fix(builtin): properly parse status file value containing spaces (#2615)

Based on Bazel's documentation, values provided through the workspace_status_command can contain whitespace on the same line the key is defined on: > The key names can be anything but they may only use upper case letters and underscores. The > first space after the key name separates it from the value. The value is the rest of the line > (including additional whitespaces). The `parseStatusFile` utility function is updated to rely on a regex for performing this parsing to match the described behavior.
bazelbuild · Apr 19, 2021 · 406dcb5 · 406dcb5
1 parent 2ded6f9
commit 406dcb5
Show file tree

Hide file tree

Showing 5 changed files with 164 additions and 53 deletions.
diff --git a/internal/pkg_npm/packager.js b/internal/pkg_npm/packager.js
@@ -61,21 +61,32 @@ function unquoteArgs(s) {
 }
 
 /**
- * The status files are expected to look like
- * BUILD_SCM_HASH 83c699db39cfd74526cdf9bebb75aa6f122908bb
- * BUILD_SCM_LOCAL_CHANGES true
- * STABLE_BUILD_SCM_VERSION 6.0.0-beta.6+12.sha-83c699d.with-local-changes
- * BUILD_TIMESTAMP 1520021990506
- *
- * @param {string} p the path to the status file
- * @returns a two-dimensional array of key/value pairs
- */
+* The status files are expected to look like
+* BUILD_SCM_HASH 83c699db39cfd74526cdf9bebb75aa6f122908bb
+* BUILD_SCM_LOCAL_CHANGES true
+* STABLE_BUILD_SCM_VERSION 6.0.0-beta.6+12.sha-83c699d.with-local-changes
+* BUILD_TIMESTAMP 1520021990506
+*
+* Parsing regex is created based on Bazel's documentation describing the status file schema:
+*   The key names can be anything but they may only use upper case letters and underscores. The
+*   first space after the key name separates it from the value. The value is the rest of the line
+*   (including additional whitespaces).
+*
+* @param {string} p the path to the status file
+* @returns a two-dimensional array of key/value pairs
+*/
 function parseStatusFile(p) {
   if (!p) return [];
-  return fs.readFileSync(p, {encoding: 'utf-8'})
-      .split('\n')
-      .filter(t => !!t)
-      .map(t => t.split(' '));
+  const results = {};
+  const statusFile = fs.readFileSync(p, {encoding: 'utf-8'});
+  for (const match of `\n${statusFile}`.matchAll(/^([A-Z_]+) (.*)/gm)) {
+    // Lines which go unmatched define an index value of `0` and should be skipped.
+    if (match.index === 0) {
+      continue;
+    }
+    results[match[1]] = match[2];
+  }
+  return results;
 }
 
 function main(args) {
@@ -95,15 +106,16 @@ function main(args) {
   // Replace statuses last so that earlier substitutions can add
   // status-related placeholders
   if (volatileFile || infoFile) {
-    const statusEntries = parseStatusFile(volatileFile)
-    statusEntries.push(...parseStatusFile(infoFile))
+    const statuses = {
+      ...parseStatusFile(volatileFile),
+      ...parseStatusFile(infoFile),
+    };
     // Looks like {'BUILD_SCM_VERSION': 'v1.2.3'}
-    const statuses = new Map(statusEntries)
     for (let idx = 0; idx < substitutions.length; idx++) {
       const match = substitutions[idx][1].match(/\{(.*)\}/);
       if (!match) continue;
       const statusKey = match[1];
-      let statusValue = statuses.get(statusKey);
+      let statusValue = statuses[statusKey];
       if (statusValue) {
         // npm versions must be numeric, so if the VCS tag starts with leading 'v', strip it
         // See https://github.com/bazelbuild/rules_nodejs/pull/1591

diff --git a/internal/pkg_web/assembler.js b/internal/pkg_web/assembler.js
@@ -34,15 +34,33 @@ function unquoteArgs(s) {
   return s.replace(/^'(.*)'$/, '$1');
 }
 
-function getBazelStatusMappings(statusFilePath) {
-  if (!statusFilePath) return {};
-  const stampFileLines = fs.readFileSync(statusFilePath, {encoding: 'utf-8'}).trim().split('\n');
-  const stampMap = {};
-  for (const line of stampFileLines) {
-    const [key, value] = line.split(' ');
-    stampMap[key] = value;
+/**
+* The status files are expected to look like
+* BUILD_SCM_HASH 83c699db39cfd74526cdf9bebb75aa6f122908bb
+* BUILD_SCM_LOCAL_CHANGES true
+* STABLE_BUILD_SCM_VERSION 6.0.0-beta.6+12.sha-83c699d.with-local-changes
+* BUILD_TIMESTAMP 1520021990506
+*
+* Parsing regex is created based on Bazel's documentation describing the status file schema:
+*   The key names can be anything but they may only use upper case letters and underscores. The
+*   first space after the key name separates it from the value. The value is the rest of the line
+*   (including additional whitespaces).
+*
+* @param {string} p the path to the status file
+* @returns a two-dimensional array of key/value pairs
+*/
+function parseStatusFile(p) {
+  if (!p) return [];
+  const results = {};
+  const statusFile = fs.readFileSync(p, {encoding: 'utf-8'});
+  for (const match of `\n${statusFile}`.matchAll(/^([A-Z_]+) (.*)/gm)) {
+    // Lines which go unmatched define an index value of `0` and should be skipped.
+    if (match.index === 0) {
+      continue;
+    }
+    results[match[1]] = match[2];
   }
-  return stampMap;
+  return results;
 }
 
 function normalizeSubstitutions(substitutionsArg, stampMap) {
@@ -74,8 +92,8 @@ function main(params) {
   const rawSubstitutions = params.shift().replace(/^'(.*)'$/, '$1');
 
   const stampMap = {
-    ...getBazelStatusMappings(volatileFilePath),
-    ...getBazelStatusMappings(stableFilePath),
+    ...parseStatusFile(volatileFilePath),
+    ...parseStatusFile(stableFilePath),
   };
 
   const normalizedSubstitutions = normalizeSubstitutions(rawSubstitutions, stampMap)

diff --git a/packages/rollup/install.md b/packages/rollup/install.md
@@ -99,15 +99,43 @@ To use these files, you write JS code in your `rollup.config.js` to read from th
 Each line is a space-separated key/value pair.
 
 ```javascript
+/**
+* The status files are expected to look like
+* BUILD_SCM_HASH 83c699db39cfd74526cdf9bebb75aa6f122908bb
+* BUILD_SCM_LOCAL_CHANGES true
+* STABLE_BUILD_SCM_VERSION 6.0.0-beta.6+12.sha-83c699d.with-local-changes
+* BUILD_TIMESTAMP 1520021990506
+*
+* Parsing regex is created based on Bazel's documentation describing the status file schema:
+*   The key names can be anything but they may only use upper case letters and underscores. The
+*   first space after the key name separates it from the value. The value is the rest of the line
+*   (including additional whitespaces).
+*
+* @param {string} p the path to the status file
+* @returns a two-dimensional array of key/value pairs
+*/
+function parseStatusFile(p) {
+  if (!p) return [];
+  const results = {};
+  const statusFile = require('fs').readFileSync(p, {encoding: 'utf-8'});
+  for (const match of `\n${statusFile}`.matchAll(/^([A-Z_]+) (.*)/gm)) {
+    // Lines which go unmatched define an index value of `0` and should be skipped.
+    if (match.index === 0) {
+      continue;
+    }
+    results[match[1]] = match[2];
+  }
+  return results;
+}
+
+const statuses = parseStatusFile(bazel_version_file);
 // Parse the stamp file produced by Bazel from the version control system
 let version = '<unknown>';
-if (bazel_info_file) {
-  const versionTag = require('fs')
-                         .readFileSync(bazel_info_file, {encoding: 'utf-8'})
-                         .split('\n')
-                         .find(s => s.startsWith('STABLE_GIT_COMMIT'));
-  if (versionTag) {
-    version = 'v' + versionTag.split(' ')[1].trim();
+// Don't assume BUILD_SCM_VERSION exists
+if (statuses['BUILD_SCM_VERSION']) {
+  version = 'v' + statuses['BUILD_SCM_VERSION'];
+  if (DEBUG) {
+    version += '_debug';
   }
 }
 ```

diff --git a/packages/rollup/test/integration/rollup.config.js b/packages/rollup/test/integration/rollup.config.js
@@ -2,16 +2,44 @@ import commonjs from '@rollup/plugin-commonjs';
 import json from '@rollup/plugin-json';
 import nodeResolve from '@rollup/plugin-node-resolve';
 
+/**
+* The status files are expected to look like
+* BUILD_SCM_HASH 83c699db39cfd74526cdf9bebb75aa6f122908bb
+* BUILD_SCM_LOCAL_CHANGES true
+* STABLE_BUILD_SCM_VERSION 6.0.0-beta.6+12.sha-83c699d.with-local-changes
+* BUILD_TIMESTAMP 1520021990506
+*
+* Parsing regex is created based on Bazel's documentation describing the status file schema:
+*   The key names can be anything but they may only use upper case letters and underscores. The
+*   first space after the key name separates it from the value. The value is the rest of the line
+*   (including additional whitespaces).
+*
+* @param {string} p the path to the status file
+* @returns a two-dimensional array of key/value pairs
+*/
+function parseStatusFile(p) {
+  if (!p) return [];
+  const results = {};
+  const statusFile = require('fs').readFileSync(p, {encoding: 'utf-8'});
+  for (const match of `\n${statusFile}`.matchAll(/^([A-Z_]+) (.*)/gm)) {
+    // Lines which go unmatched define an index value of `0` and should be skipped.
+    if (match.index === 0) {
+      continue;
+    }
+    results[match[1]] = match[2];
+  }
+  return results;
+}
+
 // Parse the stamp file produced by Bazel from the version control system
 let version = '<unknown>';
-if (bazel_version_file) {
-  const versionTag = require('fs')
-                         .readFileSync(bazel_version_file, {encoding: 'utf-8'})
-                         .split('\n')
-                         .find(s => s.startsWith('BUILD_SCM_VERSION'));
-  // Don't assume BUILD_SCM_VERSION exists
-  if (versionTag) {
-    version = 'v' + versionTag.split(' ')[1].trim();
+
+const statuses = parseStatusFile(bazel_version_file);
+// Don't assume BUILD_SCM_VERSION exists
+if (statuses['BUILD_SCM_VERSION']) {
+  version = 'v' + statuses['BUILD_SCM_VERSION'];
+  if (DEBUG) {
+    version += '_debug';
   }
 }
 

diff --git a/packages/rollup/test/version_stamp/rollup.config.js b/packages/rollup/test/version_stamp/rollup.config.js
@@ -1,18 +1,43 @@
+/**
+* The status files are expected to look like
+* BUILD_SCM_HASH 83c699db39cfd74526cdf9bebb75aa6f122908bb
+* BUILD_SCM_LOCAL_CHANGES true
+* STABLE_BUILD_SCM_VERSION 6.0.0-beta.6+12.sha-83c699d.with-local-changes
+* BUILD_TIMESTAMP 1520021990506
+*
+* Parsing regex is created based on Bazel's documentation describing the status file schema:
+*   The key names can be anything but they may only use upper case letters and underscores. The
+*   first space after the key name separates it from the value. The value is the rest of the line
+*   (including additional whitespaces).
+*
+* @param {string} p the path to the status file
+* @returns a two-dimensional array of key/value pairs
+*/
+function parseStatusFile(p) {
+  if (!p) return [];
+  const results = {};
+  const statusFile = require('fs').readFileSync(p, {encoding: 'utf-8'});
+  for (const match of `\n${statusFile}`.matchAll(/^([A-Z_]+) (.*)/gm)) {
+    // Lines which go unmatched define an index value of `0` and should be skipped.
+    if (match.index === 0) {
+      continue;
+    }
+    results[match[1]] = match[2];
+  }
+  return results;
+}
+
 const DEBUG = process.env['COMPILATION_MODE'] === 'dbg';
 
 // Parse the stamp file produced by Bazel from the version control system
 let version = '<unknown>';
-if (bazel_version_file) {
-  const versionTag = require('fs')
-                         .readFileSync(bazel_version_file, {encoding: 'utf-8'})
-                         .split('\n')
-                         .find(s => s.startsWith('BUILD_SCM_VERSION'));
-  // Don't assume BUILD_SCM_VERSION exists
-  if (versionTag) {
-    version = 'v' + versionTag.split(' ')[1].trim();
-    if (DEBUG) {
-      version += '_debug';
-    }
+
+const statuses = parseStatusFile(bazel_version_file);
+// Don't assume BUILD_SCM_VERSION exists
+if (statuses['BUILD_SCM_VERSION']) {
+  version = 'v' + statuses['BUILD_SCM_VERSION'];
+  if (DEBUG) {
+    version += '_debug';
   }
 }