Repository navigation
test(eval): compare the floor as the plugin does, to a whole percent - #4
Merged
Merged
Conversation
Claude Code reports context as a whole percentage, so at a 5% floor a compaction at 46,335 tokens of a 1M window is past it: the plugin saw 5%. The other floor checks already allow a point either side; this one compared exact tokens and failed a run that did what it should. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
make eval-mechanismfailed "nothing is compacted below the floor" on a run that behaved correctly: with the floor at 5%, the first turn ended at 46,335 tokens of a 1M window, Claude Code reported that as 5%, and taskcut compacted. The check compared exact tokens (50,000) while the plugin compares Claude Code's whole percentage.The other floor checks in the same function already allow a point of rounding either side; this one now does too.
The same run's two other failures (recall, the interrupted work) are not addressed here: from the second compaction on, the compaction model refused to summarise, and each later compaction followed its own earlier refusal. That is Claude Code's compaction, not the check.
🤖 Generated with Claude Code