Hi there!
I'm trying to use the chatgpt pr review pipe and it fails with "too many tokens in your message (over 128000)" whenever I edit a large file. The file is an sql dump of test data and generally contains nothing to review. The changes in PR are +350/-250 across 20 files, and 90% of the line changes are in that file.
Is there a way to add an ignore list to avoid situations like this?
Also, would be really nice to be able not to fail pipeline if this pipe fails. It has to be in the "PR" pipe, so when it fails there is no way to merge and it blocks the workflow. Very disruptive for no reason.
A