REGR: Revert #793 (sorting in sparse to arrays)#824
Conversation
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## main #824 +/- ##
=====================================
Coverage 85.4% 85.4%
=====================================
Files 151 151
Lines 16015 16025 +10
=====================================
+ Hits 13679 13689 +10
Misses 2336 2336
🚀 New features to boost your workflow:
|
|
The issue in CI is linked to numpy 1.26, which is already out of SPEC 0 support window. Shall we just bump the pin? |
I there a way to have a soft pin? Like |
|
We're either following SPEC 0 or we are not.
As long as we don't make any change that would be incompatible, you can pip install with Given we adopted SPEC 0 specifically to avoid discussions like this, I find it a bit unfortunate that we have it. |
|
This is not backwards compatible so we don't have to deal with that here. |
|
I think tracking spec 0 is important. Trying to support outside of that window increases the maintenance cost. |
sjsrey
left a comment
There was a problem hiding this comment.
This is bringing back memories from dealing with sorting in Weights. I feel your pain :-).
Reverting the rest of #793 as #822 apparently was not enough. I managed to cause two regressions by a single PR 🎉.
This was caught by momepy CI - https://github.com/pysal/momepy/actions/runs/20742499625/job/59552088014