Skip to content

Add a test expectations file for to-upstream tests - #3846

Merged
domenic merged 2 commits into
mainfrom
to-upstream-expectations.yml
Mar 23, 2025
Merged

Add a test expectations file for to-upstream tests#3846
domenic merged 2 commits into
mainfrom
to-upstream-expectations.yml

Conversation

@domenic

@domenic domenic commented Mar 23, 2025

Copy link
Copy Markdown
Member

No description provided.

domenic added 2 commits March 23, 2025 16:45
No functional changes.

Currently they only process the to-run.yml file, but we plan to introduce an expectations file for the to-upstream tests as part of #3835.
@domenic

domenic commented Mar 23, 2025

Copy link
Copy Markdown
Member Author

Implementation report:

This task consisted of a lot of factoring out existing code to make it more general, so it could be reused from two places. I suspect this task could have benefited from AI a decent amount, although it's also possible I would have pushed it beyond its limits and gotten frustrated.

There was a lot of refactoring that was on the border between mechanical and interesting, and I made some nontrivial decisions on how to simplify and layer the components after the refactoring. Just pointing Cursor at it and saying "please factor this out so that it can be called from two places" might have worked, but might have produced something not-quite-right.

I definitely noticed a drag from the lack of autocomplete and Cursor-tab, as even if I do the major refactorings myself, the micro-refactorings of updating moved code for its new environment would have benefited a lot of from those AI tools.

@domenic
domenic merged commit a2b7c22 into main Mar 23, 2025
@domenic
domenic deleted the to-upstream-expectations.yml branch March 23, 2025 09:22
@domenic domenic added the metr uplift A candidate for the METR Uplift experiment label Apr 19, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

metr uplift A candidate for the METR Uplift experiment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant