Conversation
compress: add message mode
Treat the native compaction summary as stale token data so context nudges wait for the first fresh assistant response.
Clear DCP message alias state when native compaction is detected so refs rebuild from the compacted conversation instead of continuing toward m9999.
Keep message references visible on every text chunk and completed tool output so multi-part messages stay fully addressable.
Keep completed compress calls visible as high-priority message targets so message-mode guidance preserves them consistently.
Match message token estimates to the tool arguments OpenCode stores so large built-in, custom, and plugin tool inputs do not slip past sizing.
Move the user-role guard into the helper so callers can use it consistently across message flows. Add a regression test to keep non-user messages from being treated as ignored.
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Closes #450
Closes #452
compress.modeconfig setting, defaults to"range"(previous behavior) and new experimental"message"option available."message"mode allows DCP to more surgically compress individual messages in a batch rather than a full range."message"mode allows DCP to determine compression priority for each message based on how many tokens the message is (low/med/high for 0-500/500-5000/5000+ tokens), not configurable yet but maybe in the future?"message"mode to ensure they don't get left behind."range"mode now supports the model using a list of ranges instead of just 1 range if it wants.compress.flatSchemadeprecated, not needed for lower tier model performance anymore due to new tool schemas (I think...)compress.summaryBufferconfig setting, allows active summary tokens to extend the effectivecompress.maxContextLimitto prevent excessively frequent compressions in very long sessions.