Open paul-gauthier opened 3 days ago
I think I mentioned when I saw that on discord, but I have had aider work on a file and be fine and then later commit it back in a new location (IIR it was a test_xxx and it got moved from the root, to a duplicate file that aider added under tests/test_xxx
Not sure if this related, but I just caught the AI/Aider editing the wrong file:
I had given it failing test results, run a debug prompt on it and it only had the code file (debug_ic_refactor.py) in the chat. It seems like it went "meh, close enough, we'll edit that one":
Recommendations:
1. Modify the test to ensure the temporary configuration file is created and accessible.
2. Update the test to properly add and check for the "This should appear" log message.
3. Improve error handling in the CustomLogger class when loading configurations.
4. Review and update the thread management in the CustomLogger class to prevent shutdown issues.
</analysis>
<code_updates>
Let's implement these changes:
1. Update the test_configuration_reload method in test_debug_ic_refactor_quick.py:
debug_ic_refactor.py
<source>python
<<<<<<< SEARCH
def reload_configuration(self):
The only files in the chat were two read only conventions .md files, and the main code:
../prompts/coding/code-terse.md ../prompts/coding/dee/output/debug-terse.md debug_ic_refactor.py
Perhaps some way of confirming that the file the AI thinks it is editing compared to the one it actually edits would help here?
Aider v0.59.1 Main model: claude-3-5-sonnet-20240620 with diff edit format, prompt cache, infinite output Weak model: claude-3-haiku-20240307 Git repo: .git with 38 files Repo-map: using 1024 tokens, files refresh
Issue
https://discord.com/channels/1131200896827654144/1296554646789161132/1296554646789161132
Version and model info
No response