Bugfixes and upgraded to v0.0.5 (#7)

* Bugfix: Support octal-escaped filenames coming from git diff.
* Bugfix: don't try to open directories, better handling of `--- ...`
non-filename diffs.
* Bugfix: Warn instead of failing if the comment extractor sees an
unknown file.

Author: ebrevdo

package.json

@@ -1,6 +1,6 @@
 {
   "name": "ifttt-lint",
-  "version": "0.0.4",
+  "version": "0.0.5",
   "description": "A high-concurrency IFTTT (IfThisThenThat) linter",
   "homepage": "https://www.github.com/ebrevdo/ifttt-lint",
   "main": "dist/main.js",

src/DiffParser.ts

@@ -39,10 +39,18 @@ import parseDiff from 'parse-diff';
  * @returns A Map where each key is a file path and its value contains added and removed line numbers.
  */
 export function parseChangedLines(diffText: string): Map<string, FileChanges> {
-  // Remove 'diff ' headers to avoid parse-diff header misparsing
+  // Sanitize diff input: drop git diff headers and spurious '--- ' or '+++ ' lines not indicating actual file paths
   const filtered = diffText
     .split(/\r?\n/)
-    .filter(line => !line.startsWith('diff '))
+    .filter(line => {
+      // drop main diff header lines
+      if (line.startsWith('diff ')) return false;
+      // drop spurious file-header-like lines: raw '--- ' not followed by a valid file path (e.g., prefix/ or /dev/null)
+      if (/^--- /.test(line) && !/^--- [^ ]+\//.test(line)) return false;
+      // drop spurious new-file-header-like lines: raw '+++ ' not followed by a valid file path (e.g., prefix/ or /dev/null)
+      if (/^\+\+\+ /.test(line) && !/^\+\+\+ [^ ]+\//.test(line)) return false;
+      return true;
+    })
     .join('\n');
   const files = parseDiff(filtered);
   const result = new Map<string, FileChanges>();
@@ -52,7 +60,42 @@ export function parseChangedLines(diffText: string): Map<string, FileChanges> {
       continue;
     }
     // Determine file path: prefer 'to' unless it's '/dev/null', else use 'from'
-    const raw = file.to && file.to !== '/dev/null' ? file.to : file.from;
+    let raw = file.to && file.to !== '/dev/null' ? file.to : file.from;
+    raw = raw.trim();
+    // Strip surrounding quotes if present
+    if ((raw.startsWith('"') && raw.endsWith('"')) || (raw.startsWith("'") && raw.endsWith("'"))) {
+      raw = raw.slice(1, -1);
+    }
+    // Decode C-style octal escapes (e.g., \360) to actual UTF-8 characters
+    raw = ((): string => {
+      const bytes: number[] = [];
+      let i = 0;
+      while (i < raw.length) {
+        if (raw[i] === '\\') {
+          let j = i + 1;
+          let oct = '';
+          // collect up to 3 octal digits
+          while (j < raw.length && oct.length < 3 && /[0-7]/.test(raw[j])) {
+            oct += raw[j];
+            j++;
+          }
+          if (oct.length > 0) {
+            bytes.push(parseInt(oct, 8));
+            i = j;
+            continue;
+          }
+          // not an octal escape, treat literal '\'
+          bytes.push(raw.charCodeAt(i));
+          i++;
+        } else {
+          // normal character: encode as UTF-8 bytes
+          const buf = Buffer.from(raw[i], 'utf-8');
+          for (const b of buf) bytes.push(b);
+          i++;
+        }
+      }
+      return Buffer.from(bytes).toString('utf-8');
+    })();
     // Strip single-character prefix (e.g., 'a/', 'b/')
     const filePath = raw.length > 1 && raw[1] === '/' ? raw.slice(2) : raw;
     const added = new Set<number>();

src/DirectiveParser.ts

@@ -27,14 +27,29 @@ const endLabelRegex = /LINT\.EndLabel/;
 export async function parseFileDirectives(
   filePath: string
 ): Promise<LintDirective[]> {
-  const content = await fs.readFile(filePath, 'utf-8');
+  let content: string;
+  try {
+    content = await fs.readFile(filePath, 'utf-8');
+  } catch (err: unknown) {
+    // If path is a directory, skip without error
+    if (
+      typeof err === 'object' &&
+      err !== null &&
+      'code' in err &&
+      (err as { code?: string }).code === 'EISDIR'
+    ) {
+      return [];
+    }
+    // Propagate other errors (e.g., permission issues)
+    throw err;
+  }
   // Use multi-language comment extractor to find all comments in source
   // Pass filename so extractor can choose comment syntax by extension
   let commentsMap: Record<string, { content?: string }>;
+  // Try extracting comments by file extension; if unsupported, fall back or warn
   try {
     commentsMap = extractComments(content, { filename: filePath });
   } catch {
-    // Fallback for unsupported extensions (e.g., .bzl): remap to Python or JS
     const ext = path.extname(filePath).toLowerCase();
     let fallback: string;
     if (ext === '.bzl') {
@@ -44,7 +59,12 @@ export async function parseFileDirectives(
       // Default to JavaScript
       fallback = filePath.replace(path.extname(filePath), '.js');
     }
-    commentsMap = extractComments(content, { filename: fallback });
+    try {
+      commentsMap = extractComments(content, { filename: fallback });
+    } catch {
+      // Could not extract comments (e.g., JSON or unsupported extensions): ignore silently
+      return [];
+    }
   }
   // commentsMap maps starting line numbers (as strings) to comment objects
   const directives: LintDirective[] = [];

tests/DiffParser.test.ts

@@ -0,0 +1,99 @@
+import { parseChangedLines } from '../src/DiffParser';
+
+describe('parseChangedLines spurious header filtering', () => {
+  it('ignores comment lines starting with --- without valid path', () => {
+    const diff = [
+      'diff --git a/file.sql b/file.sql',
+      'index abc..def 100644',
+      '--- a/file.sql',
+      '+++ b/file.sql',
+      '@@ -1,6 +1,6 @@',
+      ' SELECT *',
+      '+SELECT hello',
+      ' --- This is a comment, not a file header',
+      '-OLD LINE',
+      '+NEW LINE'
+    ].join('\n');
+    const result = parseChangedLines(diff);
+    // Only one file should be parsed
+    expect(result.size).toBe(1);
+    expect(result.has('file.sql')).toBe(true);
+    const changes = result.get('file.sql')!;
+    // Added lines at newLine positions 2 and 4
+    expect(changes.addedLines.has(2)).toBe(true);
+    expect(changes.addedLines.has(4)).toBe(true);
+    // Removed line at oldLine position 3
+    expect(changes.removedLines.has(3)).toBe(true);
+  });
+
+  it('retains real file header lines starting with --- a/ or +++ b/', () => {
+    const diff = [
+      'diff --git a/foo.js b/foo.js',
+      'index 123..456 100644',
+      '--- a/foo.js',
+      '+++ b/foo.js',
+      '@@ -0,0 +1,1 @@',
+      '+console.log("hi");'
+    ].join('\n');
+    const result = parseChangedLines(diff);
+    expect(result.size).toBe(1);
+    expect(Array.from(result.keys())).toEqual(['foo.js']);
+    const changes = result.get('foo.js')!;
+    expect(changes.addedLines.has(1)).toBe(true);
+  });
+  it('accepts file headers with arbitrary prefixes (e.g., c/ and w/)', () => {
+    const diff = [
+      'diff --git c/path/file.txt w/path/file.txt',
+      'index 123..456 100644',
+      '--- c/path/file.txt',
+      '+++ w/path/file.txt',
+      '@@ -1 +1 @@',
+      '-old line',
+      '+new line'
+    ].join('\n');
+    const result = parseChangedLines(diff);
+    expect(result.size).toBe(1);
+    expect(result.has('path/file.txt')).toBe(true);
+    const changes = result.get('path/file.txt')!;
+    expect(changes.removedLines.has(1)).toBe(true);
+    expect(changes.addedLines.has(1)).toBe(true);
+  });
+
+  it('strips surrounding quotes and prefixes for quoted file paths', () => {
+    const diff = [
+      'diff --git "a/my file.txt" "b/my file.txt"',
+      'index 111..222 100644',
+      '--- "a/my file.txt"',
+      '+++ "b/my file.txt"',
+      '@@ -1 +1 @@',
+      '-old',
+      '+new'
+    ].join('\n');
+    const result = parseChangedLines(diff);
+    expect(result.size).toBe(1);
+    expect(result.has('my file.txt')).toBe(true);
+    const changes = result.get('my file.txt')!;
+    expect(changes.removedLines.has(1)).toBe(true);
+    expect(changes.addedLines.has(1)).toBe(true);
+  });
+
+  it('decodes octal-escaped paths with non-ASCII chars', () => {
+    const esc = '3_\\360\\237\\224\\216_test.py';
+    const diff = [
+      `diff --git a/${esc} b/${esc}`,
+      'index 1..2 100644',
+      `--- a/${esc}`,
+      `+++ b/${esc}`,
+      '@@ -1 +1 @@',
+      '-old',
+      '+new'
+    ].join('\n');
+    const result = parseChangedLines(diff);
+    const expectedName = '3_🔎_test.py';
+    expect(result.size).toBe(1);
+    expect(result.has(expectedName)).toBe(true);
+    const changes = result.get(expectedName)!;
+    expect(changes.removedLines.has(1)).toBe(true);
+    expect(changes.addedLines.has(1)).toBe(true);
+  });
+});

tests/DirectiveParserList.test.ts

@@ -48,4 +48,11 @@ describe('DirectiveParser ThenChange array literal', () => {
       { kind: 'ThenChange', line: 1, target: 'file2.ts#lbl' }
     ]);
   });
+  describe('parseFileDirectives on directories', () => {
+    it('returns empty directive list for directory paths', async () => {
+      const tmpDir = await fs.mkdtemp(path.join(os.tmpdir(), 'lint-dir-'));
+      const directives = await parseFileDirectives(tmpDir);
+      expect(directives).toEqual([]);
+    });
+  });
 });