fix: preserve files during fuzzy edits (#51)

[agent-think]

Requested by @aron-cf

Closes #51

What was wrong

applyEditsToNormalizedContent in packages/computer/src/tools/fs/edit-diff.ts used the fully fuzzy-normalized file as baseContent whenever any exact match missed. It then both wrote and diffed that normalized copy. As reproduced in #51, one trailing space in oldText therefore stripped unrelated trailing spaces and folded smart punctuation across the file while the returned diff hid those changes.

The first source-mapping implementation also grapheme-segmented the full file before every fuzzy edit. A file near the 2 MiB edit limit could therefore create millions of short-lived strings inside a Durable Object isolate even when the match touched one small line.

What changed

Fuzzy matching now maps normalized match boundaries back to offsets in the original text and always splices replacements into that original content. Exact and fuzzy edits can share a batch, and uniqueness and overlap checks still run in the matching space. The diff and patch compare the actual original and written content.

Unicode source maps are now built lazily for candidate lines. Lines whose NFKC form is unchanged map directly without grapheme segmentation; lines that need a source map validate only the prefix needed for each boundary and cache the result. Ambiguous boundaries still fail closed.

The package README and tool-interface guide document the fuzzy lookup and byte-preservation contract, and the existing patch changeset covers the user-facing fix.

Testing

• Added 17 pure mapping regressions covering byte preservation, mixed exact/fuzzy batches, NFKC expansions, combining sequences, ambiguous boundaries, trailing whitespace, multiline spans, supplementary characters, uniqueness, and a 256 KiB unrelated line that must not be grapheme-mapped.
• Added a production Workspace/createAITools read-back regression that also verifies the returned diff and patch.
• npm run check — passed.
• npm run typecheck --workspace @cloudflare/computer — passed.
• npm test --workspace @cloudflare/computer — all five suites passed: 1,025 unit, 6 proxy, 5 Worker backend, 23 script runner, and 4 stub-soak tests.
• npm run build --workspace @cloudflare/computer — passed.

Demo

Demo URL (expires after 60 mins): https://computer-fuzzy-edit-feedback-96.halved-naranja.workers.dev

Open it and press Run fuzzy edits. It verifies byte preservation and a fuzzy edit after a 256 KiB unrelated line, and it runs the packed @cloudflare/computer build from this branch. The /demo endpoint was also verified directly with passed: true.

🤖 generated by the pr-agent — please review carefully

[changeset-bot]

🦋 Changeset detected

Latest commit: 9791438

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 1 package

Name	Type
@cloudflare/computer	Patch

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

[github-actions]

Thanks for your interest in Cloudflare Computer.

This repository does not accept unsolicited pull requests. Please use one of the accepted contribution paths instead:

• Bug reports, regressions, and fix proposals: https://github.com/cloudflare/computer/issues/new/choose
• Feature requests, enhancements, and design proposals: https://github.com/cloudflare/computer/discussions

If a maintainer asked you to open this pull request, they can add the allow-pr label and reopen it.

[github-actions]

Thanks for your interest in Cloudflare Computer.

This repository does not accept unsolicited pull requests. Please use one of the accepted contribution paths instead:

• Bug reports, regressions, and fix proposals: https://github.com/cloudflare/computer/issues/new/choose
• Feature requests, enhancements, and design proposals: https://github.com/cloudflare/computer/discussions

If a maintainer asked you to open this pull request, they can add the allow-pr label and reopen it.

[pkg-pr-new]

Open in StackBlitz

npm i https://pkg.pr.new/@cloudflare/computer@96

commit: 9791438

[aron-cf]

@agent-think can you resolve the outstanding PR feedback.

[lucky-verma]

Confirmed on 9c21426 with Node 22 using lines containing the fi ligature and oldText: "f": 20,000 lines took 566 ms and 60,000 took 3.72 s. The new preservation regressions pass locally, but this retry path still scales poorly. I would hold the merge until the boundary lookup is indexed or bounded.

[aron-cf]

@agent-think can you address the following feedback:

In packages/computer/src/tools/fs/edit-diff.ts, findText loops over every normalized occurrence of the search text looking for one whose boundaries map back to the original content. Each iteration maps two boundaries, and mapFuzzyBoundary calls getTextBoundary, which linearly scans the normalized line array from the first line. For content where no occurrence maps cleanly (for example many lines containing an NFKC-expanding character such as the ligature, with a one-character oldText), the loop visits every occurrence and each visit costs O(lines), giving quadratic behavior: roughly 4.3 s for a 240 KB file and far worse near the 2 MiB edit cap, all synchronous inside the Durable Object isolate. Note the duplicate/uniqueness check runs only after findText returns, so it does not bound this work. Consider making boundary lookup cheap (occurrences are returned in increasing offset order, so the line index can be advanced incrementally, or line starts can be binary-searched), and/or capping the number of unmappable candidates tried before failing closed.