mirror of
https://github.com/openai/codex.git
synced 2026-10-31 00:12:59 +00:00
Fixes #10869 - Gate TUI rate-limit polling on ChatGPT-auth providers only. - `prefetch_rate_limits()` now checks `should_prefetch_rate_limits()`. - New gate requires: - `config.model_provider.requires_openai_auth` - cached auth is ChatGPT (`CodexAuth::is_chatgpt_auth`) - Prevents `/wham/usage` polling in API/custom-endpoint profiles.