Skip to content

Re-enable multifield copy->view change#4357

Merged
TomAugspurger merged 1 commit intodask:masterfrom
detrout:numpy-1.16-dtype-compatibility
Jan 14, 2019
Merged

Re-enable multifield copy->view change#4357
TomAugspurger merged 1 commit intodask:masterfrom
detrout:numpy-1.16-dtype-compatibility

Conversation

@detrout
Copy link
Contributor

@detrout detrout commented Jan 7, 2019

Numpy attempted to release a feature in 1.14.0, but rolled it
back because it caused too many problems. After fixing issues, they
rereleased it in 1.16.0.

This also makes a minor enhancement to determines which numpy function
should be used at import time instead of run time.

cc @hmaarrfk

@hmaarrfk
Copy link
Contributor

hmaarrfk commented Jan 7, 2019

I think this warrants a short explanation and links to the 3PRs that touched the code.
and maybe rename the functions to _16???

@detrout
Copy link
Contributor Author

detrout commented Jan 8, 2019

Those requests seem quite reasonable but it may take me a bit to get back to this, I'm going to a conference for the rest of the week and am likely to be quite busy. If I have a chance I'll try to fix the PR. (But I didn't want to leave you hanging if it took me until the weekend)

@hmaarrfk
Copy link
Contributor

hmaarrfk commented Jan 8, 2019

I think that is fine. take your time! you found the bug and the fix!

Numpy attempted to release a feature in 1.14.0,
numpy/numpy#6053 which was then reverted in
1.14.1 with numpy/numpy#10411 And then was
finally released in 1.16 by numpy/numpy#12447

This also makes a minor enhancement to determines which numpy function
should be used at import time instead of run time.

The function name was also then updated from
_make_sliced_dtype_np_ge_14 to _make_sliced_dtype_np_ge_16 as thats
the primary version the feature was released in
@detrout detrout force-pushed the numpy-1.16-dtype-compatibility branch from bdbc76b to 527d97f Compare January 10, 2019 02:14
@hmaarrfk
Copy link
Contributor

@TomAugspurger this PR looks good to me. Sorry I didn't notice the changes were made as requested.

@TomAugspurger
Copy link
Member

Thanks @detrout, sorry I missed this earlier.

@TomAugspurger TomAugspurger merged commit 3e33c87 into dask:master Jan 14, 2019
@detrout
Copy link
Contributor Author

detrout commented Jan 14, 2019

No worries. I'd felt the discussion implied the patches were good enough and I'd already shipped this fix and the one for #4352 in Debian.

Glad to have been able to help.

jorge-pessoa pushed a commit to jorge-pessoa/dask that referenced this pull request May 14, 2019
Numpy attempted to release a feature in 1.14.0,
numpy/numpy#6053 which was then reverted in
1.14.1 with numpy/numpy#10411 And then was
finally released in 1.16 by numpy/numpy#12447

This also makes a minor enhancement to determines which numpy function
should be used at import time instead of run time.

The function name was also then updated from
_make_sliced_dtype_np_ge_14 to _make_sliced_dtype_np_ge_16 as thats
the primary version the feature was released in
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Numpy 1.16.0 test failure, dtype construction seems wrong

3 participants