[codex] Split leading reasoning blocks

[HisenWeb]

Summary

• Added shared text helpers for leading reasoning blocks in Chat content.
• Split non-streaming leading reasoning blocks into Responses reasoning plus visible message output.
• Added a streaming leading-block detector so partial opening/closing tags do not leak reasoning into output_text.
• Added regression tests for non-stream and streaming conversion behavior.

Why

Some OpenAI-compatible Chat upstreams may return reasoning embedded at the start of content rather than in explicit reasoning_content / reasoning fields. The adapter should preserve that as Responses reasoning output and avoid emitting hidden reasoning as ordinary assistant text.

Scope

This keeps the adapter narrow:

• no provider registry changes
• no cc-switch module copy
• no codex_chat_common.rs extraction yet
• no README changes, because the current main already lists codex_chat_common.rs as a shared helper reference

Validation

Not run locally in this environment. Recommended:

cargo fmt --check
cargo test --lib
cargo test --test reasoning_block_regression
cargo test

[chatgpt-codex-connector]

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.
To continue using code reviews, you can upgrade your account or add credits to your account and enable them for code reviews in your settings.