fix(cli): correct the auto mode provider list - #396
Open
Gakshith wants to merge 2 commits into
Open
Conversation
The list was one flat set, but the default set differs by language: Python bundles 7 adapters, TypeScript bundles 9. So the old line was wrong either way. LiteLLM is opt-in in both. Ollama and Google GenAI were missing for Python, OpenRouter and Writer for TypeScript, and Cohere was listed as automatic when it only is on TypeScript. Also says the instrumentation adapter is bundled rather than the provider SDK, to match the paragraph further down.
PR respanai#323 edited skills/references/tracing.md but did not rerun the generators, so the CLI bundle and both plugin copies still shipped the old table. This is generator output only, plus the release intent.
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
Line 33 of
skills/references/tracing.mdgives one flat list of the providers a bareRespan()picks up. The default set isn't the same in both languages though — Python bundles 7 adapters, TypeScript bundles 9, and they don't line up — so a single list can't be right for both. It wasn't:_FRAMEWORK_REASON, so it doesn't double-count spans the OpenAI adapter already traces), and there's no LiteLLM entry in the JS registry at all.Quick check for Python:
prints
['anthropic', 'aws-bedrock', 'google-genai', 'ollama', 'openai', 'together', 'vertexai'], the same set therespan-ai auto-instrument smoke checkstep in ci.yml asserts. The JS side comes out ofjavascript-sdks/respan/src/_auto_instrumentation_registry.tsand matches the table injavascript-sdks/respan/README.md.The per-language table at line 108 already has all of this right. Line 33 just never got updated along with it in #323.
This one matters a bit more than a normal doc line, because
respan setupwrites the file straight into~/.agents/skills/and~/.claude/skills/(setup-base-command.ts:738). Right now it tells a Python user on Cohere or LiteLLM that one line is enough. They get no spans and no error.What changed
pip install openai.scripts/build-plugins.mjsandgenerate:skill-refs. feat: expand auto-instrumentation onboarding #323 changed the source but didn't rerun them, so the CLI bundle and both plugin copies have been shipping the pre-feat: expand auto-instrumentation onboarding #323 table since Aug 7 — including| Cohere | cohere | — |, which is backwards in both columns now. That's most of the diff here and it's all generator output.Release Intent
.release-intents/20260822-auto-provider-list.jsonwith{"@respan/cli": "patch"}, sinceskill-refs.generated.tslives in that package. Same shape as #346.Validation
.release-intents/*.jsonfile if this PR changes a release-managed packageWhat I ran:
release_inventory.py --validateandrelease_intents.py validate— both passrelease_intents.py missing --changed-from main --changed-to HEAD— emptypython3 -m unittest discover -s scripts/tests -p 'test_*.py'— 11 tests, OKtsc --noEmiton the regenerated file — clean, and it round-trips byte-identical to the markdown@respan/clionly, Python matrix is 0Notes
Refs #257 and #323. Not closing #257, since #285 already did the bundling half of it.
Two follow-ups I'm happy to pick up, left out to keep this small: