Compact bug

The report names no files, tests, or Command Code version. Start by reproducing the context-length failure in the Windows and VSCode setup after a long session, then gather the missing version and session details. Done means the compaction or retry flow no longer exceeds the model's context limit and the behavior is covered by a regression test.

Summary

⚠ Error: 400 This model's maximum context length is 1048576 tokens. However, you requested 1049054 tokens (985054 in the messages, 64000 in the completion). Please reduce the
length of the messages or completi...

Type "continue" to try again. If the issue persists, contact support: https://commandcode.ai/discord
Trace ID: 30a6bb1d0b4d8bcf90b28f564ad5e5a2

◼ Conversation is already compact.

◼ Opening the /feedback GitHub issue in your browser…

✻ Worked for 3s

Expected Behavior

I do now know.

Actual Behavior

⚠ Error: 400 This model's maximum context length is 1048576 tokens. However, you requested 1049054 tokens (985054 in the messages, 64000 in the completion). Please reduce the
length of the messages or completi...

Type "continue" to try again. If the issue persists, contact support: https://commandcode.ai/discord
Trace ID: 30a6bb1d0b4d8bcf90b28f564ad5e5a2

◼ Conversation is already compact.

◼ Opening the /feedback GitHub issue in your browser…

✻ Worked for 3s

Steps to reproduce the issue

Working on coding stuff. Gave it docs to master and lot of chnages over 2 days time. It hit this error.

Command Code Version

Unknown

Operating System

Windows

Terminal/IDE

VSCode

Shell

cmd.exe

Session file (optional)

No response

Fix prompt (optional)

No response

Additional context

No response