Skip to content

acpmux: the upload-link test names no literal ~/.acpmux state root (base red) - #26087

Closed
austinywang wants to merge 1 commit into
feat-cmux-nextfrom
next-state-roots-upload-fixture
Closed

austinywang wants to merge 1 commit into
feat-cmux-nextfrom
next-state-roots-upload-fixture

Conversation

@austinywang

@austinywang austinywang commented Oct 27, 2026 •

Copy link
Copy Markdown
Contributor

Summary

cmux-next checks fails on feat-cmux-next and on every PR, at the step "No literal state roots (per-build isolation S7, cx-is1kf.8)":

check-state-roots: cmux-tui/crates/acpmux/src/prompt_uploads.rs has 1 literal state root line(s), baseline 0:
  cmux-tui/crates/acpmux/src/prompt_uploads.rs:345:            link(format!("/home/me/.acpmux/uploads/chat/{id}/a.png")),

That line is a fixture in the inline #[cfg(test)] module, which the ratchet counts. It came in with #25840. upload_ids only takes the path's parent folder as the upload id, so this PR changes the fixture to /home/me/uploads/chat/{id}/a.png. The test still checks the same behaviour, and the baseline stays at 0 instead of gaining a row.

Testing

  • bash scripts/cmux-next/check-state-roots.sh . exits 1 on the base tip and 0 on this head (check-state-roots: ok).

Observed / Not run

  • Observed: the state-roots check above.
  • Not run locally: cargo test -p acpmux a_link_names_its_upload_on_any_platform. Rust isn't built on this Mac, so this PR's cmux-tui CI runs it. Only a string literal in the test changed, and the parent folder is still the uuid.

Changelog

none

🤖 Generated with Claude Code


View with [code]smith Autofix with [code]smith
Need help on this PR? Tag @codesmith-bot with what you need.

…ase red)

check-state-roots (cmux-next checks, "No literal state roots") counts inline
#[cfg(test)] modules, and a_link_names_its_upload_on_any_platform used
/home/me/.acpmux/uploads/... as its fixture path since #25840, so the check
is red on feat-cmux-next and on every PR. upload_ids reads only the path's
parent folder as the upload id, so a path without .acpmux tests the same thing.

Test: bash scripts/cmux-next/check-state-roots.sh . exits 1 on the base, 0 here.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@github-actions

github-actions Bot commented Oct 27, 2026 •

Copy link
Copy Markdown
Contributor

cmux status (one comment, edited in place; each line is updated by the job that owns it)

Review 771c6ab: HQ feed card · gallery before/after runs · checks

App of this head (app.zip, kept 1 day)

gh workflow run reload-build.yml -R manaflow-ai/cmux --ref feat-cmux-next -f tag=pr26087 -f ref=next-state-roots-upload-fixture -f platform=macos -f backend_mode=local

web and size checks: green

@blacksmith-sh

blacksmith-sh Bot commented Oct 27, 2026 •

Copy link
Copy Markdown

Found 6 test failures on Blacksmith runners:

Failures

Test View Logs
main.DeveloperIdAppTests.test_download_prints_the_active_profile_only/
test_download_prints_the_active_profile_only
View Logs
stubs: /System/Library/Frameworks/Security.framework/Versions/A/Security has no stub/
Security has no stub
View Logs
stubs: CoreFoundation.tbd: install-name /System/Library/Frameworks/
CoreFoundation.framework/Versions/A/CoreFoundation, missing 1: ['_CFRetain']/
CoreFoundation, missing 1: ['_CFRetain']
View Logs
stubs: CoreFoundation.tbd: install-name /System/Library/Frameworks/
CoreFoundation.framework/Versions/A/CoreFoundation, missing 1: ['_log2']/
CoreFoundation, missing 1: ['_log2']
View Logs
test_the_nightly_is_covered (main.NightlyStepCoverageTests.test_the_nightly_is_cove
red)/
test_the_nightly_is_covered (main.NightlyStepCoverageTests.test_the_nightly_is_cove
red)
View Logs
test_the_repository_highlights_validate (main.DigestTests.test_the_repository_highl
ights_validate)/
test_the_repository_highlights_validate (main.DigestTests.test_the_repository_highl
ights_validate)
View Logs

Fix with [code]smith
Need help on this PR? Tag @codesmith-bot with what you need.

@github-actions github-actions Bot added the risk:normal Risk tier (docs/ci/risk-tiers.md) label Oct 27, 2026
@austinywang

Copy link
Copy Markdown
Contributor Author

Review (claude-subagent correctness review): SHIP at 771c6ab

An independent subagent reviewed the exact head and found no defects.

  • The test checks the same behaviour. upload_ids (prompt_uploads.rs:91-99) takes the uuid from the path's parent folder. Dropping .acpmux/ changes only folders higher up, so the parent of a.png is still {id}. The Windows fixture, the bare a.png case and the assertion are unchanged.
  • The check clears. bash scripts/cmux-next/check-state-roots.sh . passes on this head. On the base tip 3a04f3c29ba it fails, and prompt_uploads.rs:345 is the only line it reports.
  • Nothing else matches. --list finds no other line in the file, so state-roots-baseline.tsv needs no row. The Windows path C:\Users\me\acpmux\... doesn't match the pattern.
  • No duplicate. No open PR fixes the same line.
  • Base moved. feat-cmux-next has moved one commit since the merge base (nightly-next: read the Dev ring switch live when publishing #26077, nightly ring files only). That commit doesn't touch this file.

Not run locally: cargo test -p acpmux. This PR's CI runs it.

Resolution: nothing to change.

@austinywang austinywang added the ready-to-land Reviewed at this head; next-autoland merges it into feat-cmux-next when its only reds are base reds label Oct 27, 2026
austinywang added a commit that referenced this pull request Oct 27, 2026
…ase red)

Same change as #26087, combined here because each is red on the other's red:
check-state-roots counts the inline test module's /home/me/.acpmux/... fixture
(since #25840). upload_ids reads only the parent folder, so the test is the same.

Test: bash scripts/cmux-next/check-state-roots.sh .: exit 1 on the base, ok here.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@austinywang

Copy link
Copy Markdown
Contributor Author

The same fixture fix is now part of #25839 (commit d66d8d5). Each PR was red on the other's base red, so they're combined. Once 25839 lands, this PR's diff against the base becomes empty, and I'll close it.

@austinywang

Copy link
Copy Markdown
Contributor Author

Closing as redundant. #26278 fixed the same state-roots fixture on feat-cmux-next, using /home/me/state/uploads/..., so 'No literal state roots' is no longer red on the base. #25839 dropped this file when it merged the base.

cmux-bot Bot pushed a commit that referenced this pull request Oct 30, 2026
* docs(next): the Cloud start-session highlight links its docs (release-notary red on base)

test_the_repository_highlights_validate fails on every feat-cmux-next PR: the
highlight from #25625 has no tryIt action or docs link.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* test(nightly): read the legacy publish job with its moved step scripts inline

#25879 moved "Generate the old feed's
appcast" into scripts/ci/nightly-steps/legacy-identity-appcast.sh, so
test_the_appcast_comes_from_the_tested_plist no longer found its commands in
nightly.yml (ValueError: substring not found) and release-notary went red on
feat-cmux-next. Read the workflow through nightly_workflow_text.nightly_text(),
as the other nightly workflow tests do.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* test(nightly): leave test-execution.toml as on the base

The pre-commit hook registered two base tests that this PR does not touch;
keep this PR to the two red fixes.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* release notes: two new fixed highlights get their Try it action (base red)

home-composer-border (#26119) and selected-harness-start (#26106) landed with
neither an action nor a docs link, which scripts/whats-new/validate.py requires,
so release-notary's test_the_repository_highlights_validate is red on
feat-cmux-next. The Chief field one opens Home; the harness one starts a new
agent chat.

Test: python3 scripts/whats-new/test_digest.py: 1 failure before, OK after.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* nightly: the Sparkle key test reads nightly.yml with its moved steps inline (base red)

#26196 and #26210 moved two of the four SPARKLE_PRIVATE_KEY="$(./scripts/ci/nightly-sparkle-key.sh)"
call sites into scripts/ci/nightly-steps/, so the raw workflow has two and
tests/test_nightly_sparkle_key.sh (release-notary) fails its at-least-three check.
It now greps nightly_workflow_text.nightly_text(), the expansion the other
workflow tests read.

Test: bash tests/test_nightly_sparkle_key.sh: FAIL on the base, PASS here.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* cmux-next routing: a test's home-folder Claude settings path is not a repository read (base red)

CloudCredentialTransferTests named ".claude/settings.json" under a temp home
in one literal; ci-target-graph.py found the repository's own file at that
path and listed it as a CmuxNextAppTests read (base autofix c14a002), which
cmux-next.yml's pull_request paths do not cover, so Tier routing's
test_every_routed_input_triggers_the_workflow is red on feat-cmux-next. The
test builds the path from folder and file, and the graph loses the read
(literal_reads recomputed: the target's other 14 reads are unchanged).

Test: python3 tests/test_cmux_next_route.py: 1 failure on the base, OK here.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* acpmux: the upload-link test names no literal ~/.acpmux state root (base red)

Same change as #26087, combined here because each is red on the other's red:
check-state-roots counts the inline test module's /home/me/.acpmux/... fixture
(since #25840). upload_ids reads only the parent folder, so the test is the same.

Test: bash scripts/cmux-next/check-state-roots.sh .: exit 1 on the base, ok here.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* acpmux: collapse the Codex broker's account checks into let chains (clippy base red)

#26191 nested two if-lets in codex_auth_broker.rs; clippy -D warnings
(collapsible_if) fails on it for every PR that touches acpmux. Same checks,
one let chain each.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* release notes: the compact thought-row highlight gets its Try it action (base red)

#26270 added it with neither an action nor a docs link, so release-notary's
test_the_repository_highlights_validate is red again.

Test: python3 scripts/whats-new/test_digest.py: 1 failure before, OK after.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* tests: leave test-execution.toml as the base has it

The pre-commit hook registered unrelated tests while committing the base merge.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* cmux-next tests: drop the settings-path split; the graph skips .claude paths itself

#26248 made ci-target-graph.py ignore .git, .claude and .codex paths a test
names, so the one-literal form no longer routes CmuxNextAppTests on the
repository's settings file. Back to the base's line.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* docs(next): the Cloud start-session highlight links its docs (release-notary red on base)

test_the_repository_highlights_validate fails on every feat-cmux-next PR: the
highlight from #25625 has no tryIt action or docs link.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* docs(next): give the two base highlights release-notary rejects a live action

feat-cmux-next's inline-pr-pills and inspector-error-colours highlights have
no action or docs line, so scripts/whats-new/test_digest.py
(release-notary) fails on the base. Use existing palette actions rather than
https://cmux.com/docs/agents, which returns 404.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* docs(next): give the new Find in Chat highlight a live action (release-notary red on base since 93ff53c)

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Leo <cheerleaderleo@outlook.com>

This branch was previously deployed

3 inactive deployments
artifacts — 771c6ab6 Deployed Oct 27, 2026 by austinywang via add the Windows daemon to the tree #63480
ci-routing-agent — 771c6ab6 Deployed Oct 27, 2026 by github-actions[bot] via cmux-next macOS placement #39123
sentry-dsyms — 771c6ab6 Deployed Oct 27, 2026 by austinywang via upload cmux-tui dSYMs to Sentry #63480
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

ready-to-land Reviewed at this head; next-autoland merges it into feat-cmux-next when its only reds are base reds risk:normal Risk tier (docs/ci/risk-tiers.md)

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant