Compressing Chat don t work properly well after so many Updates from OpenCode #8648

Open
opened 2026-02-16 18:10:29 -05:00 by yindo · 3 comments
Owner

Originally created by @emco1234 on GitHub (Feb 5, 2026).

Originally assigned to: @rekram1-node on GitHub.

Description

I am using Oh my OpenCode Extension with OpenCode and for somehow the Context is not more compressed automatically, sometimes he compress but sometimes he didn t if he tries manually to patch large codebase files he stuck with that error in the Screenshot that i am uploaded. Meaning he stack in the compress process and coudn t compress the window therefore it loads and loads like that and nothing happens anymore.
Before from 2 months ago OpenCode worked there much much better, OpenCode takes the whole context window and creates a new one if the old one getting to be full, now OpenCode tries that some behavior in the same window to compress the messages but it not working so well i run as exmaple with OpenAI GPT-5.2 often in this kinda issues.

What you recommend me ?
How i can fix that ?

Plugins

Oh my OpenCode

OpenCode version

Newest: 1.1.53

Steps to reproduce

Check the Screenshot, just tell him to edit large html css pages manually 20 pages, and he tries in thinking mode to do everything and stuck after that without compressing the window and continue his work.

Image

Screenshot and/or share link

No response

Operating System

No response

Terminal

No response

Originally created by @emco1234 on GitHub (Feb 5, 2026). Originally assigned to: @rekram1-node on GitHub. ### Description I am using Oh my OpenCode Extension with OpenCode and for somehow the Context is not more compressed automatically, sometimes he compress but sometimes he didn t if he tries manually to patch large codebase files he stuck with that error in the Screenshot that i am uploaded. Meaning he stack in the compress process and coudn t compress the window therefore it loads and loads like that and nothing happens anymore. Before from 2 months ago OpenCode worked there much much better, OpenCode takes the whole context window and creates a new one if the old one getting to be full, now OpenCode tries that some behavior in the same window to compress the messages but it not working so well i run as exmaple with OpenAI GPT-5.2 often in this kinda issues. What you recommend me ? How i can fix that ? ### Plugins Oh my OpenCode ### OpenCode version Newest: 1.1.53 ### Steps to reproduce Check the Screenshot, just tell him to edit large html css pages manually 20 pages, and he tries in thinking mode to do everything and stuck after that without compressing the window and continue his work. <img width="952" height="470" alt="Image" src="https://github.com/user-attachments/assets/6bda112f-4e3e-4a84-b0aa-fa669e55b9de" /> ### Screenshot and/or share link _No response_ ### Operating System _No response_ ### Terminal _No response_
yindo added the bugperf labels 2026-02-16 18:10:29 -05:00
Author
Owner

@github-actions[bot] commented on GitHub (Feb 5, 2026):

This issue might be a duplicate of existing issues. Please check:

  • #11450: ContextWindowExceededError during compaction (litellm & claude-opus-4.5) - Similar problem where compaction fails when context is too large
  • #12191: Running Compaction halluncated into looking at dog - Compaction process producing unexpected behavior in large sessions
  • #10634: Compaction overflow check doesn't account for large tool outputs until the next step, causing context overflow - Describes overflow detection issues that may cause freezing
  • #11086: GPT 5.2 Codex - Compaction happens way too early and context limit is wrong - Related compression timing issues
  • #8089: Auto-compaction enabled by default, but context_length_exceeded errors still occur in agent workflows - Auto-compaction not triggering properly

Feel free to ignore if none of these address your specific case.

@github-actions[bot] commented on GitHub (Feb 5, 2026): This issue might be a duplicate of existing issues. Please check: - #11450: ContextWindowExceededError during compaction (litellm & claude-opus-4.5) - Similar problem where compaction fails when context is too large - #12191: Running Compaction halluncated into looking at dog - Compaction process producing unexpected behavior in large sessions - #10634: Compaction overflow check doesn't account for large tool outputs until the next step, causing context overflow - Describes overflow detection issues that may cause freezing - #11086: GPT 5.2 Codex - Compaction happens way too early and context limit is wrong - Related compression timing issues - #8089: Auto-compaction enabled by default, but context_length_exceeded errors still occur in agent workflows - Auto-compaction not triggering properly Feel free to ignore if none of these address your specific case.
Author
Owner

@emco1234 commented on GitHub (Feb 8, 2026):

@CalvinTsangv1 and @rekram1-node have you any fixes for that issue ? :(

@emco1234 commented on GitHub (Feb 8, 2026): @CalvinTsangv1 and @rekram1-node have you any fixes for that issue ? :(
Author
Owner

@CalvinTsangv1 commented on GitHub (Feb 11, 2026):

@CalvinTsangv1 and @rekram1-node have you any fixes for that issue ? :(

opencode.jsonc setting change / add to -> "compaction": {"auto": false }
change configuration and see if works for you.
If it still doesn’t work, please send me the full error log

@CalvinTsangv1 commented on GitHub (Feb 11, 2026): > [@CalvinTsangv1](https://github.com/CalvinTsangv1) and [@rekram1-node](https://github.com/rekram1-node) have you any fixes for that issue ? :( opencode.jsonc setting change / add to -> "compaction": {"auto": false } change configuration and see if works for you. If it still doesn’t work, please send me the full error log
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: anomalyco/opencode#8648