Skip to content

chore(release): prepare 0.8.3 - #216

Merged
pierreraby merged 1 commit into
mainfrom
release/0.8.3
Nov 13, 2026
Merged

pierreraby merged 1 commit into
mainfrom
release/0.8.3

Conversation

@pierreraby

Copy link
Copy Markdown
Collaborator

Release 0.8.3 (patch, latest)

Third patch on the 0.8 line. Two user-facing fixes and one pricing refresh, no breaking changes, no migration.

Included

Change Origin
Responses wire repair for zai-org/GLM-5.3 (400 input item is missing type) and for the inert reasoning.effort: "none" that zai-org/GLM-5.3 and mistral/mistral-large-4 reject #215, reported in #214 by @ThomasByr
Reviewed zero pricing for thinkingmachines/inkling-2-small:free, API/pricing snapshots refreshed to 98 models (2026-11-13) #213 by @sandexzx

No runtime, routing-default, workflow, dependency or version-scheme changes. No new configuration, no migration, no renamed IDs.

Verification recorded for this release

  • Preflight: dist-tags latest=0.8.2 / next=0.7.3-next.0; latest GitHub Release v0.8.2; npm view pi-commandcode-provider@0.8.3 returned E404 (version unused); no local or remote v0.8.3 tag; no active release run.
  • Release contract: the real validateRelease() accepted the tree — version 0.8.3, dist-tag latest, notes 2 288 bytes, contributors present, no leakage of previous sections. tests/test-release.mjs 49/49.
  • Local gates: npm test with PI_LOCAL_REQUIRED=1 OMP_COMPAT_REQUIRED=1 on real hosts (pi 1.5.2, @oh-my-pi 18.18.3) PASS, 11 host suite lines green, 0 failures. prettier --check clean, npm audit --audit-level=moderate 0 vulnerabilities, npm pack --dry-run 32 files / 84.3 kB, git diff --check clean. actionlint is not installed locally; CI covers it.
  • Live catalog/pricing: the scheduled workflow passed on 6094904 (Models checked: 98, Status: PASS, no drift), and the pricing snapshots in this branch are the reviewed 2026-11-13 ones. No expected upstream drift was suppressed.
  • Live API checks (separately authorized, paid account): the two repaired models were exercised against POST /provider/v1/responses before and after, giving 400 → 200 with a real answer. Not part of CI; mock suites still do not prove upstream behaviour.
  • Scope note: known limitation unchanged — /provider/v1/models still advertises two ids that answer 404 on every route, reported upstream as CommandCodeAI/command-code#1144. Nothing is masked locally on purpose.

Merge and publication

Merge is a squash onto main, then the annotated v0.8.3 tag goes on the merge commit and the OIDC release workflow publishes. Tag is pushed only after main is green.

Patch release carrying two user-facing fixes and one pricing refresh:

- Responses wire repair for zai-org/GLM-5.3 and mistral/mistral-large-4 (#215,
  reported in #214).
- Reviewed zero pricing for thinkingmachines/inkling-2-small:free (#213).
@github-actions

Copy link
Copy Markdown
Contributor

Dependency Review

✅ No vulnerabilities or license issues or OpenSSF Scorecard issues found.

Scanned Files

None

@github-actions

Copy link
Copy Markdown
Contributor

Memory benchmark

Compared base 6094904 with PR head 4796eb8 on the same GitHub-hosted darwin runner. Lower values are better.

Metric Base PR PR − Base Change
Stable RSS 104.9 MiB ± 0.9 MiB 105.6 MiB ± 0.6 MiB +0.6 MiB ± 0.5 MiB +0.6%
Physical footprint 60.8 MiB ± 0.6 MiB 61.3 MiB ± 0.6 MiB +0.2 MiB ± 0.8 MiB +0.3%
Physical peak 71.6 MiB ± 0.6 MiB 73.1 MiB ± 0.7 MiB +1.3 MiB ± 0.4 MiB +1.8%

Interpretation: Differences larger than their measured MAD: Stable RSS +0.6 MiB ± 0.5 MiB, Physical peak +1.3 MiB ± 0.4 MiB.

Extension overhead above pi baseline

Metric Base overhead PR overhead Difference
Stable RSS +3.9 MiB ± 1.0 MiB +4.5 MiB ± 0.3 MiB +0.6 MiB ± 0.5 MiB
Physical footprint +2.9 MiB ± 0.9 MiB +3.1 MiB ± 0.8 MiB +0.2 MiB ± 0.8 MiB
Physical peak +1.8 MiB ± 0.8 MiB +3.1 MiB ± 2.0 MiB +1.3 MiB ± 0.4 MiB

Values are medians of 6 alternating, paired runs. The value after ± is the median absolute deviation (MAD). Each process was sampled 12 times after a 2500 ms warm-up.

Environment: pi 0.84.4, Bun 1.4.4, darwin arm64. RSS comes from ps; physical footprint and peak come from macOS footprint.

This is a comparative signal, not a pass/fail threshold. GitHub-hosted runner noise can affect absolute values.

@pierreraby
pierreraby merged commit 2a5c55d into main Nov 13, 2026
15 checks passed
@pierreraby
pierreraby deleted the release/0.8.3 branch November 13, 2026 12:25
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant