agentHost: improve Agent Merge CI diagnostics - #335501
Merged
Benjamin Christopher Simmonds (benibenj) merged 2 commits intoSep 10, 2026
Merged
agentHost: improve Agent Merge CI diagnostics#335501Benjamin Christopher Simmonds (benibenj) merged 2 commits into
Benjamin Christopher Simmonds (benibenj) merged 2 commits into
Conversation
Replace prefix-only CI dumps with bounded summaries and authorized, cached tail/range/search reads. Preserve true EOF completeness, redaction, cancellation, and workflow-attempt identity while keeping responses and retained evidence bounded. Serialize concurrent reads and paginate summaries within cache capacity so follow-up evidence remains actionable without changing repair scheduling. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Copilot started reviewing on behalf of
Benjamin Christopher Simmonds (benibenj)
September 10, 2026 16:24
View session
Contributor
There was a problem hiding this comment.
Copilot review overview
🟡 Changes recommended
Paginated searches repeatedly rescan prior lines, making late matches in newline-dense logs potentially exceed the tool deadline.
Once you've addressed the issues Copilot identified, you can request another Copilot review.
Review tier: Balanced
Findings: 1
New issues introduced by this change (1)
| Severity | Finding |
|---|---|
 |
src/vs/platform/agentHost/node/agentMergeCIEvidence.ts — Each search continuation calls linesInRange with a later first, but that generator still walks… |
What changed in this PR
Improves Agent Merge CI diagnostics with bounded summaries, navigable cached evidence, and safer workflow-log handling.
Changes:
- Adds summary, tail, range, search, and continuation support.
- Expands log capture to 16 MiB with redaction and authorization safeguards.
- Adds comprehensive transport, lifecycle, caching, and diagnostics tests.
| File | Description |
|---|---|
src/vs/platform/github/test/node/pullRequestMutationService.test.ts |
Tests attempts, job metadata, limits, and redaction. |
src/vs/platform/github/test/node/githubTransport.test.ts |
Tests bounded streaming, cancellation, and redirects. |
src/vs/platform/github/common/pullRequestMutationService.ts |
Expands and redacts workflow-log capture. |
src/vs/platform/github/common/githubTransport.ts |
Hardens bounded downloads and cleanup. |
src/vs/platform/github/common/githubPullRequestMutationService.ts |
Extends workflow diagnostics contracts. |
src/vs/platform/agentHost/test/node/agentMergeServerTools.test.ts |
Tests tool descriptions and request parsing. |
src/vs/platform/agentHost/test/node/agentMergeRerun.test.ts |
Updates rerun integration expectations and disposal. |
src/vs/platform/agentHost/test/node/agentMergeCITools.test.ts |
Covers diagnostics, caching, paging, and authorization. |
src/vs/platform/agentHost/node/shared/agentMergeServerTools.ts |
Defines diagnostic modes and validation. |
src/vs/platform/agentHost/node/agentServiceComposition.ts |
Owns the disposable Agent Merge tools instance. |
src/vs/platform/agentHost/node/agentMergeTools.ts |
Implements summary-first authorized CI reads. |
src/vs/platform/agentHost/node/agentMergeCIEvidence.ts |
Implements evidence caching and navigation. |
💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.
Cache sparse line-start offsets so search continuations, context excerpts, ranges, and tails do not repeatedly scan logs from the beginning. Cover 500,000-line pagination, newline-dense 16 MiB logs, checkpoint boundaries, and empty evidence. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Benjamin Christopher Simmonds (benibenj)
marked this pull request as ready for review
September 10, 2026 17:49
Henning Dieterichs (hediet)
previously approved these changes
Sep 10, 2026
Ladislau Szomoru (lszomoru)
approved these changes
Sep 10, 2026
Benjamin Christopher Simmonds (benibenj)
merged commit Sep 10, 2026
0b2d79f
into
main
33 checks passed
Benjamin Christopher Simmonds (benibenj)
deleted the
agents/agent-merge-ci-diagnostics-improvements
branch
September 10, 2026 20:26
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Agent Merge previously returned only the first 100,000 characters of job logs whose downloads were already limited to a 2 MiB prefix. This could hide the actual failing-test summary with no supported way to read further.
readAgentMergeCIsummary-first, with failed checks, annotations, jobs/steps, and bounded failure excerpts under a 48,000-byte response budget.No GitHub permission expansion, repair retry scheduling / blocked-outcome changes, or archive-nudge/onboarding changes.
Validation
npm run gulp compile-client— passed with zero errors.npm run typecheck-client— passed.The aggregate
npm run compile-clientalso builds unrelated extensions and encountered a missing extension-sideesbuilddependency in this isolated worktree. The source-only compile task above and the final compiled unit tests passed. No remote CI logs were retrieved for this change.