Clear inactive cache files #121
Conversation
isidentical
force-pushed
the
inactive-file-removal
branch
5 times, most recently
from
71e903a
to
a0126c0
Compare
|
I might need a few days to merge this (want to review it well and I'm unfortunately not a student anymore :)). |
No problem at all, in the meanwhile, I need to make this patch working on 3.5< with my force-pushes to use this CI. |
isidentical
force-pushed
the
inactive-file-removal
branch
3 times, most recently
from
676000a
to
3dcf75c
Compare
Cache files that wasn't accessed in last 30 days will be automatically garbage collected. This collection happens when the `save_module` is called via a lock system that would make it happen only one time per day.
isidentical
force-pushed
the
inactive-file-removal
branch
from
3dcf75c
to
21e54c3
Compare
|
All green |
|
Great work altogether. I just think we should be a bit more lenient with errors. |
|
Thanks for the review @davidhalter, I'll try to look these next week |
isidentical
force-pushed
the
inactive-file-removal
branch
2 times, most recently
from
6c917cb
to
2356d56
Compare
parso/cache.py
Outdated
| cache_path=None, | ||
| inactivity_threshold=_CACHED_FILE_MAXIMUM_SURVIVAL, | ||
| ): | ||
| cache_path = _get_default_cache_path() |
There was a problem hiding this comment.
There should probably be an if cache_path is None before this.
|
And as a minor nitpick: I actually prefer if people don't force push in pull requests. It's very hard to understand what has been changed. I can still squash if I really need to when I merge. |
|
I try to keep the commit log clean, like for the last force pushes I was using trying to fix a CI failure on the same commit so I didn't want to create new commits just for that. But I respect your preference, so try to avoid them. EDIT: ups, sorry for the last push, muscle memory :/ Try to improve it... |
isidentical
force-pushed
the
inactive-file-removal
branch
from
2356d56
to
dbbaf09
Compare
|
As I said, it's a nitpick, feel free to do it however you want to. I feel it's obviously easier to spot the differences in PRs, but I can still work around that. I also personally think that commit logs don't matter nearly as much as people think they do. If I want to write a longer text in a commit log I usually abort and make the exact comment in code. But here opinions tend to vary a lot and I'm probably not even part of the majority. In the end all I want is readable concise error-free code. |
|
(and as a small addition, I like to actually have the whole history of changes. For example I think your first iteration was really interesting, because it was simply another approach, but as I said, a lot of people don't feel that way) Feel free to merge when you think you're done. |
Cache files that weren't accessed in the last 30 days will be automatically
garbage collected. This collection happens when the
save_moduleis calledvia a lock system that would make it happen only one time per day. Resolves #120