You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I'm not sure I understand your problem description. Could you update it with some details? Are you suggesting that there is a problem with checking out different branches or something like that? What does asv refer to here? A link would help.
I'm not sure I understand your problem description. Could you update it with some details? Are you suggesting that there is a problem with checking out different branches or something like that?
Refresher on git worktree:
~/Desktop/pd/ is where I keep all my pandas branches
pd/pandas/ has master checked out
pd/pandas/.git is the actual repo
git worktree add ../foo foo creates a foo branch and checks it out in pd/foo/, and the pd/foo/.git file is a single-line file telling git where to find the actual repo directory.
If I were to just check out a new branch directly inside the pd/pandas/ directory, the cythonized .c files would still be there, so would not need to be re-cythonized. But when a new branch is instantiated in a worktree directory, the .pyx files need to be cythonized from scratch, which is what I'd like to avoid.
A benchmarking tool. The relevant behavior is that it will also check out existing branches in fresh directories and build from scratch, which can be a real pain point.
cython is pretty good about avoiding unnecessary compilations, but that doesn't play nicely with git worktree or asv
Two ways this could work:
An explicit cache that sits next to the "real" .git repo
If in a worktree directory, scan other directories in the worktree to see if we can re-use already-cythonized results from there.
I'm not sure how to make version 2 work with asv.
The text was updated successfully, but these errors were encountered: