release: v0.1.1242 - #3828
Merged
Merged
release: v0.1.1242#3828
Conversation
Ships six fixes that have been stranded on main. 0.1.1241 was published on 2026-08-17 at ~11:00 UTC and main stayed at that version, so every fix merged after it could not reach any environment: the release job reads deno.json, sees no version change, and correctly does nothing. Included: #3817 skip Azure content-filter preamble frames (issue-inbox#543) #3818 route Azure GPT models through chat (issue-inbox#548) #3820 replay provider metadata across tool steps (issue-inbox#549) #3821 preserve every signed Gemini tool step #3826 withhold replay metadata when a tool call was suppressed #3822 record upstream status on failed-closed metadata lookups #3823 surface platform errors from durable run setup as their real status #3799 publish the embedded root route once, never as a dotfile #3796 report ignored feature file trees #3818 and #3817 must ship together: routing Azure GPT models to chat completions is what makes Azure's prompt_filter_results preamble reachable, and #3817 is what stops that preamble failing the stream parser. After this publishes, veryfront-agent needs its pin bumped separately. The release dispatches to veryfront-server, veryfront-job-runner and veryfront-sandbox but not to veryfront-agent, which pins its own npm copy -- so agent runs keep the old framework until that bump lands.
📦 Client bundle boundary
A server module in a client graph aborts hydration in the browser. New leaks fail CI; known leaks are tracked in |
CodeQL default setup failed at init with "No server is currently available" during the 2026-08-17 GitHub incident. Default-setup runs cannot be re-run via the API, so this empty commit retriggers the checks. No file changes.
|
Important Review available on request
Reviews should be triggered manually for repositories with fewer than 10 stars. Select Trigger review above or comment ⚙️ Run configurationConfiguration used: Repository UI Review profile: CHILL Plan: Pro Plus Run ID: Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
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
1 participant
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.
Ships six fixes that have been stranded on
main.Why they were stuck
0.1.1241was published at ~11:00 UTC today, andmainstayed at0.1.1241. Every fix merged after that point could not reach any environment — the release job readsdeno.json, sees no version change, and correctly does nothing.src/utils/version.test.tsasserts thatVERSIONmatchesdeno.json, not that it increased. So main can accumulate merged fixes at an already-published version indefinitely with CI fully green. That is worth fixing separately; this PR just unblocks what is queued.What ships
#3817 and #3818 must ship together. Routing Azure GPT models to Chat Completions is what makes Azure's
prompt_filter_resultspreamble reachable, and #3817 is what stops that preamble failing the stream parser. Shipping #3818 alone would trade a 404 forinvalid successful stream (empty choices event had no usage).Impact once deployed
This is what unblocks the model catalogue. Measured on staging today, 32 real agent runs (16 models × text + tool): 26 passed, 6 failed, and every failure is fixed by a PR in this release.
Follow-up required — this release alone is not enough
veryfront-agentneeds a separate pin bump. The release dispatchesveryfront-code-releasedtoveryfront-server,veryfront-job-runnerandveryfront-sandbox— but notveryfront-agent, which pins its own npm copy inpackage.json. Two runtimes execute agent runs, and today only one of them auto-updates. Earlier today Kimi stayed broken onveryfront-agentfor 20 minutes afterveryfront-serverhad the fix, which is how this gap was found.So: merge → publish →
veryfront-serverdeploys automatically → manually bumpveryfront-agent(all fiveveryfront/@veryfront/ext-*packages move together;frameworkDependency.test.tsenforces it) → re-run the model matrix.Verification
src/utils/version.test.ts— 1 passed (21 steps). All three version files agree at0.1.1242:deno.json,src/utils/version-constant.ts, and the prebundled hydration runtime that embeds the version string.0.1.1242is above the published0.1.1241.