Repository navigation
Print a run of tool calls as one block - #27
Merged
Merged
Conversation
Each call of a coding turn took four lines: a blank, its header, an "ok" and another blank. A turn of ten calls scrolled forty lines to say that ten things went as expected. A run is now one line per call, with nothing between them. A result keeps a line only when it is not a plain success: refused, failed, or a non-zero exit status. A blank line closes the run before the answer.
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
A coding turn printed four lines per tool call: a blank, the header,
⎿ ok, a blank. It now prints one line per call, nothing between them, and a result line only for what is not a plain success (refused, failed, non-zero exit). A blank line sets the answer apart.Tried in a real session in the container against the inference host. Tests added. The suite passes except the two launcher tests that #26 fixes (they read the user's own configuration file).