feat: add support for capturing and overriding the exit code within run_node

internal/node/launcher.sh

@@ -170,6 +170,7 @@ USER_NODE_OPTIONS=()
 ALL_ARGS=(TEMPLATED_args $NODE_REPOSITORY_ARGS "$@")
 STDOUT_CAPTURE=""
 STDERR_CAPTURE=""
+EXIT_CODE_CAPTURE=""
 
 RUN_LINKER=true
 NODE_PATCHES=true
@@ -179,8 +180,12 @@ for ARG in "${ALL_ARGS[@]:-}"; do
   case "$ARG" in
     # Supply custom linker arguments for first-party dependencies
     --bazel_node_modules_manifest=*) MODULES_MANIFEST="${ARG#--bazel_node_modules_manifest=}" ;;
+    # Captures stdout of the node process to the file specified
     --bazel_capture_stdout=*) STDOUT_CAPTURE="${ARG#--bazel_capture_stdout=}" ;;
+    # Captures stderr of the node process to the file specified
     --bazel_capture_stderr=*) STDERR_CAPTURE="${ARG#--bazel_capture_stderr=}" ;;
+    # Captures the exit code of the node process to the file specified
+    --bazel_capture_exit_code=*) EXIT_CODE_CAPTURE="${ARG#--bazel_capture_exit_code=}" ;;
     # Disable the node_loader.js monkey patches for require()
     # Note that this means you need an explicit runfiles helper library
     --nobazel_patch_module_resolver) PATCH_REQUIRE=false ;;
@@ -327,6 +332,10 @@ wait "${child}"
 RESULT="$?"
 set -e
 
+if [[ -n "${EXIT_CODE_CAPTURE}" ]]; then
+  echo "${RESULT}" > "${EXIT_CODE_CAPTURE}"
+fi
+
 if [ "${EXPECTED_EXIT_CODE}" != "0" ]; then
   if [ ${RESULT} != ${EXPECTED_EXIT_CODE} ]; then
     echo "Expected exit code to be ${EXPECTED_EXIT_CODE}, but got ${RESULT}" >&2
@@ -360,4 +369,8 @@ if [[ -n "${COVERAGE_DIR:-}" ]]; then
   fi
 fi
 
-exit ${RESULT}
+if [[ -n "${EXIT_CODE_CAPTURE}" ]]; then
+  exit 0
+else
+  exit ${RESULT}
+fi

internal/node/npm_package_bin.bzl

@@ -10,6 +10,7 @@ _ATTRS = {
     "args": attr.string_list(mandatory = True),
     "configuration_env_vars": attr.string_list(default = []),
     "data": attr.label_list(allow_files = True, aspects = [module_mappings_aspect, node_modules_aspect]),
+    "exit_code_out": attr.output(),
     "output_dir": attr.bool(),
     "outs": attr.output_list(),
     "stderr": attr.output(),
@@ -64,6 +65,9 @@ def _impl(ctx):
     if ctx.outputs.stderr:
         tool_outputs.append(ctx.outputs.stderr)
 
+    if ctx.outputs.exit_code_out:
+        tool_outputs.append(ctx.outputs.exit_code_out)
+
     run_node(
         ctx,
         executable = "tool",
@@ -73,6 +77,7 @@ def _impl(ctx):
         configuration_env_vars = ctx.attr.configuration_env_vars,
         stdout = ctx.outputs.stdout,
         stderr = ctx.outputs.stderr,
+        exit_code_out = ctx.outputs.exit_code_out,
     )
 
     return [DefaultInfo(files = depset(outputs + tool_outputs))]
@@ -104,6 +109,9 @@ def npm_package_bin(tool = None, package = None, package_bin = None, data = [],
                 subject to the same semantics as `outs`
         stdout: set to capture the stdout of the binary to a file, which can later be used as an input to another target
                 subject to the same semantics as `outs`
+        exit_code_out: set to capture the exit code of the binary to a file, which can later be used as an input to another target
+                subject to the same semantics as `outs`. Note that setting this will force the binary to exit 0.
+                If the binary creates outputs and these are declared, they must still be created
 
         args: Command-line arguments to the tool.
 

internal/node/test/BUILD.bazel

@@ -260,6 +260,7 @@ generated_file_test(
 
 # capture stderr to diagnostics.out, then analyze it in terser_diagnostics_stats printing some stats
 # stdout is captured into greeter.min.js (again, without the --output $@ flags)
+# the exit code is also capture to a file with exit_code_out, allowing downstream actions to read it
 npm_package_bin(
     name = "diagnostics_producing_bin",
     args = [
@@ -269,6 +270,7 @@ npm_package_bin(
         "--warn",
     ],
     data = ["terser_input_with_diagnostics.js"],
+    exit_code_out = "exit.code",
     package = "terser",
     stderr = "diagnostics.out",
     stdout = "greeter.min.js",
@@ -299,6 +301,12 @@ generated_file_test(
     generated = ":diagnostics.scrubbed.out",
 )
 
+generated_file_test(
+    name = "exit_code_test",
+    src = "exit_code_test.golden",
+    generated = ":exit.code",
+)
+
 nodejs_binary(
     name = "copy_to_directory",
     entry_point = "copy_to_directory.js",

internal/node/test/exit_code_test.golden

@@ -0,0 +1 @@
+0

internal/providers/node_runtime_deps_info.bzl

@@ -87,6 +87,12 @@ def run_node(ctx, inputs, arguments, executable, **kwargs):
         add_arg(arguments, "--bazel_capture_stderr=%s" % stderr_file.path)
         outputs = outputs + [stderr_file]
 
+    exit_code_file = kwargs.pop("exit_code_out", None)
+    if exit_code_file:
+        # this will force the script to exit 0, all declared outputs must still be created
+        add_arg(arguments, "--bazel_capture_exit_code=%s" % exit_code_file.path)
+        outputs = outputs + [exit_code_file]
+
     # By using the run_node helper, you suggest that your program
     # doesn't implicitly use runfiles to require() things
     # To access runfiles, you must use a runfiles helper in the program instead