Repository navigation
Conversation
# Conflicts: # apps/tester-federation-v2/package.json # pnpm-lock.yaml
…ures-to-mf2-testing-app
…es-to-mf2-testing-app
|
@MrMuzyk is attempting to deploy a commit to the Callstack Team on Vercel. A member of the Team first needs to authorize it. |
|
Purely for test pruposes |
|
Apex SentinelVerdict: 1 finding(s): 1 minor. 🟡 Minor (1)
7 findings below Not in the reviewers' prompt (8 files)
Reviewed with this repository's own Sentinel configuration ( Feedback goes in the review comment threads, as a reply under the inline comment of a finding: Reply Useful? Reply Comment |
There was a problem hiding this comment.
Apex Sentinel: 1 new finding on 7aeb22c. The verdict and the full list are in the summary comment.
| useAnimatedReaction( | ||
| () => width.value, | ||
| (_previous, next) => { | ||
| 'worklet'; | ||
| if (next !== null) { | ||
| scheduleOnRN(onWidthChange, Math.round(next)); | ||
| } | ||
| } | ||
| ); | ||
|
|
There was a problem hiding this comment.
🟡 Minor · useAnimatedReaction callback reports the previous frame's width
useAnimatedReaction passes (current, previous), but the parameters are named _previous, next, so the code schedules next — the previous frame's width. reanimated.e2e.ts passes only because the last two frames of a settling spring round to the same integer.
| useAnimatedReaction( | |
| () => width.value, | |
| (_previous, next) => { | |
| 'worklet'; | |
| if (next !== null) { | |
| scheduleOnRN(onWidthChange, Math.round(next)); | |
| } | |
| } | |
| ); | |
| useAnimatedReaction( | |
| () => width.value, | |
| (next, _previous) => { | |
| 'worklet'; | |
| if (next !== null) { | |
| scheduleOnRN(onWidthChange, Math.round(next)); | |
| } | |
| } | |
| ); |
Details
persona: maintainability
Useful? Reply @apex-sentinel-staging like, or @apex-sentinel-staging dislike <why> if it is not.
Reply @apex-sentinel-staging note ... to add context, or @apex-sentinel-staging pivot ... if it is wrong — I will re-check the cited code.
artus9033
left a comment
There was a problem hiding this comment.
A few notes on the persona's decision rules. The main ones are mixed PRs and how strict covered is: both can produce a confident wrong "already covered".
| lane apps, opt-in features the lanes do not enable). Also when the PR | ||
| changes only the tester apps themselves (`apps/**`) and no Re.Pack package: | ||
| the lanes test Re.Pack, not their own apps. | ||
| 2. **`covered`**: the lines this PR edits execute during an existing e2e |
There was a problem hiding this comment.
What happens when only part of the PR is covered? If a PR edits two runtime files and a lane runs only one of them, "pick the first that fits" will most likely land on covered. Then the green run looks like proof for the other file too. Suggest: covered only when every edit that runs on the device is covered; otherwise answer for the uncovered part and mention the covered files in the reason.
| Judge the code this PR edits, not only the bug it says it fixes. A PR can fix | ||
| a bug that shows only on real devices and still edit code the lanes run in | ||
| every test. A PR can also fix one thing and break another. So ask: if this | ||
| edit were wrong in any way, including side effects on behaviour the PR does |
There was a problem hiding this comment.
"Wrong in any way" is a bar nothing passes. Meanwhile line 50 ("runs on every build and a wrong edit would stop … loading") is a bar almost everything passes in a bundler, even though a subtle wrong edit (a chunk name, a lost option) won't stop loading. Maybe "the plausible ways this edit could be wrong", and have the every-build route name what, if wrong, breaks loading?
| not mention, would an existing e2e file go red? | ||
|
|
||
| 1. **`not_useful`**: the change cannot alter what runs on the device | ||
| (docs, CI, types only, tests only, tooling, log text, examples outside the |
There was a problem hiding this comment.
Two items in this list:
- "tooling": Re.Pack is build tooling, so its CLI, plugins and loaders could all be read as this. Maybe "repo tooling (lint, release scripts)".
- "opt-in features the lanes do not enable": these do change what runs on the device, just not in the lanes. Since step 1 comes first, they can never reach
would_help(a lane screen could enable the option) orout_of_reach. I'd let steps 3 and 5 decide them.
| the lanes test Re.Pack, not their own apps. | ||
| 2. **`covered`**: the lines this PR edits execute during an existing e2e | ||
| file's run, and that file checks what they produce, so a wrong edit would | ||
| turn it red. This holds even when the bug the PR fixes shows only outside |
There was a problem hiding this comment.
When the bug only shows outside the lanes, a reader will still take "already covered" to mean the fix was checked. Maybe have the reason say "a regression here would show; the fix itself isn't exercised".
| lane screen could make them execute: the lanes never run that setup (see | ||
| "What the lanes never reach" in `AGENTS.md`), or no emulator test could | ||
| exercise it at all. A green e2e run says nothing here. Name the reason. | ||
| 4. **`not_useful`**: the change is logic that does not depend on the device, |
There was a problem hiding this comment.
Some cases with no clear answer:
- Deletions: there are no edited lines left to execute. Ask instead whether a test checked the removed behaviour.
- PRs that add or weaken an e2e test: is a newly added test "existing"? And "tests only → not useful" hides a PR that loosens an assertion.
- Device-independent logic with no unit test: this falls through to
would_helpwith an e2e sketch, when "add a unit test" is the honest answer.
Summary
Test plan