fix(tui): keep conversation model overrides sticky#1238
Merged
jnjpng merged 1 commit intoletta-ai:mainfrom Mar 6, 2026
Merged
Conversation
Contributor
|
@paulbettner How were you observing this bug? E.g. changing model on either non-default/default conversation would immediately snap back after the first message sent after model switch? Or was it changing back non-deterministically after some amount of turns? |
This was referenced Mar 5, 2026
Contributor
Author
|
@jnjpng it seemed to be happening after the first model response. |
paulbettner
force-pushed
the
fix/conversation-model-override-sticky
branch
from
43a0adf to
2ea886e
Compare
Keep the conversation override flag ref synchronized for async callbacks so the first chunk agent sync can't clobber a /model selection. Persist /model and reasoning tier changes when running in the virtual "default" conversation by updating the agent config, preventing snap-back. Cache conversation model_settings while an override is active so reasoning effort derivation matches the override on resume. Also align default conversation API calls with the current client types: use the agent id at conversations endpoints and avoid passing agent_id params. 👾 Generated with [Letta Code](https://letta.com) Co-Authored-By: Letta Code <noreply@letta.com>
paulbettner
force-pushed
the
fix/conversation-model-override-sticky
branch
from
2ea886e to
8c6320b
Compare
jnjpng
added a commit
that referenced
this pull request
Mar 6, 2026
The no-op function placeholders introduced in #1238 silently swallow calls if the overwrite invariant ever breaks. null! crashes loudly, which is the safer default for a synchronous forward-reference pattern. 👾 Generated with [Letta Code](https://letta.com) Co-Authored-By: Letta Code <noreply@letta.com>
1 task
jnjpng
added a commit
that referenced
this pull request
Mar 6, 2026
…nore The no-op function placeholders introduced in #1238 silently swallow calls if the overwrite invariant ever breaks. null! crashes loudly, which is the safer default for a synchronous forward-reference pattern. Also expand the biome-ignore comments on processConversation and onSubmit to flag that these are blanket suppressions hiding ~16+ omitted deps — safe for refs and stable callbacks, but could mask genuinely missing reactive deps causing stale-closure bugs. 👾 Generated with [Letta Code](https://letta.com) Co-Authored-By: Letta Code <noreply@letta.com>
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
2 participants
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
model_settingswhile an override is active and uses it to derive reasoning effort, fixing missing footer reasoning tags on resume.Test plan
bun run checkbun test src/tests/agent/model-preset-refresh.wiring.test.ts/model <some-tiered-model>then immediately send a message; ensure the UI does not flip back to the agent's base model.👾 Generated with Letta Code