Initial diff format incorrect (command succeeds after model corrects itself)

Resolved 💬 3 comments Opened Aug 9, 2025 by Frulfump Closed Oct 27, 2025

💡 Likely answer: A maintainer (seratch, member) responded on this thread — see the highlighted reply below.

What version of Codex is running?

0.20.0

Which model were you using?

gpt-5

What platform is your computer?

Linux 6.6.87.2-microsoft-standard-WSL2 x86_64 x86_64

What steps can reproduce the bug?

codex /init

What is the expected behavior?

It can successfully get the diff format correct the first time.

What do you see instead?

thinking Correcting file format I see that for adding a file in the patch format, every line needs to be prefixe d with a '+'. It looks like I overlooked adding those '+' signs. I'll retry the process, ensuring I include the correct prefixes this time. Plus, I need to make sure to include the start and end tokens: " Begin Patch" and " End Patch. " Okay, I'm getting ready to re-run it with these corrections!

Additional information

N/A

View original on GitHub ↗

3 Comments

seratch member · 1 year ago

What's the final result of this scenario? Did /init command fail?

Frulfump · 1 year ago

No it could correct itself so the /init command did succeed, this is just about the waste of tokens and time.

etraut-openai contributor · 10 months ago

I think this issue is no longer relevant, so I'm going to close it out. If you're still seeing this problem with recent versions of codex, please let us know, and I'll reopen.