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
MRG: Release candidate 2.0.2 #377
Conversation
Omar pointed out the integer overflow in calculating the image size of an image of size > 4GB, on a 32-bit platform.
Windows .bat file shadow of Python script fails for paths with spaces. Thanks to P. Renner for the fix - see matthew-brett/myscripter#2
Always returned `False`.
Numpy 1.10 cannot broadcast a strided array that is set as writeable
@@ -24,6 +24,23 @@ and Stephan Gerhard (SG). | |||
|
|||
References like "pr/298" refer to github pull request numbers. | |||
|
|||
* Upcoming | |||
|
|||
* Trackvis reader will now allow final streamline to have fewer points that |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Trackvis changes not included in fact.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Would you rather pull in the changes or drop the changelog entry?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Just drop the changelog entry I guess, it is a minor API change.
Make TripWireError subclass AttributeError. This prevents the Python 3.5 doctest machinery triggering a TripWireError when inspecting the module containing a TripWire instance during doctest initialization.
The doctest string markup was for allowing doctests running on Python 3 and Python < 2.6. We haven't used it for a few years. This commit removes that code in the hope that no-one else is using it.
b40d11c
to
b27c954
Compare
Removed the changelog entries. @yarikoptic @ignatenkobrain Any extra commits in Debian/Fedora packages that you'd like to see in this release? |
On Fri, 06 Nov 2015, Chris Markiewicz wrote:
I don't think I have any patches BUT I would appreciate if you give me Yaroslav O. Halchenko |
Sounds good. I don't think we're in a rush, so just ping this thread when you're ready to go. |
I don't see new problems, but would be great if you will fix import for pydicom. Because we need to do import pydicom instead of import dicom. |
That means something like #351. I may have a few minutes to give that a try, but we'll need to test. @matthew-brett Is it easy to make a new Travis test that would test against pydicom 1.0? |
FWIW -- everything looked splendid across debians for me with 2.0.1-148-gdd4e725 |
@yarikoptic Not sure quite how to interpret the string |
On Fri, 06 Nov 2015, Chris Markiewicz wrote:
that was the output of git describe on master at that point, e.g. as
|
Oh, I see. Just FYI this PR is just cherry-picked bug fixes on top of the maint/2.0.x branch, not everything that's gone into master. If people would rather just package master into a new release, that's fine, too. These were almost clean cherry-picks, so I've sunk very little work into this branch. |
@effigies can you also backport pydicom's patch to this PR? It's required oneline massaging only. |
Chris - any interest in being the release manager for this one? Instructions here : https://github.com/nipy/nibabel/blob/master/doc/source/devel/make_release.rst - but I am sure they could be improved. |
TST: Use consistent import style in tests Conflicts: nibabel/nicom/dicomwrappers.py Backport retains BinOpener
Worked like a champ! |
Cool. Thanks for the validation. |
Otherwise it won't be copied across to the sdist.
56e3e9b
to
587f4a1
Compare
NB: This does not resolve the underlying issue (see conda/conda#1824), but incidentally prevents nibabel from triggering the bug during testing. -CJM
587f4a1
to
7926c98
Compare
Squashed for 2.0.2 release. See nipygh-383 for full history. BF: Use buffered gzip read in Py3.5.0, specifically TST: Add high-memory usage test for large nifti1 files Conflicts: nibabel/testing/__init__.py
7926c98
to
487a432
Compare
Okay. I think code-wise, this is ready to go. @matthew-brett could you walk me through the build-bots? I'm not quite sure how to go about this. (Also And should I Edit: To clarify, I've done everything on the checklist up to |
Chris - I updated the instructions here : #386 - does that doc make sense to you? |
For merging, my preference is always to do non-ff merge unless there's a good reason - to keep the set of commits as an identifiable set of work. |
Those instructions look good. I've been able to force builds (from the buildbot machine, not via the web interface), but haven't figured out pushing the |
Did you manage to get the web interface working in the end? It's less useful as you have to push something to the main repo before you can build it - for example you can build a tag once it's pushed, but you usually want to try building stuff before it's pushed. There's no easy way to push an environment variable, but in any case, the smaller Windows machine probably can't run the big test - the larger 64-bit machine probably can. I'll try enabling that by default on the larger Windows machine. |
See for example env vars defined here : http://nipy.bic.berkeley.edu/builders/nibabel-bdist64-27/builds/43/steps/shell/logs/stdio |
I had to do this by logging into the machine remotely and setting the enviroment variable in the users profile. Can give you access to this machine if you need it. |
Oops - even for the 6GB windows machine it was too much: http://nipy.bic.berkeley.edu/builders/nibabel-bdist64-27/builds/43/steps/shell_6/logs/stdio
|
Yeah. I can run the test on a 24GB machine but not an 8GB machine. Maybe if we just have one of the build slaves set to always run with Also, I set up a pypi user |
OK - done for pypi. I enabled slow,bigmem test for the big Mac buildslave: (I had to log on and update the relevant file on that machine, then restart the buildbot slave). I updated the release instructions adding the suggestion to run on this machine first: https://github.com/nipy/nibabel/pull/386/files#diff-d13ddc96596291d914cd84c6165cc10bR130 |
Great. Build successful. I think I should be good to go. I'll let you know if I run into any more issues. |
487a432
to
1eca265
Compare
@matthew-brett Release pushed. Hopefully I didn't somehow miss anything too egregious. Will make notes on #386. @ignatenkobrain 2.0.2 is released. Source is also on pypi |
This is a cherry-picked branch off of maint/2.0.x of the bug-fix pull-requests since 2.0.1. The only PR that needed a little massaging was #363.
There are a number of little DOC and TST fixes that could probably also fit in, if people think it makes sense.
Closes #373