perf(web): skip base64 for oversized image candidates

[tarik02]

What Changed

• Check the estimated base64 data URL length from the encoded blob size.
• Skip base64 conversion for candidates that already exceed the size budget.

Why

Image compression previously converted every canvas candidate to base64 before checking whether it fit. Large images repeated this work across each quality and downscale attempt.

In CI, the existing 12-test image compression suite dropped from 28.99 seconds and a timeout to 5.48 seconds. A focused local run completed in 1.19 seconds.

Checklist

• This PR is small and focused
• I explained what changed and why
• UI screenshots are not applicable
• UI motion video is not applicable

Model: GPT-5.6; harness: T3 Code.

---

Note

Low Risk
Localized performance change in image encoding with the same budget checks; no auth, security, or API surface changes.

Overview
Image compression no longer converts every encode attempt to a base64 data URL before checking the stash/send size budget.

encodeCanvas (renamed from encodeToDataUrl) takes a character budget and rejects oversized results early: toDataURL paths compare string length, and OffscreenCanvas paths estimate base64 length from the blob size before calling blobToDataUrl. The quality ladder in encodeWithinBudget only returns encodings that fit the budget instead of keeping the smallest over-budget candidate from each step.

This cuts redundant base64 work across quality and downscale retries, which drove a large drop in CI time for the image compression test suite.

^{Reviewed by Cursor Bugbot for commit 2a67bcf. Bugbot is set up for automated code reviews on this repo. Configure here.}

Note

Skip base64 encoding for oversized image candidates in encodeWithinBudget

• encodeCanvas in imageCompression.ts now accepts a budgetChars parameter and returns null for dataUrl when the encoded size exceeds the budget, avoiding unnecessary base64 conversion for large blobs.
• encodeWithinBudget now returns null when no encoding fits within the budget, rather than falling back to the smallest over-budget encoding.
• Behavioral Change: callers that previously received an over-budget data URL will now receive null when no candidate fits.

^{Macroscope summarized 2a67bcf.}

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro Plus

Run ID: 79077dd6-6fe3-4e6d-abbe-fae7e7c0703b

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[macroscopeapp]

Approvability

Verdict: Approved 2a67bcf

This is a straightforward performance optimization that adds early-exit logic to avoid expensive base64 encoding when the output would clearly exceed the budget. The caller already handles null returns correctly, and user-facing behavior is unchanged.

^{You can customize Macroscope's approvability policy. Learn more.}