Merge pull request #3181 from tree-sitter/handle-wasm-oom

When loading languages via WASM, gracefully handle memory errors and leaks in external scanners

script/build-wasm-stdlib

@@ -2,27 +2,33 @@
 
 set -e
 
-# Remove quotes, add leading underscores, remove newlines, remove trailing comma.
+# Remove quotes and commas
 EXPORTED_FUNCTIONS=$(                     \
     cat lib/src/wasm/stdlib-symbols.txt | \
-    sed -e 's/"//g'                     | \
-    sed -e 's/^/_/g'                    | \
-    tr -d '\n"'                         | \
-    sed -e 's/,$//'                       \
+    tr -d ',"'                            \
 )
 
-emcc                                                          \
-    -o stdlib.wasm                                            \
-    -Os                                                       \
-    --no-entry                                                \
-    -s MAIN_MODULE=2                                          \
-    -s "EXPORTED_FUNCTIONS=${EXPORTED_FUNCTIONS}" \
-    -s 'ALLOW_MEMORY_GROWTH'                                  \
-    -s 'TOTAL_MEMORY=4MB'                                     \
-    -fvisibility=hidden                                       \
-    -fno-exceptions                                           \
-    -xc                                                       \
-    /dev/null
+EXPORT_FLAGS=""
+for function in ${EXPORTED_FUNCTIONS}; do
+    EXPORT_FLAGS+=" -Wl,--export=${function}"
+done
+
+target/wasi-sdk-21.0/bin/clang-17   \
+    -o stdlib.wasm                  \
+    -Os                             \
+    -fPIC                           \
+    -Wl,--no-entry                  \
+    -Wl,--stack-first               \
+    -Wl,-z -Wl,stack-size=65536     \
+    -Wl,--import-undefined          \
+    -Wl,--import-memory             \
+    -Wl,--import-table              \
+    -Wl,--strip-debug               \
+    -Wl,--export=reset_heap         \
+    -Wl,--export=__wasm_call_ctors  \
+    -Wl,--export=__stack_pointer    \
+    ${EXPORT_FLAGS}                 \
+    lib/src/wasm/stdlib.c
 
 xxd -C -i stdlib.wasm > lib/src/wasm/wasm-stdlib.h
 mv stdlib.wasm target/

script/test

@@ -12,7 +12,7 @@ OPTIONS
 
   -h  Print this message
 
-  -a  Compile C code with the Clang static analyzer
+  -a  Compile C code with the Clang address sanitizer
 
   -e  Run only the corpus tests whose name contain the given string
 
@@ -41,9 +41,17 @@ while getopts "adDghl:e:s:i:" option; do
       exit
       ;;
     a)
-      export RUSTFLAGS="-Z sanitizer=address"
-      # Specify a `--target` explicitly. For some reason, this is required for
-      # address sanitizer support.
+      export CFLAGS="-fsanitize=undefined,address"
+
+      # When the Tree-sitter C library is compiled with the address sanitizer, the address sanitizer
+      # runtime library needs to be linked into the final test executable. When using Xcode clang,
+      # the Rust linker doesn't know where to find that library, so we need to specify linker flags directly.
+      runtime_dir=$(cc -print-runtime-dir)
+      if [[ $runtime_dir == */Xcode.app/* ]]; then
+        export RUSTFLAGS="-C link-arg=-L${runtime_dir} -C link-arg=-lclang_rt.asan_osx_dynamic -C link-arg=-Wl,-rpath,${runtime_dir}"
+      fi
+
+      # Specify a `--target` explicitly. This is required for address sanitizer support.
       toolchain=$(rustup show active-toolchain)
       toolchain_regex='(stable|beta|nightly)-([_a-z0-9-]+).*'
       if [[ $toolchain =~ $toolchain_regex ]]; then
@@ -52,7 +60,8 @@ while getopts "adDghl:e:s:i:" option; do
       else
         echo "Failed to parse toolchain '${toolchain}'"
       fi
-      test_flags="${test_flags} --target ${current_target}"
+
+      test_flags+=" --target ${current_target}"
       ;;
     e)
       export TREE_SITTER_EXAMPLE=${OPTARG}