `source` and `sink`,
+ * write
+ *
+ * ```ql
+ * exists(MyAnalysisConfiguration cfg | cfg.hasFlow(source, sink))
+ * ```
+ *
+ * Multiple configurations can coexist, but two classes extending
+ * `DataFlow::Configuration` should never depend on each other. One of them
+ * should instead depend on a `DataFlow2::Configuration`, a
+ * `DataFlow3::Configuration`, or a `DataFlow4::Configuration`.
+ */
+abstract class Configuration extends string {
+  bindingset[this]
+  Configuration() { any() }
+
+  /**
+   * Holds if `source` is a relevant data flow source.
+   */
+  predicate isSource(Node source) { none() }
+
+  /**
+   * Holds if `source` is a relevant data flow source with the given initial
+   * `state`.
+   */
+  predicate isSource(Node source, FlowState state) { none() }
+
+  /**
+   * Holds if `sink` is a relevant data flow sink.
+   */
+  predicate isSink(Node sink) { none() }
+
+  /**
+   * Holds if `sink` is a relevant data flow sink accepting `state`.
+   */
+  predicate isSink(Node sink, FlowState state) { none() }
+
+  /**
+   * Holds if data flow through `node` is prohibited. This completely removes
+   * `node` from the data flow graph.
+   */
+  predicate isBarrier(Node node) { none() }
+
+  /**
+   * Holds if data flow through `node` is prohibited when the flow state is
+   * `state`.
+   */
+  predicate isBarrier(Node node, FlowState state) { none() }
+
+  /** Holds if data flow into `node` is prohibited. */
+  predicate isBarrierIn(Node node) { none() }
+
+  /** Holds if data flow out of `node` is prohibited. */
+  predicate isBarrierOut(Node node) { none() }
+
+  /**
+   * DEPRECATED: Use `isBarrier` and `BarrierGuard` module instead.
+   *
+   * Holds if data flow through nodes guarded by `guard` is prohibited.
+   */
+  deprecated predicate isBarrierGuard(BarrierGuard guard) { none() }
+
+  /**
+   * DEPRECATED: Use `isBarrier` and `BarrierGuard` module instead.
+   *
+   * Holds if data flow through nodes guarded by `guard` is prohibited when
+   * the flow state is `state`
+   */
+  deprecated predicate isBarrierGuard(BarrierGuard guard, FlowState state) { none() }
+
+  /**
+   * Holds if data may flow from `node1` to `node2` in addition to the normal data-flow steps.
+   */
+  predicate isAdditionalFlowStep(Node node1, Node node2) { none() }
+
+  /**
+   * Holds if data may flow from `node1` to `node2` in addition to the normal data-flow steps.
+   * This step is only applicable in `state1` and updates the flow state to `state2`.
+   */
+  predicate isAdditionalFlowStep(Node node1, FlowState state1, Node node2, FlowState state2) {
+    none()
+  }
+
+  /**
+   * Holds if an arbitrary number of implicit read steps of content `c` may be
+   * taken at `node`.
+   */
+  predicate allowImplicitRead(Node node, ContentSet c) { none() }
+
+  /**
+   * Gets the virtual dispatch branching limit when calculating field flow.
+   * This can be overridden to a smaller value to improve performance (a
+   * value of 0 disables field flow), or a larger value to get more results.
+   */
+  int fieldFlowBranchLimit() { result = 2 }
+
+  /**
+   * Gets a data flow configuration feature to add restrictions to the set of
+   * valid flow paths.
+   *
+   * - `FeatureHasSourceCallContext`:
+   *    Assume that sources have some existing call context to disallow
+   *    conflicting return-flow directly following the source.
+   * - `FeatureHasSinkCallContext`:
+   *    Assume that sinks have some existing call context to disallow
+   *    conflicting argument-to-parameter flow directly preceding the sink.
+   * - `FeatureEqualSourceSinkCallContext`:
+   *    Implies both of the above and additionally ensures that the entire flow
+   *    path preserves the call context.
+   *
+   * These features are generally not relevant for typical end-to-end data flow
+   * queries, but should only be used for constructing paths that need to
+   * somehow be pluggable in another path context.
+   */
+  FlowFeature getAFeature() { none() }
+
+  /** Holds if sources should be grouped in the result of `hasFlowPath`. */
+  predicate sourceGrouping(Node source, string sourceGroup) { none() }
+
+  /** Holds if sinks should be grouped in the result of `hasFlowPath`. */
+  predicate sinkGrouping(Node sink, string sinkGroup) { none() }
+
+  /**
+   * Holds if data may flow from `source` to `sink` for this configuration.
+   */
+  predicate hasFlow(Node source, Node sink) { hasFlow(source, sink, this) }
+
+  /**
+   * Holds if data may flow from `source` to `sink` for this configuration.
+   *
+   * The corresponding paths are generated from the end-points and the graph
+   * included in the module `PathGraph`.
+   */
+  predicate hasFlowPath(PathNode source, PathNode sink) { hasFlowPath(source, sink, this) }
+
+  /**
+   * Holds if data may flow from some source to `sink` for this configuration.
+   */
+  predicate hasFlowTo(Node sink) { hasFlowTo(sink, this) }
+
+  /**
+   * Holds if data may flow from some source to `sink` for this configuration.
+   */
+  predicate hasFlowToExpr(DataFlowExpr sink) { this.hasFlowTo(exprNode(sink)) }
+
+  /**
+   * DEPRECATED: Use `FlowExploration<explorationLimit>` instead.
+   *
+   * Gets the exploration limit for `hasPartialFlow` and `hasPartialFlowRev`
+   * measured in approximate number of interprocedural steps.
+   */
+  deprecated int explorationLimit() { none() }
+
+  /**
+   * Holds if hidden nodes should be included in the data flow graph.
+   *
+   * This feature should only be used for debugging or when the data flow graph
+   * is not visualized (for example in a `path-problem` query).
+   */
+  predicate includeHiddenNodes() { none() }
+}
+
+/**
+ * This class exists to prevent mutual recursion between the user-overridden
+ * member predicates of `Configuration` and the rest of the data-flow library.
+ * Good performance cannot be guaranteed in the presence of such recursion, so
+ * it should be replaced by using more than one copy of the data flow library.
+ */
+abstract private class ConfigurationRecursionPrevention extends Configuration {
+  bindingset[this]
+  ConfigurationRecursionPrevention() { any() }
+
+  override predicate hasFlow(Node source, Node sink) {
+    strictcount(Node n | this.isSource(n)) < 0
+    or
+    strictcount(Node n | this.isSource(n, _)) < 0
+    or
+    strictcount(Node n | this.isSink(n)) < 0
+    or
+    strictcount(Node n | this.isSink(n, _)) < 0
+    or
+    strictcount(Node n1, Node n2 | this.isAdditionalFlowStep(n1, n2)) < 0
+    or
+    strictcount(Node n1, Node n2 | this.isAdditionalFlowStep(n1, _, n2, _)) < 0
+    or
+    super.hasFlow(source, sink)
+  }
+}
+
+/** A bridge class to access the deprecated `isBarrierGuard`. */
+private class BarrierGuardGuardedNodeBridge extends Unit {
+  abstract predicate guardedNode(Node n, Configuration config);
+
+  abstract predicate guardedNode(Node n, FlowState state, Configuration config);
+}
+
+private class BarrierGuardGuardedNode extends BarrierGuardGuardedNodeBridge {
+  deprecated override predicate guardedNode(Node n, Configuration config) {
+    exists(BarrierGuard g |
+      config.isBarrierGuard(g) and
+      n = g.getAGuardedNode()
+    )
+  }
+
+  deprecated override predicate guardedNode(Node n, FlowState state, Configuration config) {
+    exists(BarrierGuard g |
+      config.isBarrierGuard(g, state) and
+      n = g.getAGuardedNode()
+    )
+  }
+}
+
+private FlowState relevantState(Configuration config) {
+  config.isSource(_, result) or
+  config.isSink(_, result) or
+  config.isBarrier(_, result) or
+  config.isAdditionalFlowStep(_, result, _, _) or
+  config.isAdditionalFlowStep(_, _, _, result)
+}
+
+private newtype TConfigState =
+  TMkConfigState(Configuration config, FlowState state) {
+    state = relevantState(config) or state instanceof FlowStateEmpty
+  }
+
+private Configuration getConfig(TConfigState state) { state = TMkConfigState(result, _) }
+
+private FlowState getState(TConfigState state) { state = TMkConfigState(_, result) }
+
+private predicate singleConfiguration() { 1 = strictcount(Configuration c) }
+
+private module Config implements FullStateConfigSig {
+  class FlowState = TConfigState;
+
+  predicate isSource(Node source, FlowState state) {
+    getConfig(state).isSource(source, getState(state))
+    or
+    getConfig(state).isSource(source) and getState(state) instanceof FlowStateEmpty
+  }
+
+  predicate isSink(Node sink, FlowState state) {
+    getConfig(state).isSink(sink, getState(state))
+    or
+    getConfig(state).isSink(sink) and getState(state) instanceof FlowStateEmpty
+  }
+
+  predicate isBarrier(Node node) { none() }
+
+  predicate isBarrier(Node node, FlowState state) {
+    getConfig(state).isBarrier(node, getState(state)) or
+    getConfig(state).isBarrier(node) or
+    any(BarrierGuardGuardedNodeBridge b).guardedNode(node, getState(state), getConfig(state)) or
+    any(BarrierGuardGuardedNodeBridge b).guardedNode(node, getConfig(state))
+  }
+
+  predicate isBarrierIn(Node node) { any(Configuration config).isBarrierIn(node) }
+
+  predicate isBarrierOut(Node node) { any(Configuration config).isBarrierOut(node) }
+
+  predicate isAdditionalFlowStep(Node node1, Node node2) {
+    singleConfiguration() and
+    any(Configuration config).isAdditionalFlowStep(node1, node2)
+  }
+
+  predicate isAdditionalFlowStep(Node node1, FlowState state1, Node node2, FlowState state2) {
+    getConfig(state1).isAdditionalFlowStep(node1, getState(state1), node2, getState(state2)) and
+    getConfig(state2) = getConfig(state1)
+    or
+    not singleConfiguration() and
+    getConfig(state1).isAdditionalFlowStep(node1, node2) and
+    state2 = state1
+  }
+
+  predicate allowImplicitRead(Node node, ContentSet c) {
+    any(Configuration config).allowImplicitRead(node, c)
+  }
+
+  int fieldFlowBranchLimit() { result = min(any(Configuration config).fieldFlowBranchLimit()) }
+
+  FlowFeature getAFeature() { result = any(Configuration config).getAFeature() }
+
+  predicate sourceGrouping(Node source, string sourceGroup) {
+    any(Configuration config).sourceGrouping(source, sourceGroup)
+  }
+
+  predicate sinkGrouping(Node sink, string sinkGroup) {
+    any(Configuration config).sinkGrouping(sink, sinkGroup)
+  }
+
+  predicate includeHiddenNodes() { any(Configuration config).includeHiddenNodes() }
+}
+
+private import Impl<Config> as I
+import I
+
+/**
+ * A `Node` augmented with a call context (except for sinks), an access path, and a configuration.
+ * Only those `PathNode`s that are reachable from a source, and which can reach a sink, are generated.
+ */
+class PathNode instanceof I::PathNode {
+  /** Gets a textual representation of this element. */
+  final string toString() { result = super.toString() }
+
+  /**
+   * Gets a textual representation of this element, including a textual
+   * representation of the call context.
+   */
+  final string toStringWithContext() { result = super.toStringWithContext() }
+
+  /**
+   * Holds if this element is at the specified location.
+   * The location spans column `startcolumn` of line `startline` to
+   * column `endcolumn` of line `endline` in file `filepath`.
+   * For more information, see
+   * [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
+   */
+  final predicate hasLocationInfo(
+    string filepath, int startline, int startcolumn, int endline, int endcolumn
+  ) {
+    super.hasLocationInfo(filepath, startline, startcolumn, endline, endcolumn)
+  }
+
+  /** Gets the underlying `Node`. */
+  final Node getNode() { result = super.getNode() }
+
+  /** Gets the `FlowState` of this node. */
+  final FlowState getState() { result = getState(super.getState()) }
+
+  /** Gets the associated configuration. */
+  final Configuration getConfiguration() { result = getConfig(super.getState()) }
+
+  /** Gets a successor of this node, if any. */
+  final PathNode getASuccessor() { result = super.getASuccessor() }
+
+  /** Holds if this node is a source. */
+  final predicate isSource() { super.isSource() }
+
+  /** Holds if this node is a grouping of source nodes. */
+  final predicate isSourceGroup(string group) { super.isSourceGroup(group) }
+
+  /** Holds if this node is a grouping of sink nodes. */
+  final predicate isSinkGroup(string group) { super.isSinkGroup(group) }
+}
+
+private predicate hasFlow(Node source, Node sink, Configuration config) {
+  exists(PathNode source0, PathNode sink0 |
+    hasFlowPath(source0, sink0, config) and
+    source0.getNode() = source and
+    sink0.getNode() = sink
+  )
+}
+
+private predicate hasFlowPath(PathNode source, PathNode sink, Configuration config) {
+  hasFlowPath(source, sink) and source.getConfiguration() = config
+}
+
+private predicate hasFlowTo(Node sink, Configuration config) { hasFlow(_, sink, config) }
+
+predicate flowsTo = hasFlow/3;

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

@@ -3,15 +3,18 @@ private import DataFlowImplSpecific::Public
 import Cached
 
 module DataFlowImplCommonPublic {
-  /** A state value to track during data flow. */
-  class FlowState = string;
+  /** Provides `FlowState = string`. */
+  module FlowStateString {
+    /** A state value to track during data flow. */
+    class FlowState = string;
 
-  /**
-   * The default state, which is used when the state is unspecified for a source
-   * or a sink.
-   */
-  class FlowStateEmpty extends FlowState {
-    FlowStateEmpty() { this = "" }
+    /**
+     * The default state, which is used when the state is unspecified for a source
+     * or a sink.
+     */
+    class FlowStateEmpty extends FlowState {
+      FlowStateEmpty() { this = "" }
+    }
   }
 
   private newtype TFlowFeature =
@@ -707,8 +710,8 @@ private module Cached {
      * Gets a viable dispatch target of `call` in the context `ctx`. This is
      * restricted to those `call`s for which a context might make a difference.
      */
-    pragma[nomagic]
-    private DataFlowCallable viableImplInCallContextExt(DataFlowCall call, DataFlowCall ctx) {
+    cached
+    DataFlowCallable viableImplInCallContextExt(DataFlowCall call, DataFlowCall ctx) {
       result = viableImplInCallContext(call, ctx) and
       result = viableCallable(call)
       or
@@ -1391,6 +1394,9 @@ class TypedContentApprox extends MkTypedContentApprox {
   /** Gets a typed content approximated by this value. */
   TypedContent getATypedContent() { result = getATypedContent(this) }
 
+  /** Gets the content. */
+  ContentApprox getContent() { result = c }
+
   /** Gets the container type. */
   DataFlowType getContainerType() { result = t }
 
@@ -1408,6 +1414,8 @@ abstract class ApproxAccessPathFront extends TApproxAccessPathFront {
 
   abstract boolean toBoolNonEmpty();
 
+  TypedContentApprox getHead() { this = TApproxFrontHead(result) }
+
   pragma[nomagic]
   TypedContent getAHead() {
     exists(TypedContentApprox cont |

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

@@ -318,3 +318,12 @@ private class MyConsistencyConfiguration extends Consistency::ConsistencyConfigu
     // consistency alerts enough that most of them are interesting.
   }
 }
+
+/**
+ * Gets an additional term that is added to the `join` and `branch` computations to reflect
+ * an additional forward or backwards branching factor that is not taken into account
+ * when calculating the (virtual) dispatch cost.
+ *
+ * Argument `arg` is part of a path from a source to a sink, and `p` is the target parameter.
+ */
+int getAdditionalFlowIntoCallNodeTerm(ArgumentNode arg, ParameterNode p) { none() }

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

@@ -0,0 +1,63 @@
+/**
+ * Provides classes for performing local (intra-procedural) and
+ * global (inter-procedural) taint-tracking analyses.
+ */
+
+import TaintTrackingParameter::Public
+private import TaintTrackingParameter::Private
+
+private module AddTaintDefaults<DataFlowInternal::FullStateConfigSig Config> implements
+DataFlowInternal::FullStateConfigSig {
+  import Config
+
+  predicate isBarrier(DataFlow::Node node) {
+    Config::isBarrier(node) or defaultTaintSanitizer(node)
+  }
+
+  predicate isAdditionalFlowStep(DataFlow::Node node1, DataFlow::Node node2) {
+    Config::isAdditionalFlowStep(node1, node2) or
+    defaultAdditionalTaintStep(node1, node2)
+  }
+
+  predicate allowImplicitRead(DataFlow::Node node, DataFlow::ContentSet c) {
+    Config::allowImplicitRead(node, c)
+    or
+    (
+      Config::isSink(node, _) or
+      Config::isAdditionalFlowStep(node, _) or
+      Config::isAdditionalFlowStep(node, _, _, _)
+    ) and
+    defaultImplicitTaintRead(node, c)
+  }
+}
+
+/**
+ * Constructs a standard taint tracking computation.
+ */
+module Make<DataFlow::ConfigSig Config> implements DataFlow::DataFlowSig {
+  private module Config0 implements DataFlowInternal::FullStateConfigSig {
+    import DataFlowInternal::DefaultState<Config>
+    import Config
+  }
+
+  private module C implements DataFlowInternal::FullStateConfigSig {
+    import AddTaintDefaults<Config0>
+  }
+
+  import DataFlowInternal::Impl<C>
+}
+
+/**
+ * Constructs a taint tracking computation using flow state.
+ */
+module MakeWithState<DataFlow::StateConfigSig Config> implements DataFlow::DataFlowSig {
+  private module Config0 implements DataFlowInternal::FullStateConfigSig {
+    import Config
+  }
+
+  private module C implements DataFlowInternal::FullStateConfigSig {
+    import AddTaintDefaults<Config0>
+  }
+
+  import DataFlowInternal::Impl<C>
+}

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

@@ -2,4 +2,5 @@ import semmle.code.cpp.dataflow.internal.TaintTrackingUtil as Public
 
 module Private {
   import semmle.code.cpp.dataflow.DataFlow::DataFlow as DataFlow
+  import semmle.code.cpp.dataflow.internal.DataFlowImpl as DataFlowInternal
 }

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

@@ -22,5 +22,6 @@
 import cpp
 
 module DataFlow {
-  import semmle.code.cpp.ir.dataflow.internal.DataFlowImpl
+  import semmle.code.cpp.ir.dataflow.internal.DataFlow
+  import semmle.code.cpp.ir.dataflow.internal.DataFlowImpl1
 }

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

@@ -19,5 +19,6 @@ import semmle.code.cpp.ir.dataflow.DataFlow
 import semmle.code.cpp.ir.dataflow.DataFlow2
 
 module TaintTracking {
+  import semmle.code.cpp.ir.dataflow.internal.tainttracking1.TaintTracking
   import semmle.code.cpp.ir.dataflow.internal.tainttracking1.TaintTrackingImpl
 }

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

@@ -0,0 +1,245 @@
+/**
+ * Provides an implementation of global (interprocedural) data flow. This file
+ * re-exports the local (intraprocedural) data flow analysis from
+ * `DataFlowImplSpecific::Public` and adds a global analysis, mainly exposed
+ * through the `Make` and `MakeWithState` modules.
+ */
+
+private import DataFlowImplCommon
+private import DataFlowImplSpecific::Private
+import DataFlowImplSpecific::Public
+import DataFlowImplCommonPublic
+private import DataFlowImpl
+
+/** An input configuration for data flow. */
+signature module ConfigSig {
+  /**
+   * Holds if `source` is a relevant data flow source.
+   */
+  predicate isSource(Node source);
+
+  /**
+   * Holds if `sink` is a relevant data flow sink.
+   */
+  predicate isSink(Node sink);
+
+  /**
+   * Holds if data flow through `node` is prohibited. This completely removes
+   * `node` from the data flow graph.
+   */
+  default predicate isBarrier(Node node) { none() }
+
+  /** Holds if data flow into `node` is prohibited. */
+  default predicate isBarrierIn(Node node) { none() }
+
+  /** Holds if data flow out of `node` is prohibited. */
+  default predicate isBarrierOut(Node node) { none() }
+
+  /**
+   * Holds if data may flow from `node1` to `node2` in addition to the normal data-flow steps.
+   */
+  default predicate isAdditionalFlowStep(Node node1, Node node2) { none() }
+
+  /**
+   * Holds if an arbitrary number of implicit read steps of content `c` may be
+   * taken at `node`.
+   */
+  default predicate allowImplicitRead(Node node, ContentSet c) { none() }
+
+  /**
+   * Gets the virtual dispatch branching limit when calculating field flow.
+   * This can be overridden to a smaller value to improve performance (a
+   * value of 0 disables field flow), or a larger value to get more results.
+   */
+  default int fieldFlowBranchLimit() { result = 2 }
+
+  /**
+   * Gets a data flow configuration feature to add restrictions to the set of
+   * valid flow paths.
+   *
+   * - `FeatureHasSourceCallContext`:
+   *    Assume that sources have some existing call context to disallow
+   *    conflicting return-flow directly following the source.
+   * - `FeatureHasSinkCallContext`:
+   *    Assume that sinks have some existing call context to disallow
+   *    conflicting argument-to-parameter flow directly preceding the sink.
+   * - `FeatureEqualSourceSinkCallContext`:
+   *    Implies both of the above and additionally ensures that the entire flow
+   *    path preserves the call context.
+   *
+   * These features are generally not relevant for typical end-to-end data flow
+   * queries, but should only be used for constructing paths that need to
+   * somehow be pluggable in another path context.
+   */
+  default FlowFeature getAFeature() { none() }
+
+  /** Holds if sources should be grouped in the result of `hasFlowPath`. */
+  default predicate sourceGrouping(Node source, string sourceGroup) { none() }
+
+  /** Holds if sinks should be grouped in the result of `hasFlowPath`. */
+  default predicate sinkGrouping(Node sink, string sinkGroup) { none() }
+
+  /**
+   * Holds if hidden nodes should be included in the data flow graph.
+   *
+   * This feature should only be used for debugging or when the data flow graph
+   * is not visualized (as it is in a `path-problem` query).
+   */
+  default predicate includeHiddenNodes() { none() }
+}
+
+/** An input configuration for data flow using flow state. */
+signature module StateConfigSig {
+  bindingset[this]
+  class FlowState;
+
+  /**
+   * Holds if `source` is a relevant data flow source with the given initial
+   * `state`.
+   */
+  predicate isSource(Node source, FlowState state);
+
+  /**
+   * Holds if `sink` is a relevant data flow sink accepting `state`.
+   */
+  predicate isSink(Node sink, FlowState state);
+
+  /**
+   * Holds if data flow through `node` is prohibited. This completely removes
+   * `node` from the data flow graph.
+   */
+  default predicate isBarrier(Node node) { none() }
+
+  /**
+   * Holds if data flow through `node` is prohibited when the flow state is
+   * `state`.
+   */
+  predicate isBarrier(Node node, FlowState state);
+
+  /** Holds if data flow into `node` is prohibited. */
+  default predicate isBarrierIn(Node node) { none() }
+
+  /** Holds if data flow out of `node` is prohibited. */
+  default predicate isBarrierOut(Node node) { none() }
+
+  /**
+   * Holds if data may flow from `node1` to `node2` in addition to the normal data-flow steps.
+   */
+  default predicate isAdditionalFlowStep(Node node1, Node node2) { none() }
+
+  /**
+   * Holds if data may flow from `node1` to `node2` in addition to the normal data-flow steps.
+   * This step is only applicable in `state1` and updates the flow state to `state2`.
+   */
+  predicate isAdditionalFlowStep(Node node1, FlowState state1, Node node2, FlowState state2);
+
+  /**
+   * Holds if an arbitrary number of implicit read steps of content `c` may be
+   * taken at `node`.
+   */
+  default predicate allowImplicitRead(Node node, ContentSet c) { none() }
+
+  /**
+   * Gets the virtual dispatch branching limit when calculating field flow.
+   * This can be overridden to a smaller value to improve performance (a
+   * value of 0 disables field flow), or a larger value to get more results.
+   */
+  default int fieldFlowBranchLimit() { result = 2 }
+
+  /**
+   * Gets a data flow configuration feature to add restrictions to the set of
+   * valid flow paths.
+   *
+   * - `FeatureHasSourceCallContext`:
+   *    Assume that sources have some existing call context to disallow
+   *    conflicting return-flow directly following the source.
+   * - `FeatureHasSinkCallContext`:
+   *    Assume that sinks have some existing call context to disallow
+   *    conflicting argument-to-parameter flow directly preceding the sink.
+   * - `FeatureEqualSourceSinkCallContext`:
+   *    Implies both of the above and additionally ensures that the entire flow
+   *    path preserves the call context.
+   *
+   * These features are generally not relevant for typical end-to-end data flow
+   * queries, but should only be used for constructing paths that need to
+   * somehow be pluggable in another path context.
+   */
+  default FlowFeature getAFeature() { none() }
+
+  /** Holds if sources should be grouped in the result of `hasFlowPath`. */
+  default predicate sourceGrouping(Node source, string sourceGroup) { none() }
+
+  /** Holds if sinks should be grouped in the result of `hasFlowPath`. */
+  default predicate sinkGrouping(Node sink, string sinkGroup) { none() }
+
+  /**
+   * Holds if hidden nodes should be included in the data flow graph.
+   *
+   * This feature should only be used for debugging or when the data flow graph
+   * is not visualized (as it is in a `path-problem` query).
+   */
+  default predicate includeHiddenNodes() { none() }
+}
+
+/**
+ * Gets the exploration limit for `hasPartialFlow` and `hasPartialFlowRev`
+ * measured in approximate number of interprocedural steps.
+ */
+signature int explorationLimitSig();
+
+/**
+ * The output of a data flow computation.
+ */
+signature module DataFlowSig {
+  /**
+   * A `Node` augmented with a call context (except for sinks) and an access path.
+   * Only those `PathNode`s that are reachable from a source, and which can reach a sink, are generated.
+   */
+  class PathNode;
+
+  /**
+   * Holds if data can flow from `source` to `sink`.
+   *
+   * The corresponding paths are generated from the end-points and the graph
+   * included in the module `PathGraph`.
+   */
+  predicate hasFlowPath(PathNode source, PathNode sink);
+
+  /**
+   * Holds if data can flow from `source` to `sink`.
+   */
+  predicate hasFlow(Node source, Node sink);
+
+  /**
+   * Holds if data can flow from some source to `sink`.
+   */
+  predicate hasFlowTo(Node sink);
+
+  /**
+   * Holds if data can flow from some source to `sink`.
+   */
+  predicate hasFlowToExpr(DataFlowExpr sink);
+}
+
+/**
+ * Constructs a standard data flow computation.
+ */
+module Make<ConfigSig Config> implements DataFlowSig {
+  private module C implements FullStateConfigSig {
+    import DefaultState<Config>
+    import Config
+  }
+
+  import Impl<C>
+}
+
+/**
+ * Constructs a data flow computation using flow state.
+ */
+module MakeWithState<StateConfigSig Config> implements DataFlowSig {
+  private module C implements FullStateConfigSig {
+    import Config
+  }
+
+  import Impl<C>
+}

cpp/ql/lib/semmle/code/cpp/ir/dataflow/internal/DataFlowImpl1.qll

@@ -0,0 +1,396 @@
+/**
+ * DEPRECATED: Use `Make` and `MakeWithState` instead.
+ *
+ * Provides a `Configuration` class backwards-compatible interface to the data
+ * flow library.
+ */
+
+private import DataFlowImplCommon
+private import DataFlowImplSpecific::Private
+import DataFlowImplSpecific::Public
+private import DataFlowImpl
+import DataFlowImplCommonPublic
+import FlowStateString
+
+/**
+ * A configuration of interprocedural data flow analysis. This defines
+ * sources, sinks, and any other configurable aspect of the analysis. Each
+ * use of the global data flow library must define its own unique extension
+ * of this abstract class. To create a configuration, extend this class with
+ * a subclass whose characteristic predicate is a unique singleton string.
+ * For example, write
+ *
+ * ```ql
+ * class MyAnalysisConfiguration extends DataFlow::Configuration {
+ *   MyAnalysisConfiguration() { this = "MyAnalysisConfiguration" }
+ *   // Override `isSource` and `isSink`.
+ *   // Optionally override `isBarrier`.
+ *   // Optionally override `isAdditionalFlowStep`.
+ * }
+ * ```
+ * Conceptually, this defines a graph where the nodes are `DataFlow::Node`s and
+ * the edges are those data-flow steps that preserve the value of the node
+ * along with any additional edges defined by `isAdditionalFlowStep`.
+ * Specifying nodes in `isBarrier` will remove those nodes from the graph, and
+ * specifying nodes in `isBarrierIn` and/or `isBarrierOut` will remove in-going
+ * and/or out-going edges from those nodes, respectively.
+ *
+ * Then, to query whether there is flow between some `source` and `sink`,
+ * write
+ *
+ * ```ql
+ * exists(MyAnalysisConfiguration cfg | cfg.hasFlow(source, sink))
+ * ```
+ *
+ * Multiple configurations can coexist, but two classes extending
+ * `DataFlow::Configuration` should never depend on each other. One of them
+ * should instead depend on a `DataFlow2::Configuration`, a
+ * `DataFlow3::Configuration`, or a `DataFlow4::Configuration`.
+ */
+abstract class Configuration extends string {
+  bindingset[this]
+  Configuration() { any() }
+
+  /**
+   * Holds if `source` is a relevant data flow source.
+   */
+  predicate isSource(Node source) { none() }
+
+  /**
+   * Holds if `source` is a relevant data flow source with the given initial
+   * `state`.
+   */
+  predicate isSource(Node source, FlowState state) { none() }
+
+  /**
+   * Holds if `sink` is a relevant data flow sink.
+   */
+  predicate isSink(Node sink) { none() }
+
+  /**
+   * Holds if `sink` is a relevant data flow sink accepting `state`.
+   */
+  predicate isSink(Node sink, FlowState state) { none() }
+
+  /**
+   * Holds if data flow through `node` is prohibited. This completely removes
+   * `node` from the data flow graph.
+   */
+  predicate isBarrier(Node node) { none() }
+
+  /**
+   * Holds if data flow through `node` is prohibited when the flow state is
+   * `state`.
+   */
+  predicate isBarrier(Node node, FlowState state) { none() }
+
+  /** Holds if data flow into `node` is prohibited. */
+  predicate isBarrierIn(Node node) { none() }
+
+  /** Holds if data flow out of `node` is prohibited. */
+  predicate isBarrierOut(Node node) { none() }
+
+  /**
+   * DEPRECATED: Use `isBarrier` and `BarrierGuard` module instead.
+   *
+   * Holds if data flow through nodes guarded by `guard` is prohibited.
+   */
+  deprecated predicate isBarrierGuard(BarrierGuard guard) { none() }
+
+  /**
+   * DEPRECATED: Use `isBarrier` and `BarrierGuard` module instead.
+   *
+   * Holds if data flow through nodes guarded by `guard` is prohibited when
+   * the flow state is `state`
+   */
+  deprecated predicate isBarrierGuard(BarrierGuard guard, FlowState state) { none() }
+
+  /**
+   * Holds if data may flow from `node1` to `node2` in addition to the normal data-flow steps.
+   */
+  predicate isAdditionalFlowStep(Node node1, Node node2) { none() }
+
+  /**
+   * Holds if data may flow from `node1` to `node2` in addition to the normal data-flow steps.
+   * This step is only applicable in `state1` and updates the flow state to `state2`.
+   */
+  predicate isAdditionalFlowStep(Node node1, FlowState state1, Node node2, FlowState state2) {
+    none()
+  }
+
+  /**
+   * Holds if an arbitrary number of implicit read steps of content `c` may be
+   * taken at `node`.
+   */
+  predicate allowImplicitRead(Node node, ContentSet c) { none() }
+
+  /**
+   * Gets the virtual dispatch branching limit when calculating field flow.
+   * This can be overridden to a smaller value to improve performance (a
+   * value of 0 disables field flow), or a larger value to get more results.
+   */
+  int fieldFlowBranchLimit() { result = 2 }
+
+  /**
+   * Gets a data flow configuration feature to add restrictions to the set of
+   * valid flow paths.
+   *
+   * - `FeatureHasSourceCallContext`:
+   *    Assume that sources have some existing call context to disallow
+   *    conflicting return-flow directly following the source.
+   * - `FeatureHasSinkCallContext`:
+   *    Assume that sinks have some existing call context to disallow
+   *    conflicting argument-to-parameter flow directly preceding the sink.
+   * - `FeatureEqualSourceSinkCallContext`:
+   *    Implies both of the above and additionally ensures that the entire flow
+   *    path preserves the call context.
+   *
+   * These features are generally not relevant for typical end-to-end data flow
+   * queries, but should only be used for constructing paths that need to
+   * somehow be pluggable in another path context.
+   */
+  FlowFeature getAFeature() { none() }
+
+  /** Holds if sources should be grouped in the result of `hasFlowPath`. */
+  predicate sourceGrouping(Node source, string sourceGroup) { none() }
+
+  /** Holds if sinks should be grouped in the result of `hasFlowPath`. */
+  predicate sinkGrouping(Node sink, string sinkGroup) { none() }
+
+  /**
+   * Holds if data may flow from `source` to `sink` for this configuration.
+   */
+  predicate hasFlow(Node source, Node sink) { hasFlow(source, sink, this) }
+
+  /**
+   * Holds if data may flow from `source` to `sink` for this configuration.
+   *
+   * The corresponding paths are generated from the end-points and the graph
+   * included in the module `PathGraph`.
+   */
+  predicate hasFlowPath(PathNode source, PathNode sink) { hasFlowPath(source, sink, this) }
+
+  /**
+   * Holds if data may flow from some source to `sink` for this configuration.
+   */
+  predicate hasFlowTo(Node sink) { hasFlowTo(sink, this) }
+
+  /**
+   * Holds if data may flow from some source to `sink` for this configuration.
+   */
+  predicate hasFlowToExpr(DataFlowExpr sink) { this.hasFlowTo(exprNode(sink)) }
+
+  /**
+   * DEPRECATED: Use `FlowExploration<explorationLimit>` instead.
+   *
+   * Gets the exploration limit for `hasPartialFlow` and `hasPartialFlowRev`
+   * measured in approximate number of interprocedural steps.
+   */
+  deprecated int explorationLimit() { none() }
+
+  /**
+   * Holds if hidden nodes should be included in the data flow graph.
+   *
+   * This feature should only be used for debugging or when the data flow graph
+   * is not visualized (for example in a `path-problem` query).
+   */
+  predicate includeHiddenNodes() { none() }
+}
+
+/**
+ * This class exists to prevent mutual recursion between the user-overridden
+ * member predicates of `Configuration` and the rest of the data-flow library.
+ * Good performance cannot be guaranteed in the presence of such recursion, so
+ * it should be replaced by using more than one copy of the data flow library.
+ */
+abstract private class ConfigurationRecursionPrevention extends Configuration {
+  bindingset[this]
+  ConfigurationRecursionPrevention() { any() }
+
+  override predicate hasFlow(Node source, Node sink) {
+    strictcount(Node n | this.isSource(n)) < 0
+    or
+    strictcount(Node n | this.isSource(n, _)) < 0
+    or
+    strictcount(Node n | this.isSink(n)) < 0
+    or
+    strictcount(Node n | this.isSink(n, _)) < 0
+    or
+    strictcount(Node n1, Node n2 | this.isAdditionalFlowStep(n1, n2)) < 0
+    or
+    strictcount(Node n1, Node n2 | this.isAdditionalFlowStep(n1, _, n2, _)) < 0
+    or
+    super.hasFlow(source, sink)
+  }
+}
+
+/** A bridge class to access the deprecated `isBarrierGuard`. */
+private class BarrierGuardGuardedNodeBridge extends Unit {
+  abstract predicate guardedNode(Node n, Configuration config);
+
+  abstract predicate guardedNode(Node n, FlowState state, Configuration config);
+}
+
+private class BarrierGuardGuardedNode extends BarrierGuardGuardedNodeBridge {
+  deprecated override predicate guardedNode(Node n, Configuration config) {
+    exists(BarrierGuard g |
+      config.isBarrierGuard(g) and
+      n = g.getAGuardedNode()
+    )
+  }
+
+  deprecated override predicate guardedNode(Node n, FlowState state, Configuration config) {
+    exists(BarrierGuard g |
+      config.isBarrierGuard(g, state) and
+      n = g.getAGuardedNode()
+    )
+  }
+}
+
+private FlowState relevantState(Configuration config) {
+  config.isSource(_, result) or
+  config.isSink(_, result) or
+  config.isBarrier(_, result) or
+  config.isAdditionalFlowStep(_, result, _, _) or
+  config.isAdditionalFlowStep(_, _, _, result)
+}
+
+private newtype TConfigState =
+  TMkConfigState(Configuration config, FlowState state) {
+    state = relevantState(config) or state instanceof FlowStateEmpty
+  }
+
+private Configuration getConfig(TConfigState state) { state = TMkConfigState(result, _) }
+
+private FlowState getState(TConfigState state) { state = TMkConfigState(_, result) }
+
+private predicate singleConfiguration() { 1 = strictcount(Configuration c) }
+
+private module Config implements FullStateConfigSig {
+  class FlowState = TConfigState;
+
+  predicate isSource(Node source, FlowState state) {
+    getConfig(state).isSource(source, getState(state))
+    or
+    getConfig(state).isSource(source) and getState(state) instanceof FlowStateEmpty
+  }
+
+  predicate isSink(Node sink, FlowState state) {
+    getConfig(state).isSink(sink, getState(state))
+    or
+    getConfig(state).isSink(sink) and getState(state) instanceof FlowStateEmpty
+  }
+
+  predicate isBarrier(Node node) { none() }
+
+  predicate isBarrier(Node node, FlowState state) {
+    getConfig(state).isBarrier(node, getState(state)) or
+    getConfig(state).isBarrier(node) or
+    any(BarrierGuardGuardedNodeBridge b).guardedNode(node, getState(state), getConfig(state)) or
+    any(BarrierGuardGuardedNodeBridge b).guardedNode(node, getConfig(state))
+  }
+
+  predicate isBarrierIn(Node node) { any(Configuration config).isBarrierIn(node) }
+
+  predicate isBarrierOut(Node node) { any(Configuration config).isBarrierOut(node) }
+
+  predicate isAdditionalFlowStep(Node node1, Node node2) {
+    singleConfiguration() and
+    any(Configuration config).isAdditionalFlowStep(node1, node2)
+  }
+
+  predicate isAdditionalFlowStep(Node node1, FlowState state1, Node node2, FlowState state2) {
+    getConfig(state1).isAdditionalFlowStep(node1, getState(state1), node2, getState(state2)) and
+    getConfig(state2) = getConfig(state1)
+    or
+    not singleConfiguration() and
+    getConfig(state1).isAdditionalFlowStep(node1, node2) and
+    state2 = state1
+  }
+
+  predicate allowImplicitRead(Node node, ContentSet c) {
+    any(Configuration config).allowImplicitRead(node, c)
+  }
+
+  int fieldFlowBranchLimit() { result = min(any(Configuration config).fieldFlowBranchLimit()) }
+
+  FlowFeature getAFeature() { result = any(Configuration config).getAFeature() }
+
+  predicate sourceGrouping(Node source, string sourceGroup) {
+    any(Configuration config).sourceGrouping(source, sourceGroup)
+  }
+
+  predicate sinkGrouping(Node sink, string sinkGroup) {
+    any(Configuration config).sinkGrouping(sink, sinkGroup)
+  }
+
+  predicate includeHiddenNodes() { any(Configuration config).includeHiddenNodes() }
+}
+
+private import Impl<Config> as I
+import I
+
+/**
+ * A `Node` augmented with a call context (except for sinks), an access path, and a configuration.
+ * Only those `PathNode`s that are reachable from a source, and which can reach a sink, are generated.
+ */
+class PathNode instanceof I::PathNode {
+  /** Gets a textual representation of this element. */
+  final string toString() { result = super.toString() }
+
+  /**
+   * Gets a textual representation of this element, including a textual
+   * representation of the call context.
+   */
+  final string toStringWithContext() { result = super.toStringWithContext() }
+
+  /**
+   * Holds if this element is at the specified location.
+   * The location spans column `startcolumn` of line `startline` to
+   * column `endcolumn` of line `endline` in file `filepath`.
+   * For more information, see
+   * [Locations](https://codeql.github.com/docs/writing-codeql-queries/providing-locations-in-codeql-queries/).
+   */
+  final predicate hasLocationInfo(
+    string filepath, int startline, int startcolumn, int endline, int endcolumn
+  ) {
+    super.hasLocationInfo(filepath, startline, startcolumn, endline, endcolumn)
+  }
+
+  /** Gets the underlying `Node`. */
+  final Node getNode() { result = super.getNode() }
+
+  /** Gets the `FlowState` of this node. */
+  final FlowState getState() { result = getState(super.getState()) }
+
+  /** Gets the associated configuration. */
+  final Configuration getConfiguration() { result = getConfig(super.getState()) }
+
+  /** Gets a successor of this node, if any. */
+  final PathNode getASuccessor() { result = super.getASuccessor() }
+
+  /** Holds if this node is a source. */
+  final predicate isSource() { super.isSource() }
+
+  /** Holds if this node is a grouping of source nodes. */
+  final predicate isSourceGroup(string group) { super.isSourceGroup(group) }
+
+  /** Holds if this node is a grouping of sink nodes. */
+  final predicate isSinkGroup(string group) { super.isSinkGroup(group) }
+}
+
+private predicate hasFlow(Node source, Node sink, Configuration config) {
+  exists(PathNode source0, PathNode sink0 |
+    hasFlowPath(source0, sink0, config) and
+    source0.getNode() = source and
+    sink0.getNode() = sink
+  )
+}
+
+private predicate hasFlowPath(PathNode source, PathNode sink, Configuration config) {
+  hasFlowPath(source, sink) and source.getConfiguration() = config
+}
+
+private predicate hasFlowTo(Node sink, Configuration config) { hasFlow(_, sink, config) }
+
+predicate flowsTo = hasFlow/3;

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

@@ -3,15 +3,18 @@ private import DataFlowImplSpecific::Public
 import Cached
 
 module DataFlowImplCommonPublic {
-  /** A state value to track during data flow. */
-  class FlowState = string;
+  /** Provides `FlowState = string`. */
+  module FlowStateString {
+    /** A state value to track during data flow. */
+    class FlowState = string;
 
-  /**
-   * The default state, which is used when the state is unspecified for a source
-   * or a sink.
-   */
-  class FlowStateEmpty extends FlowState {
-    FlowStateEmpty() { this = "" }
+    /**
+     * The default state, which is used when the state is unspecified for a source
+     * or a sink.
+     */
+    class FlowStateEmpty extends FlowState {
+      FlowStateEmpty() { this = "" }
+    }
   }
 
   private newtype TFlowFeature =
@@ -707,8 +710,8 @@ private module Cached {
      * Gets a viable dispatch target of `call` in the context `ctx`. This is
      * restricted to those `call`s for which a context might make a difference.
      */
-    pragma[nomagic]
-    private DataFlowCallable viableImplInCallContextExt(DataFlowCall call, DataFlowCall ctx) {
+    cached
+    DataFlowCallable viableImplInCallContextExt(DataFlowCall call, DataFlowCall ctx) {
       result = viableImplInCallContext(call, ctx) and
       result = viableCallable(call)
       or
@@ -1391,6 +1394,9 @@ class TypedContentApprox extends MkTypedContentApprox {
   /** Gets a typed content approximated by this value. */
   TypedContent getATypedContent() { result = getATypedContent(this) }
 
+  /** Gets the content. */
+  ContentApprox getContent() { result = c }
+
   /** Gets the container type. */
   DataFlowType getContainerType() { result = t }
 
@@ -1408,6 +1414,8 @@ abstract class ApproxAccessPathFront extends TApproxAccessPathFront {
 
   abstract boolean toBoolNonEmpty();
 
+  TypedContentApprox getHead() { this = TApproxFrontHead(result) }
+
   pragma[nomagic]
   TypedContent getAHead() {
     exists(TypedContentApprox cont |

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

@@ -414,3 +414,12 @@ private class MyConsistencyConfiguration extends Consistency::ConsistencyConfigu
     any()
   }
 }
+
+/**
+ * Gets an additional term that is added to the `join` and `branch` computations to reflect
+ * an additional forward or backwards branching factor that is not taken into account
+ * when calculating the (virtual) dispatch cost.
+ *
+ * Argument `arg` is part of a path from a source to a sink, and `p` is the target parameter.
+ */
+int getAdditionalFlowIntoCallNodeTerm(ArgumentNode arg, ParameterNode p) { none() }

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

@@ -97,23 +97,23 @@ private string getNodeProperty(DataFlow::Node node, string key) {
     |
       kind, ", "
     )
-  or
-  // Is there partial flow from a source to this node?
-  // This property will only be emitted if partial flow is enabled by overriding
-  // `DataFlow::Configuration::explorationLimit()`.
-  key = "pflow" and
-  result =
-    strictconcat(DataFlow::PartialPathNode sourceNode, DataFlow::PartialPathNode destNode, int dist,
-      int order1, int order2 |
-      any(DataFlow::Configuration cfg).hasPartialFlow(sourceNode, destNode, dist) and
-      destNode.getNode() = node and
-      // Only print flow from a source in the same function.
-      sourceNode.getNode().getEnclosingCallable() = node.getEnclosingCallable()
-    |
-      nodeId(sourceNode.getNode(), order1, order2) + "+" + dist.toString(), ", "
-      order by
-        order1, order2, dist desc
-    )
+  // or
+  // // Is there partial flow from a source to this node?
+  // // This property will only be emitted if partial flow is enabled by overriding
+  // // `DataFlow::Configuration::explorationLimit()`.
+  // key = "pflow" and
+  // result =
+  //   strictconcat(DataFlow::PartialPathNode sourceNode, DataFlow::PartialPathNode destNode, int dist,
+  //     int order1, int order2 |
+  //     any(DataFlow::Configuration cfg).hasPartialFlow(sourceNode, destNode, dist) and
+  //     destNode.getNode() = node and
+  //     // Only print flow from a source in the same function.
+  //     sourceNode.getNode().getEnclosingCallable() = node.getEnclosingCallable()
+  //   |
+  //     nodeId(sourceNode.getNode(), order1, order2) + "+" + dist.toString(), ", "
+  //     order by
+  //       order1, order2, dist desc
+  //   )
 }
 
 /**

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

@@ -169,19 +169,11 @@ predicate defaultTaintSanitizer(DataFlow::Node node) { none() }
  */
 predicate modeledTaintStep(Operand nodeIn, Instruction nodeOut) {
   exists(CallInstruction call, TaintFunction func, FunctionInput modelIn, FunctionOutput modelOut |
-    (
-      nodeIn = callInput(call, modelIn)
-      or
-      exists(int n |
-        modelIn.isParameterDerefOrQualifierObject(n) and
-        if n = -1
-        then nodeIn = callInput(call, any(InQualifierObject inQualifier))
-        else nodeIn = callInput(call, any(InParameter inParam | inParam.getIndex() = n))
-      )
-    ) and
-    nodeOut = callOutput(call, modelOut) and
     call.getStaticCallTarget() = func and
     func.hasTaintFlow(modelIn, modelOut)
+  |
+    nodeIn = callInput(call, modelIn) and
+    nodeOut = callOutput(call, modelOut)
   )
   or
   // Taint flow from one argument to another and data flow from an argument to a

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

@@ -0,0 +1,63 @@
+/**
+ * Provides classes for performing local (intra-procedural) and
+ * global (inter-procedural) taint-tracking analyses.
+ */
+
+import TaintTrackingParameter::Public
+private import TaintTrackingParameter::Private
+
+private module AddTaintDefaults<DataFlowInternal::FullStateConfigSig Config> implements
+DataFlowInternal::FullStateConfigSig {
+  import Config
+
+  predicate isBarrier(DataFlow::Node node) {
+    Config::isBarrier(node) or defaultTaintSanitizer(node)
+  }
+
+  predicate isAdditionalFlowStep(DataFlow::Node node1, DataFlow::Node node2) {
+    Config::isAdditionalFlowStep(node1, node2) or
+    defaultAdditionalTaintStep(node1, node2)
+  }
+
+  predicate allowImplicitRead(DataFlow::Node node, DataFlow::ContentSet c) {
+    Config::allowImplicitRead(node, c)
+    or
+    (
+      Config::isSink(node, _) or
+      Config::isAdditionalFlowStep(node, _) or
+      Config::isAdditionalFlowStep(node, _, _, _)
+    ) and
+    defaultImplicitTaintRead(node, c)
+  }
+}
+
+/**
+ * Constructs a standard taint tracking computation.
+ */
+module Make<DataFlow::ConfigSig Config> implements DataFlow::DataFlowSig {
+  private module Config0 implements DataFlowInternal::FullStateConfigSig {
+    import DataFlowInternal::DefaultState<Config>
+    import Config
+  }
+
+  private module C implements DataFlowInternal::FullStateConfigSig {
+    import AddTaintDefaults<Config0>
+  }
+
+  import DataFlowInternal::Impl<C>
+}
+
+/**
+ * Constructs a taint tracking computation using flow state.
+ */
+module MakeWithState<DataFlow::StateConfigSig Config> implements DataFlow::DataFlowSig {
+  private module Config0 implements DataFlowInternal::FullStateConfigSig {
+    import Config
+  }
+
+  private module C implements DataFlowInternal::FullStateConfigSig {
+    import AddTaintDefaults<Config0>
+  }
+
+  import DataFlowInternal::Impl<C>
+}

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

@@ -2,4 +2,5 @@ import semmle.code.cpp.ir.dataflow.internal.TaintTrackingUtil as Public
 
 module Private {
   import semmle.code.cpp.ir.dataflow.DataFlow::DataFlow as DataFlow
+  import semmle.code.cpp.ir.dataflow.internal.DataFlowImpl as DataFlowInternal
 }

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

@@ -412,6 +412,9 @@ class CallTargetOperand extends RegisterOperand {
  */
 class ArgumentOperand extends RegisterOperand {
   override ArgumentOperandTag tag;
+
+  /** Gets the `CallInstruction` for which this is an argument. */
+  CallInstruction getCall() { result.getAnArgumentOperand() = this }
 }
 
 /**

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

@@ -412,6 +412,9 @@ class CallTargetOperand extends RegisterOperand {
  */
 class ArgumentOperand extends RegisterOperand {
   override ArgumentOperandTag tag;
+
+  /** Gets the `CallInstruction` for which this is an argument. */
+  CallInstruction getCall() { result.getAnArgumentOperand() = this }
 }
 
 /**

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

@@ -412,6 +412,9 @@ class CallTargetOperand extends RegisterOperand {
  */
 class ArgumentOperand extends RegisterOperand {
   override ArgumentOperandTag tag;
+
+  /** Gets the `CallInstruction` for which this is an argument. */
+  CallInstruction getCall() { result.getAnArgumentOperand() = this }
 }
 
 /**

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

@@ -107,130 +107,34 @@ private FunctionInput getIteratorArgumentInput(Operator op, int index) {
   )
 }
 
-/**
- * A non-member prefix `operator*` function for an iterator type.
- */
-private class IteratorPointerDereferenceOperator extends Operator, TaintFunction,
-  IteratorReferenceFunction {
-  FunctionInput iteratorInput;
-
-  IteratorPointerDereferenceOperator() {
-    this.hasName("operator*") and
-    iteratorInput = getIteratorArgumentInput(this, 0)
-  }
-
-  override predicate hasTaintFlow(FunctionInput input, FunctionOutput output) {
-    input = iteratorInput and
-    output.isReturnValue()
-    or
-    input.isReturnValueDeref() and
-    output.isParameterDeref(0)
-  }
-}
-
 /**
  * A non-member `operator++` or `operator--` function for an iterator type.
+ *
+ * Note that this class _only_ matches non-member functions. To find both
+ * non-member and versions, use `IteratorCrementOperator`.
  */
-private class IteratorCrementOperator extends Operator, DataFlowFunction {
-  FunctionInput iteratorInput;
-
-  IteratorCrementOperator() {
+class IteratorCrementNonMemberOperator extends Operator {
+  IteratorCrementNonMemberOperator() {
     this.hasName(["operator++", "operator--"]) and
-    iteratorInput = getIteratorArgumentInput(this, 0)
-  }
-
-  override predicate hasDataFlow(FunctionInput input, FunctionOutput output) {
-    input = iteratorInput and
-    output.isReturnValue()
-    or
-    input.isParameterDeref(0) and output.isReturnValueDeref()
-  }
-}
-
-/**
- * A non-member `operator+` function for an iterator type.
- */
-private class IteratorAddOperator extends Operator, TaintFunction {
-  FunctionInput iteratorInput;
-
-  IteratorAddOperator() {
-    this.hasName("operator+") and
-    iteratorInput = getIteratorArgumentInput(this, [0, 1])
-  }
-
-  override predicate hasTaintFlow(FunctionInput input, FunctionOutput output) {
-    input = iteratorInput and
-    output.isReturnValue()
-  }
-}
-
-/**
- * A non-member `operator-` function that takes a pointer difference type as its second argument.
- */
-private class IteratorSubOperator extends Operator, TaintFunction {
-  FunctionInput iteratorInput;
-
-  IteratorSubOperator() {
-    this.hasName("operator-") and
-    iteratorInput = getIteratorArgumentInput(this, 0) and
-    this.getParameter(1).getUnspecifiedType() instanceof IntegralType // not an iterator difference
-  }
-
-  override predicate hasTaintFlow(FunctionInput input, FunctionOutput output) {
-    input = iteratorInput and
-    output.isReturnValue()
-  }
-}
-
-/**
- * A non-member `operator+=` or `operator-=` function for an iterator type.
- */
-class IteratorAssignArithmeticOperator extends Operator {
-  IteratorAssignArithmeticOperator() {
-    this.hasName(["operator+=", "operator-="]) and
     exists(getIteratorArgumentInput(this, 0))
   }
 }
 
-private class IteratorAssignArithmeticOperatorModel extends IteratorAssignArithmeticOperator,
-  DataFlowFunction, TaintFunction {
+private class IteratorCrementNonMemberOperatorModel extends IteratorCrementNonMemberOperator,
+  DataFlowFunction {
   override predicate hasDataFlow(FunctionInput input, FunctionOutput output) {
-    input.isParameter(0) and
+    input = getIteratorArgumentInput(this, 0) and
     output.isReturnValue()
-  }
-
-  override predicate hasTaintFlow(FunctionInput input, FunctionOutput output) {
+    or
     input.isParameterDeref(0) and output.isReturnValueDeref()
-    or
-    // reverse flow from returned reference to the object referenced by the first parameter
-    input.isReturnValueDeref() and
-    output.isParameterDeref(0)
-    or
-    input.isParameterDeref(1) and
-    output.isParameterDeref(0)
-  }
-}
-
-/**
- * A prefix `operator*` member function for an iterator type.
- */
-class IteratorPointerDereferenceMemberOperator extends MemberFunction, TaintFunction,
-  IteratorReferenceFunction {
-  IteratorPointerDereferenceMemberOperator() {
-    this.getClassAndName("operator*") instanceof Iterator
-  }
-
-  override predicate hasTaintFlow(FunctionInput input, FunctionOutput output) {
-    input.isQualifierObject() and
-    output.isReturnValue()
-    or
-    input.isReturnValueDeref() and
-    output.isQualifierObject()
   }
 }
 
 /**
  * An `operator++` or `operator--` member function for an iterator type.
+ *
+ * Note that this class _only_ matches member functions. To find both
+ * non-member and member versions, use `IteratorCrementOperator`.
  */
 class IteratorCrementMemberOperator extends MemberFunction {
   IteratorCrementMemberOperator() {
@@ -258,25 +162,49 @@ private class IteratorCrementMemberOperatorModel extends IteratorCrementMemberOp
 }
 
 /**
- * A member `operator->` function for an iterator type.
+ * A (member or non-member) `operator++` or `operator--` function for an iterator type.
  */
-private class IteratorFieldMemberOperator extends Operator, TaintFunction {
-  IteratorFieldMemberOperator() { this.getClassAndName("operator->") instanceof Iterator }
+class IteratorCrementOperator extends Function {
+  IteratorCrementOperator() {
+    this instanceof IteratorCrementNonMemberOperator or
+    this instanceof IteratorCrementMemberOperator
+  }
+}
 
+/**
+ * A non-member `operator+` function for an iterator type.
+ *
+ * Note that this class _only_ matches non-member functions. To find both
+ * non-member and member versions, use `IteratorBinaryAddOperator`.
+ */
+class IteratorAddNonMemberOperator extends Operator {
+  IteratorAddNonMemberOperator() {
+    this.hasName("operator+") and
+    exists(getIteratorArgumentInput(this, [0, 1]))
+  }
+}
+
+private class IteratorAddNonMemberOperatorModel extends IteratorAddNonMemberOperator, TaintFunction {
   override predicate hasTaintFlow(FunctionInput input, FunctionOutput output) {
-    input.isQualifierObject() and
+    input = getIteratorArgumentInput(this, [0, 1]) and
     output.isReturnValue()
   }
 }
 
 /**
  * An `operator+` or `operator-` member function of an iterator class.
+ *
+ * Note that this class _only_ matches member functions. To find both
+ * non-member and member versions, use `IteratorBinaryAddOperator`.
  */
-private class IteratorBinaryArithmeticMemberOperator extends MemberFunction, TaintFunction {
+class IteratorBinaryArithmeticMemberOperator extends MemberFunction {
   IteratorBinaryArithmeticMemberOperator() {
     this.getClassAndName(["operator+", "operator-"]) instanceof Iterator
   }
+}
 
+private class IteratorBinaryArithmeticMemberOperatorModel extends IteratorBinaryArithmeticMemberOperator,
+  TaintFunction {
   override predicate hasTaintFlow(FunctionInput input, FunctionOutput output) {
     input.isQualifierObject() and
     output.isReturnValue()
@@ -284,14 +212,84 @@ private class IteratorBinaryArithmeticMemberOperator extends MemberFunction, Tai
 }
 
 /**
- * An `operator+=` or `operator-=` member function of an iterator class.
+ * A (member or non-member) `operator+` or `operator-` function for an iterator type.
  */
-private class IteratorAssignArithmeticMemberOperator extends MemberFunction, DataFlowFunction,
-  TaintFunction {
+class IteratorBinaryArithmeticOperator extends Function {
+  IteratorBinaryArithmeticOperator() {
+    this instanceof IteratorAddNonMemberOperator or
+    this instanceof IteratorSubNonMemberOperator or
+    this instanceof IteratorBinaryArithmeticMemberOperator
+  }
+}
+
+/**
+ * A non-member `operator-` function that takes a pointer difference type as its second argument.
+ *
+ * Note that this class _only_ matches non-member functions. To find both
+ * non-member and member versions, use `IteratorBinaryArithmeticOperator` (which also
+ * includes `operator+` versions).
+ */
+class IteratorSubNonMemberOperator extends Operator {
+  IteratorSubNonMemberOperator() {
+    this.hasName("operator-") and
+    exists(getIteratorArgumentInput(this, 0)) and
+    this.getParameter(1).getUnspecifiedType() instanceof IntegralType // not an iterator difference
+  }
+}
+
+private class IteratorSubOperatorModel extends IteratorSubNonMemberOperator, TaintFunction {
+  override predicate hasTaintFlow(FunctionInput input, FunctionOutput output) {
+    input = getIteratorArgumentInput(this, 0) and
+    output.isReturnValue()
+  }
+}
+
+/**
+ * A non-member `operator+=` or `operator-=` function for an iterator type.
+ *
+ * Note that this class _only_ matches non-member functions. To find both
+ * non-member and member versions, use `IteratorAssignArithmeticOperator`.
+ */
+class IteratorAssignArithmeticNonMemberOperator extends Operator {
+  IteratorAssignArithmeticNonMemberOperator() {
+    this.hasName(["operator+=", "operator-="]) and
+    exists(getIteratorArgumentInput(this, 0))
+  }
+}
+
+private class IteratorAssignArithmeticNonMemberOperatorModel extends IteratorAssignArithmeticNonMemberOperator,
+  DataFlowFunction, TaintFunction {
+  override predicate hasDataFlow(FunctionInput input, FunctionOutput output) {
+    input.isParameter(0) and
+    output.isReturnValue()
+  }
+
+  override predicate hasTaintFlow(FunctionInput input, FunctionOutput output) {
+    input.isParameterDeref(0) and output.isReturnValueDeref()
+    or
+    // reverse flow from returned reference to the object referenced by the first parameter
+    input.isReturnValueDeref() and
+    output.isParameterDeref(0)
+    or
+    (input.isParameter(1) or input.isParameterDeref(1)) and
+    output.isParameterDeref(0)
+  }
+}
+
+/**
+ * An `operator+=` or `operator-=` member function of an iterator class.
+ *
+ * Note that this class _only_ matches member functions. To find both
+ * non-member and member versions, use `IteratorAssignArithmeticOperator`.
+ */
+class IteratorAssignArithmeticMemberOperator extends MemberFunction {
   IteratorAssignArithmeticMemberOperator() {
     this.getClassAndName(["operator+=", "operator-="]) instanceof Iterator
   }
+}
 
+private class IteratorAssignArithmeticMemberOperatorModel extends IteratorAssignArithmeticMemberOperator,
+  DataFlowFunction, TaintFunction {
   override predicate hasDataFlow(FunctionInput input, FunctionOutput output) {
     input.isQualifierAddress() and
     output.isReturnValue()
@@ -305,11 +303,88 @@ private class IteratorAssignArithmeticMemberOperator extends MemberFunction, Dat
     input.isReturnValueDeref() and
     output.isQualifierObject()
     or
-    input.isParameterDeref(0) and
+    (input.isParameter(0) or input.isParameterDeref(0)) and
     output.isQualifierObject()
   }
 }
 
+/**
+ * A (member or non-member) `operator+=` or `operator-=` function for an iterator type.
+ */
+class IteratorAssignArithmeticOperator extends Function {
+  IteratorAssignArithmeticOperator() {
+    this instanceof IteratorAssignArithmeticNonMemberOperator or
+    this instanceof IteratorAssignArithmeticMemberOperator
+  }
+}
+
+/**
+ * A prefix `operator*` member function for an iterator type.
+ *
+ * Note that this class _only_ matches member functions. To find both
+ * non-member and member versions, use `IteratorPointerDereferenceOperator`.
+ */
+class IteratorPointerDereferenceMemberOperator extends MemberFunction, TaintFunction,
+  IteratorReferenceFunction {
+  IteratorPointerDereferenceMemberOperator() {
+    this.getClassAndName("operator*") instanceof Iterator
+  }
+
+  override predicate hasTaintFlow(FunctionInput input, FunctionOutput output) {
+    input.isQualifierObject() and
+    output.isReturnValue()
+    or
+    input.isReturnValueDeref() and
+    output.isQualifierObject()
+  }
+}
+
+/**
+ * A non-member prefix `operator*` function for an iterator type.
+ *
+ * Note that this class _only_ matches non-member functions. To find both
+ * non-member and member versions, use `IteratorPointerDereferenceOperator`.
+ */
+class IteratorPointerDereferenceNonMemberOperator extends Operator, IteratorReferenceFunction {
+  IteratorPointerDereferenceNonMemberOperator() {
+    this.hasName("operator*") and
+    exists(getIteratorArgumentInput(this, 0))
+  }
+}
+
+private class IteratorPointerDereferenceNonMemberOperatorModel extends IteratorPointerDereferenceNonMemberOperator,
+  TaintFunction {
+  override predicate hasTaintFlow(FunctionInput input, FunctionOutput output) {
+    input = getIteratorArgumentInput(this, 0) and
+    output.isReturnValue()
+    or
+    input.isReturnValueDeref() and
+    output.isParameterDeref(0)
+  }
+}
+
+/**
+ * A (member or non-member) prefix `operator*` function for an iterator type.
+ */
+class IteratorPointerDereferenceOperator extends Function {
+  IteratorPointerDereferenceOperator() {
+    this instanceof IteratorPointerDereferenceNonMemberOperator or
+    this instanceof IteratorPointerDereferenceMemberOperator
+  }
+}
+
+/**
+ * A member `operator->` function for an iterator type.
+ */
+private class IteratorFieldMemberOperator extends Operator, TaintFunction {
+  IteratorFieldMemberOperator() { this.getClassAndName("operator->") instanceof Iterator }
+
+  override predicate hasTaintFlow(FunctionInput input, FunctionOutput output) {
+    input.isQualifierObject() and
+    output.isReturnValue()
+  }
+}
+
 /**
  * An `operator[]` member function of an iterator class.
  */
@@ -326,17 +401,24 @@ private class IteratorArrayMemberOperator extends MemberFunction, TaintFunction,
 /**
  * An `operator=` member function of an iterator class that is not a copy or move assignment
  * operator.
- *
- * The `hasTaintFlow` override provides flow through output iterators that return themselves with
- * `operator*` and use their own `operator=` to assign to the container.
  */
-private class IteratorAssignmentMemberOperator extends MemberFunction, TaintFunction {
+class IteratorAssignmentMemberOperator extends MemberFunction {
   IteratorAssignmentMemberOperator() {
     this.getClassAndName("operator=") instanceof Iterator and
     not this instanceof CopyAssignmentOperator and
     not this instanceof MoveAssignmentOperator
   }
+}
 
+/**
+ * An `operator=` member function of an iterator class that is not a copy or move assignment
+ * operator.
+ *
+ * The `hasTaintFlow` override provides flow through output iterators that return themselves with
+ * `operator*` and use their own `operator=` to assign to the container.
+ */
+private class IteratorAssignmentMemberOperatorModel extends IteratorAssignmentMemberOperator,
+  TaintFunction {
   override predicate hasTaintFlow(FunctionInput input, FunctionOutput output) {
     input.isParameterDeref(0) and
     output.isQualifierObject()

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

@@ -27,7 +27,12 @@ private class StdSetConstructor extends Constructor, TaintFunction {
 
   override predicate hasTaintFlow(FunctionInput input, FunctionOutput output) {
     // taint flow from any parameter of an iterator type to the qualifier
-    input.isParameterDeref(this.getAnIteratorParameterIndex()) and
+    (
+      // AST dataflow doesn't have indirection for iterators.
+      // Once we deprecate AST dataflow we can delete this first disjunct.
+      input.isParameter(this.getAnIteratorParameterIndex()) or
+      input.isParameterDeref(this.getAnIteratorParameterIndex())
+    ) and
     (
       output.isReturnValue() // TODO: this is only needed for AST data flow, which treats constructors as returning the new object
       or
@@ -45,7 +50,12 @@ private class StdSetInsert extends TaintFunction {
   override predicate hasTaintFlow(FunctionInput input, FunctionOutput output) {
     // flow from last parameter to qualifier and return value
     // (where the return value is a pair, this should really flow just to the first part of it)
-    input.isParameterDeref(this.getNumberOfParameters() - 1) and
+    (
+      // AST dataflow doesn't have indirection for iterators.
+      // Once we deprecate AST dataflow we can delete this first disjunct.
+      input.isParameter(this.getNumberOfParameters() - 1) or
+      input.isParameterDeref(this.getNumberOfParameters() - 1)
+    ) and
     (
       output.isQualifierObject() or
       output.isReturnValue()

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

@@ -38,8 +38,7 @@ private class StdBasicStringIterator extends Iterator, Type {
  */
 abstract private class StdStringTaintFunction extends TaintFunction {
   /**
-   * Gets the index of a parameter to this function that is a string (or
-   * character).
+   * Gets the index of a parameter to this function that is a string.
    */
   final int getAStringParameterIndex() {
     exists(Type paramType | paramType = this.getParameter(result).getUnspecifiedType() |
@@ -50,7 +49,14 @@ abstract private class StdStringTaintFunction extends TaintFunction {
       paramType instanceof ReferenceType and
       not paramType.(ReferenceType).getBaseType() =
         this.getDeclaringType().getTemplateArgument(2).(Type).getUnspecifiedType()
-      or
+    )
+  }
+
+  /**
+   * Gets the index of a parameter to this function that is a character.
+   */
+  final int getACharParameterIndex() {
+    exists(Type paramType | paramType = this.getParameter(result).getUnspecifiedType() |
       // i.e. `std::basic_string::CharT`
       paramType = this.getDeclaringType().getTemplateArgument(0).(Type).getUnspecifiedType()
     )
@@ -79,6 +85,7 @@ private class StdStringConstructor extends Constructor, StdStringTaintFunction {
     // taint flow from any parameter of the value type to the returned object
     (
       input.isParameterDeref(this.getAStringParameterIndex()) or
+      input.isParameter(this.getACharParameterIndex()) or
       input.isParameter(this.getAnIteratorParameterIndex())
     ) and
     (
@@ -128,7 +135,7 @@ private class StdStringPush extends StdStringTaintFunction {
 
   override predicate hasTaintFlow(FunctionInput input, FunctionOutput output) {
     // flow from parameter to qualifier
-    input.isParameterDeref(0) and
+    input.isParameter(0) and
     output.isQualifierObject()
   }
 }
@@ -180,6 +187,7 @@ private class StdStringAppend extends StdStringTaintFunction {
     (
       input.isQualifierObject() or
       input.isParameterDeref(this.getAStringParameterIndex()) or
+      input.isParameter(this.getACharParameterIndex()) or
       input.isParameter(this.getAnIteratorParameterIndex())
     ) and
     (
@@ -210,6 +218,7 @@ private class StdStringInsert extends StdStringTaintFunction {
     (
       input.isQualifierObject() or
       input.isParameterDeref(this.getAStringParameterIndex()) or
+      input.isParameter(this.getACharParameterIndex()) or
       input.isParameter(this.getAnIteratorParameterIndex())
     ) and
     (
@@ -236,6 +245,7 @@ private class StdStringAssign extends StdStringTaintFunction {
     // flow from parameter to string itself (qualifier) and return value
     (
       input.isParameterDeref(this.getAStringParameterIndex()) or
+      input.isParameter(this.getACharParameterIndex()) or
       input.isParameter(this.getAnIteratorParameterIndex())
     ) and
     (

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

@@ -158,11 +158,10 @@ abstract class FormattingFunction extends ArrayFunction, TaintFunction {
 
   override predicate hasTaintFlow(FunctionInput input, FunctionOutput output) {
     exists(int arg |
-      (
-        arg = getFormatParameterIndex() or
-        arg >= getFirstFormatArgumentIndex()
-      ) and
-      input.isParameterDeref(arg) and
+      arg = getFormatParameterIndex() or
+      arg >= getFirstFormatArgumentIndex()
+    |
+      (input.isParameterDeref(arg) or input.isParameter(arg)) and
       output.isParameterDeref(getOutputParameterIndex(_))
     )
   }

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

@@ -64,7 +64,7 @@ predicate isInsecureEncryption(string name) { name.regexpMatch(getInsecureAlgori
  */
 bindingset[name]
 predicate isEncryptionAdditionalEvidence(string name) {
-  name.toUpperCase().matches("%" + ["CRYPT", "CODE", "CODING", "CBC", "KEY", "CIPHER", "MAC"] + "%")
+  name.regexpMatch("(?i).*(crypt|code|coding|cbc|key|cipher|mac).*")
 }
 
 /**

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

@@ -14,8 +14,8 @@ import cpp
  */
 bindingset[s]
 private predicate suspicious(string s) {
-  s.regexpMatch(".*(password|passwd|accountid|account.?key|accnt.?key|license.?key|trusted).*") and
-  not s.matches(["%hash%", "%crypt%", "%file%", "%path%", "%invalid%"])
+  s.regexpMatch("(?i).*(password|passwd|accountid|account.?key|accnt.?key|license.?key|trusted).*") and
+  not s.regexpMatch("(?i).*(hash|crypt|file|path|invalid).*")
 }
 
 /**
@@ -23,7 +23,7 @@ private predicate suspicious(string s) {
  */
 class SensitiveVariable extends Variable {
   SensitiveVariable() {
-    suspicious(this.getName().toLowerCase()) and
+    suspicious(this.getName()) and
     not this.getUnspecifiedType() instanceof IntegralType
   }
 }
@@ -33,7 +33,7 @@ class SensitiveVariable extends Variable {
  */
 class SensitiveFunction extends Function {
   SensitiveFunction() {
-    suspicious(this.getName().toLowerCase()) and
+    suspicious(this.getName()) and
     not this.getUnspecifiedType() instanceof IntegralType
   }
 }

cpp/ql/src/CHANGELOG.md

@@ -1,3 +1,11 @@
+## 0.5.3
+
+No user-facing changes.
+
+## 0.5.2
+
+No user-facing changes.
+
 ## 0.5.1
 
 ### Minor Analysis Improvements

cpp/ql/src/Critical/MissingCheckScanf.ql

@@ -100,7 +100,7 @@ private predicate fwdFlow(Instruction instr, ValueNumber vn) {
  */
 pragma[nomagic]
 predicate revFlow(Instruction instr, ValueNumber vn) {
-  fwdFlow(instr, vn) and
+  fwdFlow(instr, pragma[only_bind_out](vn)) and
   (
     isSink(instr, _, vn)
     or
@@ -126,7 +126,7 @@ class Node extends MkNode {
 
   final string toString() { result = instr.toString() }
 
-  final Node getASuccessor() { result = MkNode(instr.getASuccessor(), vn) }
+  final Node getASuccessor() { result = MkNode(pragma[only_bind_out](instr.getASuccessor()), vn) }
 
   final Location getLocation() { result = instr.getLocation() }
 }
@@ -167,7 +167,7 @@ predicate hasFlow(
 ) {
   exists(ValueNumber vn |
     isSource(call, index, source, vn, _) and
-    hasFlow(getNode(source, vn), getNode(sink, vn)) and
+    hasFlow(getNode(source, pragma[only_bind_into](vn)), getNode(sink, pragma[only_bind_into](vn))) and
     isSink(sink, access, vn)
   )
 }

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

@@ -22,7 +22,7 @@ import semmle.code.cpp.ir.dataflow.TaintTracking
 import semmle.code.cpp.ir.dataflow.TaintTracking2
 import semmle.code.cpp.security.FlowSources
 import semmle.code.cpp.models.implementations.Strcat
-import DataFlow::PathGraph
+import ExecTaint::PathGraph
 
 Expr sinkAsArgumentIndirection(DataFlow::Node sink) {
   result =
@@ -67,28 +67,28 @@ predicate interestingConcatenation(DataFlow::Node fst, DataFlow::Node snd) {
   )
 }
 
-class ConcatState extends DataFlow::FlowState {
-  ConcatState() { this = "ConcatState" }
+newtype TState =
+  TConcatState() or
+  TExecState(DataFlow::Node fst, DataFlow::Node snd) { interestingConcatenation(fst, snd) }
+
+class ConcatState extends TConcatState {
+  string toString() { result = "ConcatState" }
 }
 
-class ExecState extends DataFlow::FlowState {
+class ExecState extends TExecState {
   DataFlow::Node fst;
   DataFlow::Node snd;
 
-  ExecState() {
-    this =
-      "ExecState (" + fst.getLocation() + " | " + fst + ", " + snd.getLocation() + " | " + snd + ")" and
-    interestingConcatenation(pragma[only_bind_into](fst), pragma[only_bind_into](snd))
-  }
+  ExecState() { this = TExecState(fst, snd) }
 
   DataFlow::Node getFstNode() { result = fst }
 
   DataFlow::Node getSndNode() { result = snd }
 
   /** Holds if this is a possible `ExecState` for `sink`. */
-  predicate isFeasibleForSink(DataFlow::Node sink) {
-    any(ExecStateConfiguration conf).hasFlow(snd, sink)
-  }
+  predicate isFeasibleForSink(DataFlow::Node sink) { ExecState::hasFlow(snd, sink) }
+
+  string toString() { result = "ExecState" }
 }
 
 /**
@@ -96,45 +96,42 @@ class ExecState extends DataFlow::FlowState {
  * given sink. This avoids a cartesian product between all sinks and all `ExecState`s in
  * `ExecTaintConfiguration::isSink`.
  */
-class ExecStateConfiguration extends TaintTracking2::Configuration {
-  ExecStateConfiguration() { this = "ExecStateConfiguration" }
-
-  override predicate isSource(DataFlow::Node source) {
+module ExecStateConfiguration implements DataFlow::ConfigSig {
+  predicate isSource(DataFlow::Node source) {
     exists(ExecState state | state.getSndNode() = source)
   }
 
-  override predicate isSink(DataFlow::Node sink) {
-    shellCommand(sinkAsArgumentIndirection(sink), _)
-  }
+  predicate isSink(DataFlow::Node sink) { shellCommand(sinkAsArgumentIndirection(sink), _) }
 
-  override predicate isSanitizerOut(DataFlow::Node node) {
-    isSink(node, _) // Prevent duplicates along a call chain, since `shellCommand` will include wrappers
+  predicate isBarrierOut(DataFlow::Node node) {
+    isSink(node) // Prevent duplicates along a call chain, since `shellCommand` will include wrappers
   }
 }
 
-class ExecTaintConfiguration extends TaintTracking::Configuration {
-  ExecTaintConfiguration() { this = "ExecTaintConfiguration" }
+module ExecState = TaintTracking::Make<ExecStateConfiguration>;
 
-  override predicate isSource(DataFlow::Node source, DataFlow::FlowState state) {
+module ExecTaintConfiguration implements DataFlow::StateConfigSig {
+  class FlowState = TState;
+
+  predicate isSource(DataFlow::Node source, FlowState state) {
     source instanceof FlowSource and
     state instanceof ConcatState
   }
 
-  override predicate isSink(DataFlow::Node sink, DataFlow::FlowState state) {
-    any(ExecStateConfiguration conf).isSink(sink) and
+  predicate isSink(DataFlow::Node sink, FlowState state) {
+    ExecStateConfiguration::isSink(sink) and
     state.(ExecState).isFeasibleForSink(sink)
   }
 
-  override predicate isAdditionalTaintStep(
-    DataFlow::Node node1, DataFlow::FlowState state1, DataFlow::Node node2,
-    DataFlow::FlowState state2
+  predicate isAdditionalFlowStep(
+    DataFlow::Node node1, FlowState state1, DataFlow::Node node2, FlowState state2
   ) {
     state1 instanceof ConcatState and
     state2.(ExecState).getFstNode() = node1 and
     state2.(ExecState).getSndNode() = node2
   }
 
-  override predicate isSanitizer(DataFlow::Node node, DataFlow::FlowState state) {
+  predicate isBarrier(DataFlow::Node node, FlowState state) {
     (
       node.asInstruction().getResultType() instanceof IntegralType
       or
@@ -143,16 +140,18 @@ class ExecTaintConfiguration extends TaintTracking::Configuration {
     state instanceof ConcatState
   }
 
-  override predicate isSanitizerOut(DataFlow::Node node) {
+  predicate isBarrierOut(DataFlow::Node node) {
     isSink(node, _) // Prevent duplicates along a call chain, since `shellCommand` will include wrappers
   }
 }
 
+module ExecTaint = TaintTracking::MakeWithState<ExecTaintConfiguration>;
+
 from
-  ExecTaintConfiguration conf, DataFlow::PathNode sourceNode, DataFlow::PathNode sinkNode,
-  string taintCause, string callChain, DataFlow::Node concatResult
+  ExecTaint::PathNode sourceNode, ExecTaint::PathNode sinkNode, string taintCause, string callChain,
+  DataFlow::Node concatResult
 where
-  conf.hasFlowPath(sourceNode, sinkNode) and
+  ExecTaint::hasFlowPath(sourceNode, sinkNode) and
   taintCause = sourceNode.getNode().(FlowSource).getSourceType() and
   shellCommand(sinkAsArgumentIndirection(sinkNode.getNode()), callChain) and
   concatResult = sinkNode.getState().(ExecState).getSndNode()

cpp/ql/src/Security/CWE/CWE-732/DoNotCreateWorldWritable.c

@@ -1,11 +1,11 @@
-int write_default_config_bad() {
+void write_default_config_bad() {
 	// BAD - this is world-writable so any user can overwrite the config
-	FILE* out = creat(OUTFILE, 0666);
-	fprintf(out, DEFAULT_CONFIG);
+	int out = creat(OUTFILE, 0666);
+	dprintf(out, DEFAULT_CONFIG);
 }
 
-int write_default_config_good() {
+void write_default_config_good() {
 	// GOOD - this allows only the current user to modify the file
-	FILE* out = creat(OUTFILE, S_IWUSR | S_IRUSR);
-	fprintf(out, DEFAULT_CONFIG);
+	int out = creat(OUTFILE, S_IWUSR | S_IRUSR);
+	dprintf(out, DEFAULT_CONFIG);
 }

cpp/ql/src/change-notes/released/0.5.2.md

@@ -0,0 +1,3 @@
+## 0.5.2
+
+No user-facing changes.

cpp/ql/src/change-notes/released/0.5.3.md

@@ -0,0 +1,3 @@
+## 0.5.3
+
+No user-facing changes.

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

@@ -1,2 +1,2 @@
 ---
-lastReleaseVersion: 0.5.1
+lastReleaseVersion: 0.5.3

cpp/ql/src/experimental/Security/CWE/CWE-369/DivideByZeroUsingReturnValue.cpp

@@ -0,0 +1,7 @@
+
+...
+  a = getc(f);
+  if (a < 123) ret = 123/a; // BAD
+...
+  if (a != 0) ret = 123/a; // GOOD
+...

cpp/ql/src/experimental/Security/CWE/CWE-369/DivideByZeroUsingReturnValue.qhelp

@@ -0,0 +1,23 @@
+<!DOCTYPE qhelp PUBLIC
+  "-//Semmle//qhelp//EN"
+  "qhelp.dtd">
+<qhelp>
+<overview>
+<p> Possible cases of division by zero when using the return value from functions.</p>
+
+</overview>
+
+<example>
+<p>The following example shows the use of a function with an error when using the return value and without an error.</p>
+<sample src="DivideByZeroUsingReturnValue.cpp" />
+
+</example>
+<references>
+
+<li>
+  CERT Coding Standard:
+  <a href="https://wiki.sei.cmu.edu/confluence/display/c/INT33-C.+Ensure+that+division+and+remainder+operations+do+not+result+in+divide-by-zero+errors">INT33-C. Ensure that division and remainder operations do not result in divide-by-zero errors - SEI CERT C Coding Standard - Confluence</a>.
+</li>
+
+</references>
+</qhelp>

cpp/ql/src/experimental/Security/CWE/CWE-369/DivideByZeroUsingReturnValue.ql

@@ -0,0 +1,274 @@
+/**
+ * @name Divide by zero using return value
+ * @description Possible cases of division by zero when using the return value from functions.
+ * @kind problem
+ * @id cpp/divide-by-zero-using-return-value
+ * @problem.severity warning
+ * @precision medium
+ * @tags correctness
+ *       security
+ *       external/cwe/cwe-369
+ */
+
+import cpp
+import semmle.code.cpp.valuenumbering.GlobalValueNumbering
+import semmle.code.cpp.controlflow.Guards
+
+/** Holds if function `fn` can return a value equal to value `val` */
+predicate mayBeReturnValue(Function fn, float val) {
+  exists(Expr tmpExp, ReturnStmt rs |
+    tmpExp.getValue().toFloat() = val and
+    rs.getEnclosingFunction() = fn and
+    (
+      globalValueNumber(rs.getExpr()) = globalValueNumber(tmpExp)
+      or
+      exists(AssignExpr ae |
+        ae.getLValue().(VariableAccess).getTarget() =
+          globalValueNumber(rs.getExpr()).getAnExpr().(VariableAccess).getTarget() and
+        globalValueNumber(ae.getRValue()) = globalValueNumber(tmpExp)
+      )
+      or
+      exists(Initializer it |
+        globalValueNumber(it.getExpr()) = globalValueNumber(tmpExp) and
+        it.getDeclaration().(Variable).getAnAccess().getTarget() =
+          globalValueNumber(rs.getExpr()).getAnExpr().(VariableAccess).getTarget()
+      )
+    )
+  )
+}
+
+/** Holds if function `fn` can return a value equal zero */
+predicate mayBeReturnZero(Function fn) {
+  mayBeReturnValue(fn, 0)
+  or
+  fn.hasName([
+      "iswalpha", "iswlower", "iswprint", "iswspace", "iswblank", "iswupper", "iswcntrl",
+      "iswctype", "iswalnum", "iswgraph", "iswxdigit", "iswdigit", "iswpunct", "isblank", "isupper",
+      "isgraph", "isalnum", "ispunct", "islower", "isspace", "isprint", "isxdigit", "iscntrl",
+      "isdigit", "isalpha", "timespec_get", "feof", "atomic_is_lock_free",
+      "atomic_compare_exchange", "thrd_equal", "isfinite", "islessequal", "isnan", "isgreater",
+      "signbit", "isinf", "islessgreater", "isnormal", "isless", "isgreaterequal", "isunordered",
+      "ferror"
+    ])
+  or
+  fn.hasName([
+      "thrd_sleep", "feenv", "feholdexcept", "feclearexcept", "feexceptflag", "feupdateenv",
+      "remove", "fflush", "setvbuf", "fgetpos", "fsetpos", "fclose", "rename", "fseek", "raise"
+    ])
+  or
+  fn.hasName(["tss_get", "gets"])
+  or
+  fn.hasName(["getc", "atoi"])
+}
+
+/** Gets the Guard which compares the expression `bound` */
+pragma[inline]
+GuardCondition checkByValue(Expr bound, Expr val) {
+  exists(GuardCondition gc |
+    (
+      gc.ensuresEq(bound, val, _, _, _) or
+      gc.ensuresEq(val, bound, _, _, _) or
+      gc.ensuresLt(bound, val, _, _, _) or
+      gc.ensuresLt(val, bound, _, _, _) or
+      gc = globalValueNumber(bound).getAnExpr()
+    ) and
+    result = gc
+  )
+}
+
+/** Holds if there are no comparisons between the value returned by possible function calls `compArg` and the value `valArg`, or when these comparisons do not exclude equality to the value `valArg`. */
+pragma[inline]
+predicate compareFunctionWithValue(Expr guardExp, Function compArg, Expr valArg) {
+  not exists(Expr exp |
+    exp.getAChild*() = globalValueNumber(compArg.getACallToThisFunction()).getAnExpr() and
+    checkByValue(exp, valArg).controls(guardExp.getBasicBlock(), _)
+  )
+  or
+  exists(GuardCondition gc |
+    (
+      gc.ensuresEq(globalValueNumber(compArg.getACallToThisFunction()).getAnExpr(), valArg, 0,
+        guardExp.getBasicBlock(), true)
+      or
+      gc.ensuresEq(valArg, globalValueNumber(compArg.getACallToThisFunction()).getAnExpr(), 0,
+        guardExp.getBasicBlock(), true)
+      or
+      gc.ensuresLt(globalValueNumber(compArg.getACallToThisFunction()).getAnExpr(), valArg, 0,
+        guardExp.getBasicBlock(), false)
+      or
+      gc.ensuresLt(valArg, globalValueNumber(compArg.getACallToThisFunction()).getAnExpr(), 0,
+        guardExp.getBasicBlock(), false)
+    )
+    or
+    exists(Expr exp |
+      exp.getValue().toFloat() > valArg.getValue().toFloat() and
+      gc.ensuresLt(globalValueNumber(compArg.getACallToThisFunction()).getAnExpr(), exp, 0,
+        guardExp.getBasicBlock(), true)
+      or
+      exp.getValue().toFloat() < valArg.getValue().toFloat() and
+      gc.ensuresLt(exp, globalValueNumber(compArg.getACallToThisFunction()).getAnExpr(), 0,
+        guardExp.getBasicBlock(), true)
+    )
+  )
+  or
+  valArg.getValue().toFloat() = 0 and
+  exists(NotExpr ne, IfStmt ifne |
+    ne.getOperand() = globalValueNumber(compArg.getACallToThisFunction()).getAnExpr() and
+    ifne.getCondition() = ne and
+    ifne.getThen().getAChild*() = guardExp
+  )
+}
+
+/** Wraping predicate for call `compareFunctionWithValue`. */
+pragma[inline]
+predicate checkConditions1(Expr div, Function fn, float changeInt) {
+  exists(Expr val |
+    val.getEnclosingFunction() = fn and
+    val.getValue().toFloat() = changeInt and
+    compareFunctionWithValue(div, fn, val)
+  )
+}
+
+/** Holds if there are no comparisons between the value `compArg` and the value `valArg`, or when these comparisons do not exclude equality to the value `valArg`. */
+pragma[inline]
+predicate compareExprWithValue(Expr guardExp, Expr compArg, Expr valArg) {
+  not exists(Expr exp |
+    exp.getAChild*() = globalValueNumber(compArg).getAnExpr() and
+    checkByValue(exp, valArg).controls(guardExp.getBasicBlock(), _)
+  )
+  or
+  exists(GuardCondition gc |
+    (
+      gc.ensuresEq(globalValueNumber(compArg).getAnExpr(), valArg, 0, guardExp.getBasicBlock(), true)
+      or
+      gc.ensuresEq(valArg, globalValueNumber(compArg).getAnExpr(), 0, guardExp.getBasicBlock(), true)
+      or
+      gc.ensuresLt(globalValueNumber(compArg).getAnExpr(), valArg, 0, guardExp.getBasicBlock(),
+        false)
+      or
+      gc.ensuresLt(valArg, globalValueNumber(compArg).getAnExpr(), 0, guardExp.getBasicBlock(),
+        false)
+    )
+    or
+    exists(Expr exp |
+      exp.getValue().toFloat() > valArg.getValue().toFloat() and
+      gc.ensuresLt(globalValueNumber(compArg).getAnExpr(), exp, 0, guardExp.getBasicBlock(), true)
+      or
+      exp.getValue().toFloat() < valArg.getValue().toFloat() and
+      gc.ensuresLt(exp, globalValueNumber(compArg).getAnExpr(), 0, guardExp.getBasicBlock(), true)
+    )
+  )
+  or
+  valArg.getValue().toFloat() = 0 and
+  exists(NotExpr ne, IfStmt ifne |
+    ne.getOperand() = globalValueNumber(compArg).getAnExpr() and
+    ifne.getCondition() = ne and
+    ifne.getThen().getAChild*() = guardExp
+  )
+}
+
+/** Wraping predicate for call `compareExprWithValue`. */
+pragma[inline]
+predicate checkConditions2(Expr div, Expr divVal, float changeInt2) {
+  exists(Expr val |
+    (
+      val.getEnclosingFunction() =
+        div.getEnclosingFunction().getACallToThisFunction().getEnclosingFunction() or
+      val.getEnclosingFunction() = div.getEnclosingFunction()
+    ) and
+    val.getValue().toFloat() = changeInt2 and
+    compareExprWithValue(div, divVal, val)
+  )
+}
+
+/** Gets the value of the difference or summand from the expression `src`. */
+float getValueOperand(Expr src, Expr e1, Expr e2) {
+  src.(SubExpr).hasOperands(e1, e2) and
+  result = e2.getValue().toFloat()
+  or
+  src.(AddExpr).hasOperands(e1, e2) and
+  result = -e2.getValue().toFloat()
+}
+
+/** Function the return of the expression `e1` and the multiplication operands, or the left operand of division if `e1` contains a multiplication or division, respectively. */
+Expr getMulDivOperand(Expr e1) {
+  result = e1 or
+  result = e1.(MulExpr).getAnOperand() or
+  result = e1.(DivExpr).getLeftOperand()
+}
+
+/** The class that defines possible variants of the division expression or the search for the remainder. */
+class MyDiv extends Expr {
+  MyDiv() {
+    this instanceof DivExpr or
+    this instanceof RemExpr or
+    this instanceof AssignDivExpr or
+    this instanceof AssignRemExpr
+  }
+
+  Expr getRV() {
+    result = this.(AssignArithmeticOperation).getRValue() or
+    result = this.(BinaryArithmeticOperation).getRightOperand()
+  }
+}
+
+from Expr exp, string msg, Function fn, GVN findVal, float changeInt, MyDiv div
+where
+  findVal = globalValueNumber(fn.getACallToThisFunction()) and
+  (
+    // Look for divide-by-zero operations possible due to the return value of the function `fn`.
+    checkConditions1(div, fn, changeInt) and
+    (
+      // Function return value can be zero.
+      mayBeReturnZero(fn) and
+      getMulDivOperand(globalValueNumber(div.getRV()).getAnExpr()) = findVal.getAnExpr() and
+      changeInt = 0
+      or
+      // Denominator can be sum or difference.
+      changeInt = getValueOperand(div.getRV(), findVal.getAnExpr(), _) and
+      mayBeReturnValue(fn, changeInt)
+    ) and
+    exp = div and
+    msg =
+      "Can lead to division by 0, since the function " + fn.getName() + " can return a value " +
+        changeInt.toString() + "."
+    or
+    // Search for situations where division by zero is possible inside the `divFn` function if the passed argument can be equal to a certain value.
+    exists(int posArg, Expr divVal, FunctionCall divFc, float changeInt2 |
+      // Division is associated with the function argument.
+      exists(Function divFn |
+        divFn.getParameter(posArg).getAnAccess() = divVal and
+        divVal.getEnclosingStmt() = div.getEnclosingStmt() and
+        divFc = divFn.getACallToThisFunction()
+      ) and
+      (
+        divVal = div.getRV() and
+        divFc.getArgument(posArg) != findVal.getAnExpr() and
+        (
+          // Function return value can be zero.
+          mayBeReturnZero(fn) and
+          getMulDivOperand(globalValueNumber(divFc.getArgument(posArg)).getAnExpr()) =
+            findVal.getAnExpr() and
+          changeInt = 0 and
+          changeInt2 = 0
+          or
+          // Denominator can be sum or difference.
+          changeInt = getValueOperand(divFc.getArgument(posArg), findVal.getAnExpr(), _) and
+          mayBeReturnValue(fn, changeInt) and
+          changeInt2 = 0
+        )
+        or
+        // Look for a situation where the difference or subtraction is considered as an argument, and it can be used in the same way.
+        changeInt = getValueOperand(div.getRV(), divVal, _) and
+        changeInt2 = changeInt and
+        mayBeReturnValue(fn, changeInt) and
+        divFc.getArgument(posArg) = findVal.getAnExpr()
+      ) and
+      checkConditions2(div, divVal, changeInt2) and
+      checkConditions1(divFc, fn, changeInt) and
+      exp = divFc and
+      msg =
+        "Can lead to division by 0, since the function " + fn.getName() + " can return a value " +
+          changeInt.toString() + "."
+    )
+  )
+select exp, msg

cpp/ql/src/experimental/Security/CWE/CWE-415/DoubleFree.ql

@@ -14,8 +14,8 @@ import cpp
 
 from FunctionCall fc, FunctionCall fc2, LocalScopeVariable v
 where
-  freeCall(fc, v.getAnAccess()) and
-  freeCall(fc2, v.getAnAccess()) and
+  fc.(DeallocationExpr).getFreedExpr() = v.getAnAccess() and
+  fc2.(DeallocationExpr).getFreedExpr() = v.getAnAccess() and
   fc != fc2 and
   fc.getASuccessor*() = fc2 and
   not exists(Expr exptmp |

cpp/ql/src/experimental/Security/CWE/CWE-805/BufferAccessWithIncorrectLengthValue.cpp

@@ -0,0 +1,8 @@
+
+...
+char buf[256];
+X509_NAME_oneline(X509_get_subject_name(peer),buf,sizeof(buf)); // GOOD
+...
+char buf[256];
+X509_NAME_oneline(X509_get_subject_name(peer),buf,1024); // BAD
+...

cpp/ql/src/experimental/Security/CWE/CWE-805/BufferAccessWithIncorrectLengthValue.qhelp

@@ -0,0 +1,23 @@
+<!DOCTYPE qhelp PUBLIC
+  "-//Semmle//qhelp//EN"
+  "qhelp.dtd">
+<qhelp>
+<overview>
+<p>Using a size argument that is larger than the buffer size will result in an out-of-bounds memory access and possibly overflow. You need to limit the value of the length argument.</p>
+
+</overview>
+
+<example>
+<p>The following example shows the use of a function with and without an error in the size argument.</p>
+<sample src="BufferAccessWithIncorrectLengthValue.cpp" />
+
+</example>
+<references>
+
+<li>
+  CERT Coding Standard:
+  <a href="https://wiki.sei.cmu.edu/confluence/display/c/ARR38-C.+Guarantee+that+library+functions+do+not+form+invalid+pointers">ARR38-C. Guarantee that library functions do not form invalid pointers - SEI CERT C Coding Standard - Confluence</a>.
+</li>
+
+</references>
+</qhelp>

cpp/ql/src/experimental/Security/CWE/CWE-805/BufferAccessWithIncorrectLengthValue.ql

@@ -0,0 +1,76 @@
+/**
+ * @name Buffer access with incorrect length value
+ * @description Incorrect use of the length argument in some functions will result in out-of-memory accesses.
+ * @kind problem
+ * @id cpp/buffer-access-with-incorrect-length-value
+ * @problem.severity warning
+ * @precision medium
+ * @tags correctness
+ *       security
+ *       experimental
+ *       external/cwe/cwe-805
+ */
+
+import cpp
+
+/** Holds for a function `f`, which has an argument at index `bpos` that points to a buffer and an argument at index `spos` that points to a size. */
+predicate numberArgument(Function f, int bpos, int spos) {
+  f.hasGlobalOrStdName([
+      "X509_NAME_oneline", "SSL_CIPHER_description", "SSL_get_shared_ciphers",
+      "SSL_export_keying_material_early", "SSL_export_keying_material", "SSL_set_alpn_protos",
+      "SSL_CTX_set_alpn_protos", "SSL_read", "SSL_read_ex", "SSL_read_early_data",
+      "SSL_bytes_to_cipher_list", "SSL_write", "SSL_SESSION_set1_master_key",
+      "SSL_CTX_set_session_id_context", "BIO_gets", "BIO_read", "BIO_read_ex", "BIO_write",
+      "BIO_write_ex", "BIO_ctrl", "BN_bn2binpad", "BN_signed_bn2bin", "BN_signed_bn2lebin",
+      "EVP_PKEY_get_default_digest_name", "EVP_DigestUpdate", "EVP_PKEY_CTX_set1_tls1_prf_secret",
+      "EVP_KDF_derive", "EVP_CIPHER_CTX_get_updated_iv", "EVP_PKEY_get_group_name", "EVP_MAC_init",
+      "write", "read", "send", "sendto", "recv", "recvfrom", "strerror_r"
+    ]) and
+  bpos = 1 and
+  spos = 2
+  or
+  f.hasGlobalOrStdName(["X509_NAME_get_text_by_NID", "EVP_PKEY_get_utf8_string_param"]) and
+  bpos = 2 and
+  spos = 3
+  or
+  f.hasGlobalOrStdName([
+      "BIO_snprintf", "BN_signed_lebin2bn", "BIO_new_mem_buf", "BN_lebin2bn", "BN_bin2bn",
+      "EVP_read_pw_string", "EVP_read_pw_string", "strftime", "strnlen", "fgets", "snprintf",
+      "vsnprintf"
+    ]) and
+  bpos = 0 and
+  spos = 1
+  or
+  f.hasGlobalOrStdName(["AES_ige_encrypt", "memchr"]) and bpos = 0 and spos = 2
+  or
+  f.hasGlobalOrStdName("EVP_MAC_final") and bpos = 1 and spos = 3
+  or
+  f.hasGlobalOrStdName("OBJ_obj2txt") and bpos = 2 and spos = 1
+  or
+  f.hasGlobalOrStdName("EVP_CIPHER_CTX_ctrl") and bpos = 3 and spos = 2
+  or
+  f.hasGlobalOrStdName(["EVP_PKEY_get_octet_string_param", "getnameinfo"]) and bpos = 2 and spos = 3
+  or
+  f.hasGlobalOrStdName([
+      "EVP_DecryptUpdate", "EVP_EncryptUpdate", "EVP_PKEY_encrypt", "EVP_PKEY_sign",
+      "EVP_CipherUpdate"
+    ]) and
+  bpos = 3 and
+  spos = 4
+  or
+  f.hasGlobalOrStdName("getnameinfo") and bpos = 4 and spos = 5
+}
+
+from FunctionCall fc
+where
+  exists(ArrayType array, int bufArgPos, int sizeArgPos |
+    numberArgument(fc.getTarget(), bufArgPos, sizeArgPos) and
+    fc.getArgument(pragma[only_bind_into](sizeArgPos)).getValue().toInt() > array.getByteSize() and
+    fc.getArgument(pragma[only_bind_into](bufArgPos))
+        .(VariableAccess)
+        .getTarget()
+        .getADeclarationEntry()
+        .getType() = array
+  )
+select fc,
+  "Access beyond the bounds of the allocated memory is possible, the size argument used is greater than the size of the buffer."

cpp/ql/src/qlpack.yml

@@ -1,5 +1,5 @@
 name: codeql/cpp-queries
-version: 0.5.1
+version: 0.5.4-dev
 groups: 
   - cpp
   - queries

cpp/ql/test/experimental/library-tests/rangeanalysis/bitshift/BitShiftRange.expected

@@ -0,0 +1,44 @@
+| bitshift.cpp:23:3:23:9 | ... <<= ... | 0.0 | 255.0 | file://:0:0:0:0 | unsigned char | file://:0:0:0:0 | int | file://:0:0:0:0 | unsigned char | file://:0:0:0:0 | int |
+| bitshift.cpp:25:5:25:11 | ... <<= ... | 0.0 | 240.0 | file://:0:0:0:0 | unsigned char | file://:0:0:0:0 | int | file://:0:0:0:0 | unsigned char | file://:0:0:0:0 | int |
+| bitshift.cpp:29:3:29:8 | ... << ... | 0.0 | 1020.0 | file://:0:0:0:0 | unsigned char | file://:0:0:0:0 | int | file://:0:0:0:0 | int | file://:0:0:0:0 | int |
+| bitshift.cpp:32:3:32:9 | ... << ... | -2.147483648E9 | 2.147483647E9 | file://:0:0:0:0 | unsigned char | file://:0:0:0:0 | int | file://:0:0:0:0 | int | file://:0:0:0:0 | int |
+| bitshift.cpp:35:3:35:9 | ... << ... | -2.147483648E9 | 2.147483647E9 | file://:0:0:0:0 | unsigned char | file://:0:0:0:0 | int | file://:0:0:0:0 | int | file://:0:0:0:0 | int |
+| bitshift.cpp:38:3:38:22 | ... << ... | 0.0 | 32640.0 | file://:0:0:0:0 | unsigned char | file://:0:0:0:0 | unsigned char | file://:0:0:0:0 | int | file://:0:0:0:0 | int |
+| bitshift.cpp:39:3:39:22 | ... << ... | 0.0 | 32640.0 | file://:0:0:0:0 | unsigned char | file://:0:0:0:0 | unsigned char | file://:0:0:0:0 | int | file://:0:0:0:0 | int |
+| bitshift.cpp:40:3:40:22 | ... << ... | 0.0 | 32640.0 | file://:0:0:0:0 | unsigned char | file://:0:0:0:0 | unsigned char | file://:0:0:0:0 | int | file://:0:0:0:0 | int |
+| bitshift.cpp:43:3:43:19 | ... << ... | -2.147483648E9 | 2.147483647E9 | file://:0:0:0:0 | unsigned char | file://:0:0:0:0 | signed char | file://:0:0:0:0 | int | file://:0:0:0:0 | int |
+| bitshift.cpp:46:3:46:22 | ... << ... | 128.0 | 128.0 | file://:0:0:0:0 | int | file://:0:0:0:0 | unsigned char | file://:0:0:0:0 | int | file://:0:0:0:0 | int |
+| bitshift.cpp:49:3:49:8 | ... << ... | -2.147483648E9 | 2.147483647E9 | file://:0:0:0:0 | int | file://:0:0:0:0 | unsigned char | file://:0:0:0:0 | int | file://:0:0:0:0 | int |
+| bitshift.cpp:52:5:52:10 | ... << ... | 1.0 | 128.0 | file://:0:0:0:0 | int | file://:0:0:0:0 | unsigned char | file://:0:0:0:0 | int | file://:0:0:0:0 | int |
+| bitshift.cpp:57:3:57:8 | ... << ... | -2.147483648E9 | 2.147483647E9 | file://:0:0:0:0 | signed char | file://:0:0:0:0 | int | file://:0:0:0:0 | int | file://:0:0:0:0 | int |
+| bitshift.cpp:58:3:58:9 | ... << ... | -2.147483648E9 | 2.147483647E9 | file://:0:0:0:0 | signed char | file://:0:0:0:0 | int | file://:0:0:0:0 | int | file://:0:0:0:0 | int |
+| bitshift.cpp:59:3:59:9 | ... << ... | -2.147483648E9 | 2.147483647E9 | file://:0:0:0:0 | signed char | file://:0:0:0:0 | int | file://:0:0:0:0 | int | file://:0:0:0:0 | int |
+| bitshift.cpp:60:3:60:22 | ... << ... | -2.147483648E9 | 2.147483647E9 | file://:0:0:0:0 | signed char | file://:0:0:0:0 | unsigned char | file://:0:0:0:0 | int | file://:0:0:0:0 | int |
+| bitshift.cpp:61:3:61:19 | ... << ... | -2.147483648E9 | 2.147483647E9 | file://:0:0:0:0 | signed char | file://:0:0:0:0 | signed char | file://:0:0:0:0 | int | file://:0:0:0:0 | int |
+| bitshift.cpp:64:3:64:19 | ... << ... | -2.147483648E9 | 2.147483647E9 | file://:0:0:0:0 | int | file://:0:0:0:0 | signed char | file://:0:0:0:0 | int | file://:0:0:0:0 | int |
+| bitshift.cpp:67:3:67:8 | ... << ... | -2.147483648E9 | 2.147483647E9 | file://:0:0:0:0 | int | file://:0:0:0:0 | signed char | file://:0:0:0:0 | int | file://:0:0:0:0 | int |
+| bitshift.cpp:70:5:70:10 | ... << ... | 1.0 | 128.0 | file://:0:0:0:0 | int | file://:0:0:0:0 | signed char | file://:0:0:0:0 | int | file://:0:0:0:0 | int |
+| bitshift.cpp:75:5:75:10 | ... << ... | -2.147483648E9 | 2.147483647E9 | file://:0:0:0:0 | unsigned char | file://:0:0:0:0 | signed char | file://:0:0:0:0 | int | file://:0:0:0:0 | int |
+| bitshift.cpp:76:5:76:10 | ... << ... | -2.147483648E9 | 2.147483647E9 | file://:0:0:0:0 | signed char | file://:0:0:0:0 | unsigned char | file://:0:0:0:0 | int | file://:0:0:0:0 | int |
+| bitshift.cpp:90:3:90:9 | ... >>= ... | 0.0 | 63.0 | file://:0:0:0:0 | unsigned char | file://:0:0:0:0 | int | file://:0:0:0:0 | unsigned char | file://:0:0:0:0 | int |
+| bitshift.cpp:92:5:92:11 | ... >>= ... | 0.0 | 15.0 | file://:0:0:0:0 | unsigned char | file://:0:0:0:0 | int | file://:0:0:0:0 | unsigned char | file://:0:0:0:0 | int |
+| bitshift.cpp:96:3:96:8 | ... >> ... | 0.0 | 63.0 | file://:0:0:0:0 | unsigned char | file://:0:0:0:0 | int | file://:0:0:0:0 | int | file://:0:0:0:0 | int |
+| bitshift.cpp:99:3:99:9 | ... >> ... | 0.0 | 0.0 | file://:0:0:0:0 | unsigned char | file://:0:0:0:0 | int | file://:0:0:0:0 | int | file://:0:0:0:0 | int |
+| bitshift.cpp:103:3:103:9 | ... >> ... | 0.0 | 0.0 | file://:0:0:0:0 | unsigned char | file://:0:0:0:0 | int | file://:0:0:0:0 | int | file://:0:0:0:0 | int |
+| bitshift.cpp:106:3:106:22 | ... >> ... | 0.0 | 63.0 | file://:0:0:0:0 | unsigned char | file://:0:0:0:0 | unsigned char | file://:0:0:0:0 | int | file://:0:0:0:0 | int |
+| bitshift.cpp:107:3:107:22 | ... >> ... | 0.0 | 63.0 | file://:0:0:0:0 | unsigned char | file://:0:0:0:0 | unsigned char | file://:0:0:0:0 | int | file://:0:0:0:0 | int |
+| bitshift.cpp:108:3:108:22 | ... >> ... | 0.0 | 63.0 | file://:0:0:0:0 | unsigned char | file://:0:0:0:0 | unsigned char | file://:0:0:0:0 | int | file://:0:0:0:0 | int |
+| bitshift.cpp:111:3:111:19 | ... >> ... | -2.147483648E9 | 2.147483647E9 | file://:0:0:0:0 | unsigned char | file://:0:0:0:0 | signed char | file://:0:0:0:0 | int | file://:0:0:0:0 | int |
+| bitshift.cpp:114:3:114:24 | ... >> ... | 32.0 | 32.0 | file://:0:0:0:0 | int | file://:0:0:0:0 | unsigned char | file://:0:0:0:0 | int | file://:0:0:0:0 | int |
+| bitshift.cpp:117:3:117:10 | ... >> ... | -2.147483648E9 | 2.147483647E9 | file://:0:0:0:0 | int | file://:0:0:0:0 | unsigned char | file://:0:0:0:0 | int | file://:0:0:0:0 | int |
+| bitshift.cpp:120:5:120:12 | ... >> ... | 32.0 | 128.0 | file://:0:0:0:0 | int | file://:0:0:0:0 | unsigned char | file://:0:0:0:0 | int | file://:0:0:0:0 | int |
+| bitshift.cpp:126:3:126:8 | ... >> ... | -2.147483648E9 | 2.147483647E9 | file://:0:0:0:0 | signed char | file://:0:0:0:0 | int | file://:0:0:0:0 | int | file://:0:0:0:0 | int |
+| bitshift.cpp:127:3:127:9 | ... >> ... | -2.147483648E9 | 2.147483647E9 | file://:0:0:0:0 | signed char | file://:0:0:0:0 | int | file://:0:0:0:0 | int | file://:0:0:0:0 | int |
+| bitshift.cpp:128:3:128:9 | ... >> ... | -1.0 | 0.0 | file://:0:0:0:0 | signed char | file://:0:0:0:0 | int | file://:0:0:0:0 | int | file://:0:0:0:0 | int |
+| bitshift.cpp:129:3:129:22 | ... >> ... | -2.147483648E9 | 2.147483647E9 | file://:0:0:0:0 | signed char | file://:0:0:0:0 | unsigned char | file://:0:0:0:0 | int | file://:0:0:0:0 | int |
+| bitshift.cpp:130:3:130:19 | ... >> ... | -2.147483648E9 | 2.147483647E9 | file://:0:0:0:0 | signed char | file://:0:0:0:0 | signed char | file://:0:0:0:0 | int | file://:0:0:0:0 | int |
+| bitshift.cpp:133:3:133:21 | ... >> ... | -2.147483648E9 | 2.147483647E9 | file://:0:0:0:0 | int | file://:0:0:0:0 | signed char | file://:0:0:0:0 | int | file://:0:0:0:0 | int |
+| bitshift.cpp:136:3:136:10 | ... >> ... | -2.147483648E9 | 2.147483647E9 | file://:0:0:0:0 | int | file://:0:0:0:0 | signed char | file://:0:0:0:0 | int | file://:0:0:0:0 | int |
+| bitshift.cpp:139:5:139:12 | ... >> ... | 32.0 | 128.0 | file://:0:0:0:0 | int | file://:0:0:0:0 | signed char | file://:0:0:0:0 | int | file://:0:0:0:0 | int |
+| bitshift.cpp:144:5:144:10 | ... >> ... | -2.147483648E9 | 2.147483647E9 | file://:0:0:0:0 | unsigned char | file://:0:0:0:0 | signed char | file://:0:0:0:0 | int | file://:0:0:0:0 | int |
+| bitshift.cpp:145:5:145:10 | ... >> ... | -2.147483648E9 | 2.147483647E9 | file://:0:0:0:0 | signed char | file://:0:0:0:0 | unsigned char | file://:0:0:0:0 | int | file://:0:0:0:0 | int |

cpp/ql/test/experimental/library-tests/rangeanalysis/bitshift/BitShiftRange.ql

@@ -0,0 +1,24 @@
+import cpp
+import semmle.code.cpp.rangeanalysis.SimpleRangeAnalysis
+import experimental.semmle.code.cpp.rangeanalysis.extensions.ConstantShiftExprRange
+
+Expr getLOp(Operation o) {
+  result = o.(BinaryOperation).getLeftOperand() or
+  result = o.(Assignment).getLValue()
+}
+
+Expr getROp(Operation o) {
+  result = o.(BinaryOperation).getRightOperand() or
+  result = o.(Assignment).getRValue()
+}
+
+from Operation o
+where
+  (
+    o instanceof BinaryBitwiseOperation
+    or
+    o instanceof AssignBitwiseOperation
+  )
+select o, lowerBound(o), upperBound(o), getLOp(o).getUnderlyingType(),
+  getROp(o).getUnderlyingType(), getLOp(o).getFullyConverted().getUnderlyingType(),
+  getROp(o).getFullyConverted().getUnderlyingType()

cpp/ql/test/experimental/library-tests/rangeanalysis/bitshift/bitshift.cpp

@@ -0,0 +1,147 @@
+typedef signed char int8_t;
+typedef short int16_t;
+typedef int int32_t;
+typedef long int64_t;
+
+typedef unsigned char uint8_t;
+typedef unsigned short uint16_t;
+typedef unsigned int uint32_t;
+typedef unsigned long uint64_t;
+
+extern uint8_t value_known_at_runtime8();
+
+void testLShiftOperator() {
+  uint8_t unsigned_const1 = 7;
+  uint8_t unsigned_const2(7);
+  uint8_t unsigned_const3{7};
+  int8_t signed_const = -7;
+  uint8_t x = value_known_at_runtime8();
+  int8_t y = (int8_t)value_known_at_runtime8();
+  uint8_t z = value_known_at_runtime8();
+
+  // An assign left shift operator. Note that no promotion occurs here
+  z <<= 2;  // [0, 255]
+  if (z <= 60) {
+    z <<= 2;  // [0, 240]
+  }
+
+  // A normal shift
+  x << 2;  // [0, 1020]
+
+  // Possible to exceed the maximum size
+  x << 25;  // [-2147483648, 2147483648]
+
+  // Undefined behavior
+  x << 34;  // [-2147483648, 2147483648]
+
+  // A normal shift by a constant in a variable
+  x << unsigned_const1;  // [0, 32640]
+  x << unsigned_const2;  // [0, 32640]
+  x << unsigned_const3;  // [0, 32640]
+
+  // Negative shifts are undefined
+  x << signed_const;  // [-2147483648, 2147483648]
+
+  // Now the left operand is a constant
+  1 << unsigned_const1;  // [128, 128]
+
+  // x could be large enough to cause undefined behavior
+  1 << x;  // [-2147483648, 2147483647]
+  if (x < 8) {
+    // x is now constrained so the shift is defined
+    1 << x;  // [1, 128]
+  }
+
+  // We don't support shifting negative values (and some of these are undefined
+  // anyway)
+  y << 2;                // [-2147483648, 2147483647]
+  y << 25;               // [-2147483648, 2147483648]
+  y << 34;               // [-2147483648, 2147483648]
+  y << unsigned_const1;  // [-2147483648, 2147483647]
+  y << signed_const;     // [-2147483648, 2147483648]
+
+  // Negative shifts are undefined
+  1 << signed_const;  // [-2147483648, 2147483648]
+
+  // We don't handle cases where the shift range could be negative
+  1 << y;  // [-2147483648, 2147483648]
+  if (y >= 0 && y < 8) {
+    // The shift range is now positive
+    1 << y;  // [1, 128]
+  }
+
+  if (x > 0 and x < 2 and y > 0 and x < 2) {
+    // We don't support shifts where neither operand is a constant at the moment
+    x << y;  // [-2147483648, 2147483648]
+    y << x;  // [-2147483648, 2147483648]
+  }
+}
+
+void testRShiftOperator() {
+  uint8_t unsigned_const1 = 2;
+  uint8_t unsigned_const2(2);
+  uint8_t unsigned_const3{2};
+  int8_t signed_const = -2;
+  uint8_t x = value_known_at_runtime8();
+  int8_t y = (int8_t)value_known_at_runtime8();
+  uint8_t z = value_known_at_runtime8();
+
+  // An assign right shift operator. Note that no promotion occurs here
+  z >>= 2;  // [0, 63]
+  if (z <= 60) {
+    z >>= 2;  // [0, 15]
+  }
+
+  // A normal shift
+  x >> 2;  // [0, 63]
+
+  // Possible to exceed the maximum size
+  x >> 25;  // [0, 0]
+
+  // Undefined behavior, but this case is handled by the SimpleRangeAnalysis
+  // library and sets the the bounds to [0, 0], which is fine
+  x >> 34;  // [0, 0]
+
+  // A normal shift by a constant in a variable
+  x >> unsigned_const1;  // [0, 63]
+  x >> unsigned_const2;  // [0, 63]
+  x >> unsigned_const3;  // [0, 63]
+
+  // Negative shifts are undefined
+  x >> signed_const;  // [-2147483648, 2147483648]
+
+  // Now the left operand is a constant
+  128 >> unsigned_const1;  // [32, 32]
+
+  // x could be large enough to cause undefined behavior
+  128 >> x;  // [-2147483648, 2147483647]
+  if (x < 3) {
+    // x is now constrained so the shift is defined
+    128 >> x;  // [32, 128]
+  }
+
+  // We don't support shifting negative values, but the SimpleRangeAnalysis
+  // library handles the first three cases even though they're implementation
+  // defined or undefined behavior (TODO: Check ideone)
+  y >> 2;   // [-2147483648, 2147483647] (Default is [-32, 31])
+  y >> 25;  // -2147483648, 2147483647] (Default is [-1, 0])
+  y >> 34;  // [-1, 0] (My code doesn't touch this, so default code is used)
+  y >> unsigned_const1;  // [-2147483648, 2147483647]
+  y >> signed_const;     // [-2147483648, 2147483648]
+
+  // Negative shifts are undefined
+  128 >> signed_const;  // [-2147483648, 2147483648]
+
+  // We don't handle cases where the shift range could be negative
+  128 >> y;  // [-2147483648, 2147483648]
+  if (y >= 0 && y < 3) {
+    // The shift range is now positive
+    128 >> y;  // [32, 128]
+  }
+
+  if (x > 0 and x < 2 and y > 0 and x < 2) {
+    // We don't support shifts where neither operand is a constant at the moment
+    x >> y;  // [-2147483648, 2147483648]
+    y >> x;  // [-2147483648, 2147483648]
+  }
+}

cpp/ql/test/experimental/query-tests/Security/CWE/CWE-369/semmle/tests/DivideByZeroUsingReturnValue.expected

@@ -0,0 +1,27 @@
+| test.cpp:47:24:47:31 | ... / ... | Can lead to division by 0, since the function getSize can return a value 0. |
+| test.cpp:48:15:48:34 | ... / ... | Can lead to division by 0, since the function getSize2 can return a value 0. |
+| test.cpp:53:10:53:17 | ... / ... | Can lead to division by 0, since the function getSize can return a value 0. |
+| test.cpp:65:15:65:22 | ... / ... | Can lead to division by 0, since the function getSize can return a value 0. |
+| test.cpp:68:15:68:22 | ... / ... | Can lead to division by 0, since the function getSize can return a value 0. |
+| test.cpp:71:9:71:16 | ... / ... | Can lead to division by 0, since the function getSize can return a value 0. |
+| test.cpp:74:9:74:16 | ... / ... | Can lead to division by 0, since the function getSize can return a value 0. |
+| test.cpp:77:21:77:28 | ... / ... | Can lead to division by 0, since the function getSize can return a value 0. |
+| test.cpp:79:25:79:32 | ... / ... | Can lead to division by 0, since the function getSize can return a value 0. |
+| test.cpp:81:24:81:31 | ... / ... | Can lead to division by 0, since the function getSize can return a value 0. |
+| test.cpp:128:10:128:16 | ... / ... | Can lead to division by 0, since the function getSize can return a value 0. |
+| test.cpp:135:10:135:16 | ... / ... | Can lead to division by 0, since the function getSize can return a value 0. |
+| test.cpp:141:10:141:23 | ... / ... | Can lead to division by 0, since the function getSize can return a value 0. |
+| test.cpp:153:12:153:19 | ... / ... | Can lead to division by 0, since the function getSize can return a value 0. |
+| test.cpp:172:3:172:12 | ... /= ... | Can lead to division by 0, since the function getSize can return a value 0. |
+| test.cpp:173:3:173:12 | ... %= ... | Can lead to division by 0, since the function getSize can return a value 0. |
+| test.cpp:187:10:187:17 | ... / ... | Can lead to division by 0, since the function getSizeFloat can return a value 0. |
+| test.cpp:199:12:199:25 | ... / ... | Can lead to division by 0, since the function getSize can return a value -1. |
+| test.cpp:202:12:202:25 | ... / ... | Can lead to division by 0, since the function getSize can return a value 1. |
+| test.cpp:205:10:205:23 | ... / ... | Can lead to division by 0, since the function getSize can return a value 1. |
+| test.cpp:210:10:210:23 | ... / ... | Can lead to division by 0, since the function getSize can return a value 3. |
+| test.cpp:258:3:258:10 | call to badMyDiv | Can lead to division by 0, since the function getSize can return a value 0. |
+| test.cpp:259:3:259:10 | call to badMyDiv | Can lead to division by 0, since the function getSize can return a value 2. |
+| test.cpp:260:3:260:13 | call to badMySubDiv | Can lead to division by 0, since the function getSize can return a value 3. |
+| test.cpp:263:5:263:15 | call to badMySubDiv | Can lead to division by 0, since the function getSize can return a value 3. |
+| test.cpp:273:5:273:12 | call to badMyDiv | Can lead to division by 0, since the function getSize can return a value 3. |
+| test.cpp:275:5:275:12 | call to badMyDiv | Can lead to division by 0, since the function getSize can return a value -1. |