fix(opencode): sanitize history and gate options on target model switch#16630
Draft
tobwen wants to merge 1 commit intoanomalyco:devfrom
Draft
fix(opencode): sanitize history and gate options on target model switch#16630tobwen wants to merge 1 commit intoanomalyco:devfrom
tobwen wants to merge 1 commit intoanomalyco:devfrom
Conversation
Contributor
|
The following comment was made by an LLM, it may be inaccurate: Based on my search, I found one potentially related PR that overlaps with the current PR: Related PR:
This PR (16630) explicitly mentions in its description that it "overlaps with #11572" as part of covering the same bug class. The current PR appears to be a more comprehensive fix that also addresses issues #16154, #11571, and #14783 in a unified way, likely superseding or extending the earlier approach in PR #11572. |
tobwen
force-pushed
the
fix/target-model-capabilities
branch
3 times, most recently
from
063aabd to
64e8056
Compare
When a session switches to a different target model, it now sanitizes history and request options against that target model's capabilities. - Strips incompatible reasoning/tool history for targets that do not support those features - Preserves field-based reasoning metadata for interleaved targets - Filters unsupported reasoning/thinking options - Gates tool resolution plus toolChoice on target model toolcall Added regression tests for GPT reasoning history -> non-reasoning OpenAI-compatible Mistral.
tobwen
force-pushed
the
fix/target-model-capabilities
branch
from
64e8056 to
d25510b
Compare
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Issue for this PR
Fixes #16154
Also covers the same bug class as closed issue #11571 and overlaps with #11572 and #14783.
Type of change
What does this PR do?
When a session switches to a different target model, it now sanitizes history and request options against that target model's capabilities.
Added regression tests for GPT reasoning history -> non-reasoning OpenAI-compatible Mistral.
How did you verify your code works?
Ran from
packages/opencode:bun test test/provider/transform.test.ts test/session/llm.test.tsbunx tsc --noEmitScreenshots / recordings
Not a UI change.
Checklist