-
Notifications
You must be signed in to change notification settings - Fork 1.2k
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
Reimport checkout #10388
Reimport checkout #10388
Conversation
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## main #10388 +/- ##
==========================================
- Coverage 90.72% 90.46% -0.27%
==========================================
Files 501 501
Lines 38865 38885 +20
Branches 5618 5619 +1
==========================================
- Hits 35262 35176 -86
- Misses 2961 3038 +77
- Partials 642 671 +29 ☔ View full report in Codecov by Sentry. |
Ping @skshetry |
I'll need to investigate. But we set Line 154 in d510ad6
I am a bit hesitant of this PR as it goes against #9246, and adds another (complex) codepath. |
Yes, I looked into it, and the local cache dir is set, but it's not enough because it's doing download/copy and not checkout.
Lines 47 to 76 in 7df8b0c
It calls I don't think there is a much simpler way to do it than this PR, and I tried to make it as narrow as possible. It uses #9246 to rely on the local cache, building the object and passing it to dvc-data's checkout. We don't have |
|
Yes, but it is doing a copy from that fs to localfs. It is not linking, and even if I try to enable linking, it fails because it is working across filesystems. |
Hello, is there any update with this merge request? It would help us a ton if it gets merged. |
@skshetry Is there a good reason to keep this blocked? I'm happy to revert if we find time to resolve in a cleaner way later. |
@skshetry ping |
@dberenbaum, do you have a repro for this? It seems we do use |
Fixes #10255. This is a partial reversion of #9246. I'm not sure this is the cleanest or best way to do it, but it solves the problem.
It builds the hashfile object and tries to checkout from that, falling back to download if it fails.
Note that the build step is still slow for 2.x imports.