chore: release v0.8.27 - #3608
Conversation
|
Preview deployment for your docs. Learn more about Mintlify Previews.
💡 Tip: Enable Automations to automatically generate PRs for you. |
jrusso1020
left a comment
jrusso1020
left a comment
There was a problem hiding this comment.
Approving at e42a1d31a0f450def3709674a0937e26bc645be0.
A release PR's one real risk is claiming to carry something it does not, so that is what I checked first. Both commits the notes cite are ancestors of this head - compare/<commit>...<head> reports ahead for 5a5e841c (#3606) and for ae7e530f (#3598) - so the tag cut from this branch will contain them rather than merely mention them. #3606 merged at 03:53:45Z and #3598 at 03:25:00Z, both before this branch was cut.
The bump is uniform and the diff carries no code. All thirteen packages/*/package.json read 0.8.27, and so do the three plugin manifests. The remaining files are docs/changelog.mdx and the new releases/v0.8.27.md. Nothing else is in the diff, which is what a release PR should look like - a version-only change is the one shape where "the tests passed on main" is still evidence about what is shipping.
The notes describe the user-visible effect rather than the mechanism, which is right for a changelog, and both sentences are true of the merged code. On the favicon half I reviewed the implementation itself at c6f4b57 and the ranking behaves as the note says: sharpest declared candidate first, with object-fit: contain and no invented background, so a transparent mark stays transparent.
Perf: scrub and player-perf are RED at this head and I am approving over them deliberately, so here is the reasoning rather than an override with no note.
The failure is not a perf result. Perf: scrub never reached its measure step - it died in bun run build with File 'packages/core/src/generated/runtime-inline.ts' is not a module, and Run player perf - scrub (measure mode) is marked skipped underneath it. Perf: drift, fps and parity are cancelled, which is the matrix failing fast around it rather than three more findings.
Three things say that is infrastructure and not this cut:
Buildis SUCCESS on this exact commit, so the repo does build.Perf: loadis SUCCESS on this exact commit, and it runs the samebun run buildin the same matrix - one shard read a generated file another had not finished writing.- The diff is thirteen version strings, three plugin manifests and two documents. There is no path from it to the codegen of
runtime-inline.ts.
None of Perf: * or player-perf is in this branch's required set, so they will not gate the merge either way. Worth a re-run before you merge, so the release does not land with a red lane on its commit that somebody has to re-derive later. The rest of CI is still in flight - regression shards, the windows lanes, CLI smoke (required) - so I am not calling the required set green.
The merge is yours and it is gated on those finishing; I am not merging or queueing this, and this repo requires an approval on the last push, so any further push needs a fresh one.
Review by Rames
jerrai-bot-heygen
left a comment
jerrai-bot-heygen
left a comment
There was a problem hiding this comment.
Reviewed release v0.8.27 at exact head e42a1d31a0f450def3709674a0937e26bc645be0.
The release diff is limited to a uniform 0.8.27 version bump, release notes, and changelog. It contains the already-merged #3598 missing-asset accounting and #3606 favicon-ranking behavior; package/plugin versions are aligned and no lockfile update is required for workspace dependencies. Earlier transient perf shard failures were rerun; current completed checks are green and remaining jobs are pending.
— Jerrai
jerrai-bot-heygen
left a comment
jerrai-bot-heygen
left a comment
There was a problem hiding this comment.
Re-approved at exact head 00b3fad70e0927442197828f25a847264063fab4 after the main-only merge. The release bump content remains unchanged and verified; CI has restarted at this head with no failures.
3 participants
No description provided.