Repository navigation
docs(browser_cli): stop claiming UNSUPPORTED read both bundles - #20897
Conversation
UX Review (Opus 5.5) — ⏭️ skippedRevision |
GPT 6.1 Review — ✅ no blocking findingsGPT 6.1 completed its review of This comment is updated in place on each push. Review detailsNo findings. False positive or not applicable? A repository writer can comment: |
The UNSUPPORTED docstrings on cli_lifecycle_env_support and _warn_lifecycle_loss, and the test that stands in for that verdict, said both bundles were read. _seams_support returns UNSUPPORTED at the first bundle whose hook is missing, so when registry.js lacks its hook, coreBundle.js is never read. Each sentence says what holds on every branch instead: a bundle was read and its hook is gone. Prose only; no behaviour change. Fixes #10699
587a2d7 to
b07413a
Compare
Design Review (Opus 5.5) — ✅ PASSDesign-level review of Design-Verdict: PASS Nothing to check. [DESIGN-REVIEWED] b07413a Description read[DESCRIPTION-READ] 5834e79849b401fe22c511e4507ed9cde6a483243925b6edec023abeb81178e9 That sha256 names everything this verdict read: the PR description, plus the 1 evidence file(s) the workflow collected from it and handed this review. It covers that evidence because the media strip replaces every attachment URL with the same placeholder, so an edit swapping one attachment for another would leave the description's captured text identical while what this review judged changed. Recompute it with |
First Principles Review (Opus 5.5) — ✅ PASSPremise-level review of First-Principles-Verdict: PASS Nothing to check. What this change shipsInventory (3 items) — 3 justifiedIntent: a FIX. Three docstrings now say what an
[FIRST-PRINCIPLES-REVIEWED] b07413a Description read[DESCRIPTION-READ] 8b04264cfc9294232f9dcbb8a18b82343a6868a2220d8f6f8c9edf86c4b42e10 That sha256 names the PR description this verdict read. Recompute it with |
Opus 5.5 Review — ✅ no blocking findingsReviewed Verdict parsed from the review's SHA-scoped output markers for commit False positive or not applicable? A repository writer can comment: |
chenmingwei23
left a comment
There was a problem hiding this comment.
Approving: readiness: passed, all lanes green, no outstanding change requests.
Problem / Motivation
Goal: The
UNSUPPORTEDseam-verdict docstrings inbrowser_clisay only what the seam scan guarantees: a bundle was read and its hook is gone.Three docstrings say an
UNSUPPORTEDverdict means both bundles were read. The scan stops at the first missing hook, so whenregistry.jsis the bundle missing it,coreBundle.jsis never read and the claim is false.Why it matters
These docstrings tell a reader what an
UNSUPPORTEDverdict proves. Read as written, it proves both hooks were measured, so a reader can takecoreBundle.jsas checked when onlyregistry.jswas. A later change to the gate or its warning built on that reading would rest on a check that never ran.Not a goal
_seams_supportkeeps reading bundles in order and stopping at the first missing hook.What changed (motivation → approach → change)
The claim is wrong because of how
_seams_support(browser_cli/install.py) walks its seams: in order, returning at the first needle it does not find. For the lifecycle gate the first seam isregistry.jsand the second iscoreBundle.js(main @ 6d4f2874e8, install.py L1359-L1406):Each sentence is reworded to what holds on every branch. Docstrings only; no code changes.
browser_cli/install.pycli_lifecycle_env_supportbrowser_cli/launch.py_warn_lifecycle_losstest/test_browser_cli_launch.pytest_a_measured_missing_hook_still_names_the_installed_cliThe owning spec,
docs/system-specs/modules/browser.md, already saysUNSUPPORTEDmeans the serving bundle was read and a needle is gone, so it is unchanged.Backwards compatibility
Compatible: docstrings only; no code, signature, verdict or log text changes.
Removes nothing: only wording inside three docstrings changes.
Tests
No new test: the change is prose, and no test reads either docstring. In
test/test_browser_cli_launch.pyonly the docstring oftest_a_measured_missing_hook_still_names_the_installed_clichanges; its body is untouched.test/test_browser_cli_*.pyfiles: 509 passed at-n 2.-n0; 10/10 at-n 4 --dist loadgroup --max-worker-restart=2besidetest/test_browser_cli_token.pyandtest/test_browser_cli_snapshots.py; the whole file in a shuffled order, seed 10533, 44 passed.test/test_flake_pattern_ratchet.py: 56 passed.Manual verification
N/A — prose only; no runtime path changes.
git grep -n -i "both bundles" -- src test docsfinds no remaining copy of the claim.Related Issues
Fixes #10699
Checklist
feat|fix|docs|style|refactor|perf|test|chore|ci|build|revert: ...)