Unified: Extract built-in Swift types

Author: hvitved

unified/BUILD.bazel

@@ -45,15 +45,24 @@ codeql_pkg_files(
         otherwise = ["//unified/extractor"],
         win64 = ["//unified/extractor-unsupported-os:extractor"],
     ),
-    prefix = "tools/{CODEQL_PLATFORM}",
+    prefix = "{CODEQL_PLATFORM}",
+)
+
+pkg_filegroup(
+    name = "tools",
+    srcs = [
+        ":extractor-arch",
+        "//unified/tools",
+        "//unified/tools/builtins",
+    ],
+    prefix = "tools",
 )
 
 codeql_pack(
     name = "unified",
     srcs = [
         ":codeql-extractor-yml",
         ":dbscheme-group",
-        ":extractor-arch",
-        "//unified/tools",
+        ":tools",
     ],
 )

unified/extractor/src/extractor.rs

@@ -1,10 +1,10 @@
-use clap::Args;
-use std::path::PathBuf;
-
 use crate::languages;
+use clap::Args;
 use codeql_extractor::extractor::desugaring;
 use codeql_extractor::trap;
-
+use std::path::Path;
+use std::path::PathBuf;
+use std::{env, fs};
 #[derive(Args)]
 pub struct Options {
     /// Sets a custom source archive folder
@@ -31,6 +31,24 @@ pub fn run(options: Options) -> std::io::Result<()> {
         lang.prefix = "unified";
     }
 
+    let builtins_path = env::var("CODEQL_EXTRACTOR_UNIFIED_ROOT")
+        .map(|path| Path::new(&path).join("tools").join("builtins"))
+        .expect("failed to read CODEQL_EXTRACTOR_UNIFIED_ROOT environment variable");
+    let builtins_dir = fs::read_dir(builtins_path).expect("failed to read builtins directory");
+    let mut file_list = fs::OpenOptions::new()
+        .append(true)
+        .open(&options.file_list)
+        .expect("failed to open file list");
+    for entry in builtins_dir {
+        let entry = entry.expect("failed to read builtins directory");
+        let path = entry.path();
+        if path.extension().is_some_and(|ext| ext == "swift") {
+            use std::io::Write;
+            writeln!(file_list, "{}", path.display()).expect("failed to write to file list");
+        }
+    }
+    drop(file_list);
+
     let extractor = desugaring::Extractor {
         prefix: "unified".to_string(),
         languages,

unified/ql/lib/codeql/files/FileSystem.qll

@@ -37,7 +37,7 @@ module Folder = Impl::Folder;
 /** A file. */
 class File extends Container, Impl::File {
   /** Holds if this file was extracted from ordinary source code. */
-  predicate fromSource() { any() }
+  predicate fromSource() { exists(this.getRelativePath()) }
 
   /**
    * Gets the number of lines containing code in this file. This value

unified/ql/lib/codeql/unified/internal/ControlFlowGraph.qll

@@ -41,7 +41,9 @@ private module Ast implements AstSig<Location> {
 
   Callable getEnclosingCallable(AstNode node) { result = node.getEnclosingCallable() }
 
-  class Callable = U::Callable;
+  class Callable extends U::Callable {
+    Callable() { this.fromSource() }
+  }
 
   AstNode callableGetBody(Callable c) { result = c.getBody() }
 

unified/ql/lib/codeql/unified/internal/FacadeAst.qll

@@ -15,6 +15,9 @@ module Unified {
     /** Gets the file containing this AST node. */
     File getFile() { result = this.getLocation().getFile() }
 
+    /** Holds if this AST node comes from ordinary source code. */
+    predicate fromSource() { this.getFile().fromSource() }
+
     /** Holds if this AST node has a modifier with the given text. */
     predicate hasModifier(string text) {
       exists(Modifier mod |

unified/ql/test/library-tests/BasicTest/test.ql

@@ -1,6 +1,8 @@
 import unified
 
-query predicate identifier(Identifier node, string value) { value = node.getValue() }
+query predicate identifier(Identifier node, string value) {
+  node.fromSource() and value = node.getValue()
+}
 
 query predicate namedPattern(NamedPattern node, string value) { value = node.getName() }
 

unified/ql/test/library-tests/comments/comments.ql

@@ -1,3 +1,3 @@
 import unified
 
-query predicate comments(Comment c, string text) { text = c.getCommentText() }
+query predicate comments(Comment c, string text) { c.fromSource() and text = c.getCommentText() }

unified/ql/test/library-tests/static-name-binding/explicit-instance-field-access.swift

@@ -15,7 +15,7 @@ private class A {
         return self.y // $ not handled by static name binding
     }
 
-    class func z() -> Int { // name=A.type.z
+    class func z() -> Int { // $ access=Int // name=A.type.z
         return 789
     }
 
@@ -37,7 +37,7 @@ private class B : A { // $ access=A
         return self.y // $ not handled by static name binding
     }
 
-    class func z() -> Int { // name=B.type.z
+    class func z() -> Int { // $ access=Int // name=B.type.z
         return 789
     }
 

unified/tools/BUILD.bazel

@@ -6,6 +6,5 @@ codeql_pkg_files(
         "BUILD.bazel",
     ],
     exes = glob(["**/*"]),
-    prefix = "tools",
     visibility = ["//unified:__pkg__"],
 )

unified/tools/builtins/BUILD.bazel

@@ -0,0 +1,8 @@
+load("//misc/bazel:pkg.bzl", "codeql_pkg_files")
+
+codeql_pkg_files(
+    name = "builtins",
+    srcs = glob(["*.swift"]),
+    prefix = "builtins",
+    visibility = ["//unified:__subpackages__"],
+)