Show LLM errors #2019
Merged
Show LLM errors #2019
Conversation
This file contains 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
1. Tell the model not to put code changes in fences. 2. Tell it to chunk the response if too long.
kgilpin
approved these changes
Oct 1, 2024
|
🎉 This PR is included in version @appland/components-v4.38.0 🎉 The release is available on:
Your semantic-release bot 📦🚀 |
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Outdated suggestions cannot be applied.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.
Turns out errors were dropped in the UI RPC layer. It took ages to go through all the layers to figure out where the leak was, but fortunately once found the fix was trivial.
(Note this is copilot which has heavy off-topic filters.)
Error messages could use better styling, but I think this is outside of scope and this change should be merged ASAP because even this is better than silence.
(I also added another small commit adjusting the generate prompt; it helps prevent XML fencing (fixes #2012 hopefully) and tells the model to chunk response when required; this doesn't work best yet since
@generateonly looks at the last message to extract paths, so continuation requests won't have full required file contents, but I feel it's still better than stopping mid-chunk due to response size limits.)