-
Notifications
You must be signed in to change notification settings - Fork 1.8k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
BUG (?): v0.58 seems to have gotten caught in a loop, ran up $13 in charges rapidly #1842
Comments
I don't believe the code.md file is implicated in the issue as it is unique to me and other users are reporting similar issues without using code.md. If having code.md will help I can share it but I have a feeling you can reproduce the error without it. |
Let's track this issue in #1841 so we can consolidate any clues and learnings there. |
Starting with a new venv and attempting to recreate:
Then:
Then:
Then
Then
Then
Then
Then
Then
Then
/bin/bash: - : invalid option Add the output to the chat? ^C again to exit
test_board_initialization (word_search_game.tests.test_board.TestBoard.test_board_initialization) ... ok ======================================================================
|
After running the code (it works correctly, I played the game), I ran the same test command again, and triggered the cascading output issue a second time: ``
test_board_initialization (word_search_game.tests.test_board.TestBoard.test_board_initialization) ... ok ======================================================================
|
The fix is available in the main branch. You can get it by installing the latest version from github:
If you have a chance to try it, let me know if it works better for you. |
Issue
I just upgraded to 0.58.
UPDATE: As requested in the other thread on loops, here are the switches I launch aider with (in addition to two lines not shown that contain my API keys):
--sonnet
--cache-prompts
--cache-keepalive-pings 5
--no-stream
--dark-mode
--edit-format diff
I set up a new directory, created a venv, sourced in to it and launched aider.
I provided via /read my code conventions (code.md)
I provided this prompt:
/architect Respecting the rules in code.md, design an interactive, linux console text based, word search game.
It came up with a design. That worked fine. It prompted at the end of the response to edit the file, which I agreed to.
It created the file described in the design and wrote the contents out, that all worked fine.
It sent updated code.
Now this is where things went off the rails. I hadn't tested the code yet, I'm focusing on vetting workflow with the new architect/coder model. So far the code for the game looked fine in VS Code and the prompts and responses all made sense.
My next prompt was:
So as you can see, after the test details were spit out and it asked to edit the file, some kind of loop ensued and a whopping Tokens: 4514k sent, 476 received. Cost: $13.55 message, $13.81 session. were transferred.
I happened to get a phone call right after I launched the command and walked away from my PC, assuming I would come back and it would be finished. Instead when I came back it was still running which is why I interrupted it and was shocked to see how many credits it chewed up.
Any ideas where this went wrong? Also, should there be some outer limit of tokens that can be transferred before aider halts the conversation and raises an error?
Version and model info
Aider v0.58.0
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 0 files
Repo-map: using 1024 tokens, files refresh
The text was updated successfully, but these errors were encountered: