ci: Check binaries separately from building them (#7355)

Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>

docker/check-sanitizers.sh

@@ -1,48 +0,0 @@
-#!/bin/bash
-
-# Sanity-check that the sanitizer runtimes shipped with g++/clang++ work
-# end-to-end against the system loader: compile each example with both
-# compilers, run it, and confirm the expected diagnostic is emitted.
-
-set -eo pipefail
-
-cpp_files_dir="${1:?usage: $0 <cpp_files_dir>}"
-
-case "$(uname -m)" in
-    x86_64) loader=/lib64/ld-linux-x86-64.so.2 ;;
-    aarch64) loader=/lib/ld-linux-aarch64.so.1 ;;
-    *)
-        echo "Unsupported arch: $(uname -m)" >&2
-        exit 1
-        ;;
-esac
-
-declare -A sanitize=(
-    [asan]="-fsanitize=address"
-    [tsan]="-fsanitize=thread"
-    [ubsan]="-fsanitize=undefined"
-)
-declare -A expect=(
-    [asan]="heap-use-after-free"
-    [tsan]="data race"
-    [ubsan]="signed integer overflow"
-)
-
-for compiler in g++ clang++; do
-    for name in asan tsan ubsan; do
-        bin="/tmp/${name}-${compiler}"
-        echo "=== Build ${name} with ${compiler} ==="
-        "$compiler" -std=c++20 -O1 -g ${sanitize[$name]} \
-            -Wl,--dynamic-linker=$loader \
-            "${cpp_files_dir}/${name}.cpp" -o "$bin"
-        echo "=== Run ${name}-${compiler} ==="
-        output=$("$bin" 2>&1) || true
-        echo "$output"
-        echo "$output" | grep -q "${expect[$name]}" ||
-            {
-                echo "expected '${expect[$name]}' from $bin"
-                exit 1
-            }
-        rm -f "$bin"
-    done
-done

docker/loader-path.sh

@@ -0,0 +1,12 @@
+#!/bin/bash
+
+case "$(uname -m)" in
+    x86_64) LOADER=/lib64/ld-linux-x86-64.so.2 ;;
+    aarch64) LOADER=/lib/ld-linux-aarch64.so.1 ;;
+    *)
+        echo "Unsupported arch: $(uname -m)" >&2
+        exit 1
+        ;;
+esac
+
+echo "${LOADER}"

docker/nix.Dockerfile

@@ -27,7 +27,9 @@ RUN mkdir /tmp/nix-store-closure && \
     cp -R $(nix-store -qR result/) /tmp/nix-store-closure
 
 # Final image
-FROM ${BASE_IMAGE}
+FROM ${BASE_IMAGE} AS final
+
+ARG BASE_IMAGE
 
 # bash is not located at /bin/bash in nixos/nix, so we need to create a symlink to it.
 RUN if [ -d /nix ]; then \
@@ -43,25 +45,23 @@ ENTRYPOINT ["/bin/bash"]
 COPY --from=builder /tmp/nix-store-closure /nix/store
 COPY --from=builder /tmp/build/result /nix/ci-env
 
-ENV PATH="/nix/ci-env/bin:$PATH"
+ENV PATH="/nix/ci-env/bin:${PATH}"
 
 # Externally-built dynamically-linked ELF binaries hard-code the loader path
-# (e.g. /lib64/ld-linux-x86-64.so.2) in their PT_INTERP header. Copy the
-# loader from the Nix store to that path when the base image doesn't already
-# provide one (i.e. on nixos/nix).
+# (e.g. /lib64/ld-linux-x86-64.so.2) in their PT_INTERP header. Install it
+# from the Nix store when the base image doesn't already provide one.
+COPY docker/loader-path.sh /tmp/loader-path.sh
+
 RUN <<EOF
-case "$(uname -m)" in
-    x86_64)  target=/lib64/ld-linux-x86-64.so.2 ;;
-    aarch64) target=/lib/ld-linux-aarch64.so.1 ;;
-    *) echo "Unsupported arch: $(uname -m)" >&2; exit 1 ;;
-esac
-if [ ! -e "$target" ]; then
+target="$(/tmp/loader-path.sh)"
+
+if [ ! -e "${target}" ]; then
     # Use the loader from the same glibc that gcc links libc against, so
     # ld-linux and libc/libpthread share GLIBC_PRIVATE symbols at runtime.
-    src="$(dirname "$(gcc -print-file-name=libc.so.6)")/$(basename "$target")"
-    [ -e "$src" ] || { echo "ld-linux not found at $src" >&2; exit 1; }
-    mkdir -p "$(dirname "$target")"
-    cp "$src" "$target"
+    src="$(dirname "$(gcc -print-file-name=libc.so.6)")/$(basename "${target}")"
+    [ -e "${src}" ] || { echo "ld-linux not found at ${src}" >&2; exit 1; }
+    mkdir -p "$(dirname "${target}")"
+    cp "${src}" "${target}"
 fi
 EOF
 
@@ -87,9 +87,16 @@ run-clang-tidy --help
 vim --version
 EOF
 
-# Sanity-check that the sanitizer runtimes shipped with g++/clang++ work
-# end-to-end against the system loader.
-COPY docker/cpp_files/ /tmp/cpp_files/
-COPY docker/check-sanitizers.sh /tmp/check-sanitizers.sh
+# Sanity-check that the sanitizer runtimes shipped with g++/clang++ are able to build binaries
+COPY docker/test_files/cpp_sources/ /tmp/cpp_sources/
+COPY docker/test_files/compile-cpp-sources.sh /tmp/compile-cpp-sources.sh
+RUN /tmp/compile-cpp-sources.sh /tmp/cpp_sources /tmp/bins
 
-RUN grep -qi ubuntu /etc/os-release 2>/dev/null && /tmp/check-sanitizers.sh /tmp/cpp_files || true
+# Sanity-check that the built binaries are able to run.
+# We only support running the test binaries on Ubuntu and NixOS right now (will be fixed in the future)
+#
+# When build and test images will be separate, we will be to run on vanilla images.
+COPY docker/test_files/run-test-binaries.sh /tmp/run-test-binaries.sh
+RUN if echo "${BASE_IMAGE}" | grep -qiE '(ubuntu|nixos)'; then \
+        /tmp/run-test-binaries.sh /tmp/bins; \
+    fi

docker/test_files/compile-cpp-sources.sh

@@ -0,0 +1,50 @@
+#!/bin/bash
+# Compile all C++ test binaries during the Docker image build.
+# Each binary has the target system's ELF PT_INTERP (dynamic-linker path)
+# baked in so it can run on the (potentially minimal) final BASE_IMAGE.
+
+set -eo pipefail
+
+src_dir="${1:?usage: $0 <src_dir> <dst_dir>}"
+dst_dir="${2:?usage: $0 <src_dir> <dst_dir>}"
+
+loader="$(/tmp/loader-path.sh)"
+
+mkdir -p "${dst_dir}"
+
+function compile() {
+    local compiler="${1}"
+    local name="${2}"
+    local san_flag="${3:-}"
+
+    local src="${src_dir}/${name}.cpp"
+    local binary="${dst_dir}/${name}-${compiler}"
+
+    echo "=== Compile ${name} with ${compiler} ==="
+    cmd="${compiler} -std=c++23 -O1 -g \
+        -pthread \
+        -Wl,--dynamic-linker=${loader} \
+        ${san_flag} \
+        ${src} -o ${binary}"
+    echo "Command: ${cmd}"
+    eval "${cmd}"
+}
+
+declare -A sanitize=(
+    [regular]=""
+
+    [asan]="-fsanitize=address"
+    [tsan]="-fsanitize=thread"
+    [ubsan]="-fsanitize=undefined -fno-sanitize-recover=all"
+)
+
+for name in regular asan tsan ubsan; do
+    san_flag="${sanitize[${name}]}"
+    for compiler in g++ clang++; do
+        compile "${compiler}" "${name}" "${san_flag}"
+    done
+done
+
+echo "=== All binaries compiled ==="
+
+ls -la "${dst_dir}"

docker/test_files/cpp_sources/regular.cpp

@@ -0,0 +1,28 @@
+#include <iostream>
+#include <mutex>
+#include <thread>
+#include <vector>
+
+static std::mutex gMutex;
+
+void
+worker(int id)
+{
+    std::lock_guard<std::mutex> lock(gMutex);
+    std::cout << "Hello from thread " << id << "\n";
+}
+
+int
+main()
+{
+    constexpr int kNumThreads = 10;
+    std::vector<std::thread> threads;
+    threads.reserve(kNumThreads);
+    for (int i = 0; i < kNumThreads; ++i)
+        threads.emplace_back(worker, i);
+    for (auto& t : threads)
+        t.join();
+
+    std::cout << "Hello from main thread\n";
+    return 0;
+}

docker/test_files/run-test-binaries.sh

@@ -0,0 +1,62 @@
+#!/bin/bash
+# Run pre-compiled sanitizer binaries and confirm each emits its expected diagnostic.
+# Binaries must already exist in <bins_dir> with the layout:
+# <name>-g++ and <name>-clang++ for name in {regular,asan,tsan,ubsan}
+
+set -eo pipefail
+
+bins_dir="${1:?usage: $0 <bins_dir>}"
+
+# Run a binary and verify its exit code and output.
+# Usage: run <binary> <expected_output> <expected_rc>
+function run() {
+    local binary="${1}"
+    local expected_output="${2}"
+    local expected_rc="${3}"
+
+    local out_file
+    out_file="$(mktemp)"
+
+    echo "=== Run ${binary} ==="
+    local rc=0
+    "${binary}" >"${out_file}" 2>&1 || rc=$?
+
+    cat "${out_file}"
+
+    if [ "${expected_rc}" = "nonzero" ]; then
+        if [ "${rc}" -eq 0 ]; then
+            echo "ERROR: expected non-zero exit code from ${binary}, got ${rc}" >&2
+            exit 1
+        fi
+    elif [ "${rc}" -ne "${expected_rc}" ]; then
+        echo "ERROR: expected exit code ${expected_rc} from ${binary}, got ${rc}" >&2
+        exit 1
+    fi
+
+    grep -q "${expected_output}" "${out_file}" ||
+        {
+            echo "ERROR: expected '${expected_output}' from ${binary}" >&2
+            exit 1
+        }
+    echo "OK: '${expected_output}' detected"
+}
+
+declare -A expect=(
+    [regular]="Hello from main thread"
+
+    [asan]="heap-use-after-free"
+    [tsan]="data race"
+    [ubsan]="signed integer overflow"
+)
+
+for compiler in g++ clang++; do
+    for name in regular asan tsan ubsan; do
+        binary="${bins_dir}/${name}-${compiler}"
+        if [ "${name}" = "regular" ]; then
+            expected_rc=0
+        else
+            expected_rc=nonzero
+        fi
+        run "${binary}" "${expect[$name]}" "${expected_rc}"
+    done
+done