fix(web-defaults): safely parse JSON with strings containing slashes …

[shobhitagnihotri69]

Summary

Fixes #219.

When firecrawl setup defaults reads existing configuration files (such as Claude desktop config or VS Code settings), regex-based comment stripping (\/\/.*$) was aggressively removing characters inside string literals (e.g. Read(//c/Users/...)). This produced malformed JSON syntax errors (SyntaxError: Expected double-quoted property name in JSON...) and silently skipped saving web default configurations.

Changes

• String-aware comment stripping: Enhanced removeJsonComments in src/utils/web-defaults.ts with a character-by-character scanner that respects string literal boundaries and escape sequences (\", \\), ensuring // within strings is never removed.
• Direct JSON fast-path: Attempts JSON.parse(existing) first before applying comment stripping.
• Unit test coverage: Added tests in src/__tests__/utils/web-defaults.test.ts verifying that:
  • Strings containing // paths are preserved and parse cleanly.
  • JSON containing both valid string slashes and legitimate // or /* */ comments are parsed without corruption.

Testing

• pnpm test passed 506/506 tests across 28 test suites (100% green).
• pnpm run build succeeded with zero TypeScript errors.

---

Summary by cubic

Fixes firecrawl setup defaults failing to save web defaults when existing config files contain // inside string literals (e.g. Read(//c/Users/...)). The old regex-based comment stripper removed slashes inside strings, causing JSON syntax errors and silently skipping the save. The new string-aware scanner respects string boundaries and escape sequences, and direct JSON parse is attempted first.

Bug Fixes

• Replaces regex comment stripping with a character-by-character scanner that tracks string state and escape sequences.
• Tries JSON.parse on the raw content before applying comment stripping.
• Adds tests covering strings with // paths and files mixing comments with such strings.

^{Written for commit 3de37f2. Summary will update on new commits.}

[cubic-dev-ai]

1 issue found across 2 files

Confidence score: 3/5

• In src/utils/web-defaults.ts, lone \r line endings prevent // comments from terminating, so the scanner can discard the rest of a settings file and skip Claude defaults setup; treat \r as a line-comment terminator and add coverage for legacy line endings.

Prompt for AI agents (unresolved issues)

Check if these issues are valid — if so, understand the root cause of each and fix them. If appropriate, use sub-agents to investigate and fix each issue separately.


<file name="src/utils/web-defaults.ts">

<violation number="1" location="src/utils/web-defaults.ts:52">
P2: When a settings file uses lone `\r` line endings, this scanner never terminates `//` comments and drops the rest of the JSON, causing Claude defaults setup to skip the file. Treat `\r` as a line-comment terminator as well as `\n`.</violation>
</file>

<sub>Heads up: you’re close to your flex budget. Increase your flex budget so reviews don’t pause.

Shadow auto-approve: would not auto-approve because issues were found.

Fix all with cubic | Re-trigger cubic</sub>

[cubic-dev-ai]

P2: When a settings file uses lone \r line endings, this scanner never terminates // comments and drops the rest of the JSON, causing Claude defaults setup to skip the file. Treat \r as a line-comment terminator as well as \n.

Prompt for AI agents

Check if this issue is valid — if so, understand the root cause and fix it. At src/utils/web-defaults.ts, line 52:

<comment>When a settings file uses lone `\r` line endings, this scanner never terminates `//` comments and drops the rest of the JSON, causing Claude defaults setup to skip the file. Treat `\r` as a line-comment terminator as well as `\n`.</comment>

<file context>
@@ -37,10 +37,60 @@ async function writeText(filePath: string, content: string): Promise<void> {
+    const nextChar = i + 1 < content.length ? content[i + 1] : '';
+
+    if (inLineComment) {
+      if (char === '\n') {
+        inLineComment = false;
+        result += char;
</file context>

Suggested change

	if (char === '\n') {
	if (char === '\n' || char === '\r') {