Skip to content
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

--cache-ignore behavior isn't the same as invalidate before #5674

Closed
wisechengyi opened this issue Apr 8, 2018 · 8 comments
Closed

--cache-ignore behavior isn't the same as invalidate before #5674

wisechengyi opened this issue Apr 8, 2018 · 8 comments

Comments

@wisechengyi
Copy link
Contributor

wisechengyi commented Apr 8, 2018

With the removal of invalidate (2d4e617#diff-0295a620e939c7397353c955d452ae25), there is no good way to force a task to recheck for cache anymore after a successfully run because everything is validated. This is important because sometimes we need to make sure a task will hit the cache with invalidation, whereas the alternative clean-all will wipe out a lot of other info under .pants.d/ along with pants server content.

I'd like to propose to add it back. @stuhood @jsirois @benjyw, thoughts?

@stuhood
Copy link
Member

stuhood commented Apr 9, 2018

I think that for the purposes of finding caching bugs, and for pants Task implementors, clean-all is sufficient, and invalidate is too challenging for an end user to use correctly.

Also, invalidate will stop making sense as more tasks are ported to the new engine.

@jsirois
Copy link
Contributor

jsirois commented Apr 9, 2018

And if you want invalidate back, that's just alias pants-invalidate='rm -r .pants.d/build_invalidator'.

@benjyw
Copy link
Contributor

benjyw commented Apr 9, 2018

I wouldn't be opposed to rearranging directory names under ./pants.d/build_invalidator and in the cache so that it's easier to selectively delete them if you're testing invalidation/caching behavior. But I wouldn't expose invalidate to the end user again.

@stuhood
Copy link
Member

stuhood commented Apr 9, 2018

I also think that incorporating some target awareness into clean-all would be useful, as I'm sure people expect that that works.

@benjyw
Copy link
Contributor

benjyw commented Apr 9, 2018 via email

@stuhood
Copy link
Member

stuhood commented Apr 9, 2018

Yea, basically. Although I think it would be good to make more progress on #4769 in order to figure out what that would actually mean, and not implement something that we immediately need to break in order to have a sane UX there.

@benjyw
Copy link
Contributor

benjyw commented Apr 9, 2018 via email

@Eric-Arellano
Copy link
Contributor

Closing as this option was removed in Pants 2.0.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

5 participants