Skip to content

Conversation

effigies
Copy link
Member

Dug through nisext finally, and I agree with #717 (comment) that we can probably drop most of that from our setup.py. While I'm at it, I'm moving us to a setup.cfg-based setup.

No rush on this one. Would be nice to move by 2.5 or 3.0.

Closes #717.

@effigies effigies added this to the 2.5.0 milestone May 25, 2019
@effigies
Copy link
Member Author

If we are inclined to move to versioneer, then we can drop the nisext usage altogether.

@effigies effigies force-pushed the maint/setuptools branch from 666b1b3 to 8471c35 Compare May 25, 2019 21:38
@effigies effigies force-pushed the maint/setuptools branch 2 times, most recently from 4ce67a3 to e9110d7 Compare June 24, 2019 12:55
@codecov-io
Copy link

codecov-io commented Jun 24, 2019

Codecov Report

Merging #764 into master will decrease coverage by 0.01%.
The diff coverage is n/a.

Impacted file tree graph

@@            Coverage Diff             @@
##           master     #764      +/-   ##
==========================================
- Coverage   89.99%   89.97%   -0.02%     
==========================================
  Files          94       94              
  Lines       12015    11992      -23     
  Branches     2128     2128              
==========================================
- Hits        10813    10790      -23     
+ Misses        859      858       -1     
- Partials      343      344       +1
Impacted Files Coverage Δ
nibabel/info.py 100% <ø> (ø) ⬆️
nibabel/cmdline/parrec2nii.py 32.73% <0%> (ø) ⬆️

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 935079b...2345bb4. Read the comment docs.

@effigies
Copy link
Member Author

This is all set. This sets a minimum setuptools requirement of 30.3.0, due to moving most of the metadata to setup.cfg. Any objections?

I've pushed off versioneer mostly because of the upstream/2.5.0.dev tag. I'm a little inclined to move to that after the 2.5.0 tag, which will get us off of nisext entirely and reduce the fiddliness of releasing. We can discuss that later, as well.

@effigies effigies merged commit 19ec7f3 into nipy:master Jun 27, 2019
@effigies effigies deleted the maint/setuptools branch June 27, 2019 12:47
install_requires =
numpy >=1.8
six >=1.3
bz2file ; python_version < "3.0"
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Do you by any chance remember from what setuptools version this extended setup.cfg is supported? I just fear possible troubles on older systems

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

30.3.0

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

In general, will it be a problem for older systems to upgrade setuptools? That version had a python_requires='>=2.6,!=3.0.*,!=3.1.*,!=3.2.*' constraint, which is a superset of our own support level.

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@yarikoptic Sorry to bug you, but I just want to check whether that's sufficiently old, or you still think this could produce problems... I can revert to a pre-30.3-style setuptools config, if this is a serious concern.

@yarikoptic
Copy link
Member

Current debian stable stretch has 33. Oldstable Jessie only elderly 5 and we backported nibabel for it before http://neuro.debian.net/pkgs/python-nibabel.html but I think it is ok to stop carrying about it looking forward

@effigies
Copy link
Member Author

Okay. If somebody complains after 2.5 comes out, we can release a bugfix. Thanks for checking into this.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

scripts have wrong interpreter after "conda install nibabel" in Windows
3 participants