Visitar URL original
fix(workflow): terminate finite port probing without floating-point progress by BlueboxC · Pull Request #1110 · tt-a1i/archify · GitHub
Skip to content

fix(workflow): terminate finite port probing without floating-point progress - #1110

Draft
BlueboxC wants to merge 1 commit into
tt-a1i:devfrom
BlueboxC:codex/upstream-finite-workflow-probe-1072
Draft

BlueboxC wants to merge 1 commit into
tt-a1i:devfrom
BlueboxC:codex/upstream-finite-workflow-probe-1072

Conversation

@BlueboxC

@BlueboxC BlueboxC commented Oct 25, 2026 •

Copy link
Copy Markdown

Problem and value

Fixes #1072. Public Workflow validation can fail to terminate for valid finite widths such as 1e30: adding the 12px port-probe spacing no longer changes representable coordinates.

Stop automatic probing when neither signed coordinate is distinct, or when the next loop increment itself cannot progress. Existing fallback, authored dimensions, ordinary routing and schema limits remain unchanged. No arbitrary maximum width or router rewrite is introduced.

Scope

Based on inspected dev 10b4dd5c5f2264baa315ee65a9b4cc23c13ea5f0. This isolates the progress guard and public-CLI regression from the unrelated #1070/#1071 fixes in the earlier fork commit. The canonical ZIP is rebuilt from this narrow source.

Verification

  • Red: untouched dev exceeds a bounded five-second public validate workflow --json deadline at width1e30.
  • Green: widths1e18,1e25,1e26,1e30 and ordinary320 complete through the public CLI, return structured receipts, preserve input bytes and write no HTML. Timeout cleanup owns the entire spawned process group.
  • All 292 focused Workflow tests pass, with zero skips.
  • npm run test:generated, canonical ZIP build and git diff --check pass using official Node22.23.2 on macOS arm64.
  • Final-head CI completed: 16 successful checks, four conditional skips, zero failures. This includes Windows portability on Node22/24, Ubuntu/macOS/Windows package smokes, Node18/20/22/24 tests, both browser shards, WebM decode and ZIP freshness. Conditional scope skips are not a complete-suite replay. Maintainer acceptance and personal human approval remain pending; earlier installed-fork evidence retains its original revision.

AI assistance and accountability

  • AI assistance: investigation, implementation, tests and writing using Codex; exact model identifier unavailable.
  • Accountable contributor: @BlueboxC.
  • I reviewed the submitted changes and explanation, understand their impact, and can answer questions about them.

Draft: the personal human-review confirmation remains unchecked.

BlueboxC

@coderabbitai

coderabbitai Bot commented Oct 25, 2026

Copy link
Copy Markdown

Important

Draft PR not reviewed

Draft PRs are not automatically reviewed by default.

  • Trigger a manual review

To automatically review draft PRs, update your CodeRabbit configuration:

reviews:
  auto_review:
    drafts: true
  • Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

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

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

@tt-a1i tt-a1i left a comment

Copy link
Copy Markdown
Owner

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Agent-assisted review; visual inspection and local probes below were performed by agents, not claimed as human acceptance.

Reviewed head b51dda9112fe94d1d874acdf3d150898ae4ff33a against current dev e837dc4cb83042d2b6c596e039692e25ec62c2da.

The progress guard is appropriately scoped: it stops when neither signed probe can change the floating-point coordinate, or the loop increment cannot advance, while preserving the fallback and authored dimensions. I independently reproduced the current-dev timeout at width 1e30 and confirmed that the extracted candidate handles 1e18/1e25/1e26/1e30 and the ordinary 320 control promptly, preserves input bytes, and writes no HTML during validate. Four maintained inputs at both quality profiles retain identical complete compile results. No blocking source defect found. Please keep the PR in Draft until the accountable contributor completes the personal-review confirmation, then synchronize current dev and verify final-head integration CI; no arbitrary dimension limit is requested.

This branch has not been deployed

No deployments
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants