tinyarray merge requestshttps://gitlab.kwant-project.org/kwant/tinyarray/-/merge_requests2015-11-07T10:53:39Zhttps://gitlab.kwant-project.org/kwant/tinyarray/-/merge_requests/1Python32015-11-07T10:53:39ZJoseph WestonPython3Add Python 3 support to tinyarray.
Tinyarray will work with *both* Python 2 and Python 3 after this change.Add Python 3 support to tinyarray.
Tinyarray will work with *both* Python 2 and Python 3 after this change.Christoph GrothChristoph Grothhttps://gitlab.kwant-project.org/kwant/tinyarray/-/merge_requests/3implement comparison between arrays with the same shape2016-05-26T10:59:21ZJoseph Westonimplement comparison between arrays with the same shapeIf the arrays do not have the same shape then we return NotImplemented.
This will fix issue #2 for the case where the arrays are the same shape, and it should
raise a `TypeError` in the other cases.
This currently fails for Python...If the arrays do not have the same shape then we return NotImplemented.
This will fix issue #2 for the case where the arrays are the same shape, and it should
raise a `TypeError` in the other cases.
This currently fails for Python 2 because `TypeError` is not raised, even when the comparison methods
return `NotImplemented`.Christoph GrothChristoph Grothhttps://gitlab.kwant-project.org/kwant/tinyarray/-/merge_requests/2Implement `__sizeof__`2016-05-26T11:18:54ZJoseph WestonImplement `__sizeof__`Implement the `__sizeof__` member function to get the size in bytes of the full Python object.
This fixes #3.
I have also added the Gitlab-CI config to the Git repository so that we can have automated testing for Tinyarray as for Kwa...Implement the `__sizeof__` member function to get the size in bytes of the full Python object.
This fixes #3.
I have also added the Gitlab-CI config to the Git repository so that we can have automated testing for Tinyarray as for Kwant.
I can pull this out into a separate merge request if needed.
Christoph GrothChristoph Grothhttps://gitlab.kwant-project.org/kwant/tinyarray/-/merge_requests/7CI: add mirroring between kwant gitlab, gitlab.com and github.com2016-11-04T21:30:09ZJoseph WestonCI: add mirroring between kwant gitlab, gitlab.com and github.comrelated to qt/admin-notes#23
Use CI for mirroring instead of server-side hooks, as it is more maintainablerelated to qt/admin-notes#23
Use CI for mirroring instead of server-side hooks, as it is more maintainablehttps://gitlab.kwant-project.org/kwant/tinyarray/-/merge_requests/8fix memory leak when (un)pickling tinyarrays2016-12-20T10:38:03ZJoseph Westonfix memory leak when (un)pickling tinyarraysfixes #10fixes #10Christoph GrothChristoph Grothhttps://gitlab.kwant-project.org/kwant/tinyarray/-/merge_requests/9adds automatic building of conda package2017-01-30T13:51:40ZBas Nijholtadds automatic building of conda packageSame idea as: https://gitlab.kwant-project.org/kwant/kwant/merge_requests/80/
This will need an env variable `$ANACONDA_TOKEN` that I can provide for the https://anaconda.org/kwant channel.
It will automatically build a conda packa...Same idea as: https://gitlab.kwant-project.org/kwant/kwant/merge_requests/80/
This will need an env variable `$ANACONDA_TOKEN` that I can provide for the https://anaconda.org/kwant channel.
It will automatically build a conda package for each commit in master.https://gitlab.kwant-project.org/kwant/tinyarray/-/merge_requests/10Add missing space that C++11 needs for macro expansion.2017-02-13T16:40:55ZJörg BehrmannAdd missing space that C++11 needs for macro expansion.This silences the warning about "invalid suffix on literal".
The remaining warnings about strict prototypes are actually a bug in distutils that has been carried around since at least 2006 and does not seem to get a fix soon.This silences the warning about "invalid suffix on literal".
The remaining warnings about strict prototypes are actually a bug in distutils that has been carried around since at least 2006 and does not seem to get a fix soon.Christoph GrothChristoph Grothhttps://gitlab.kwant-project.org/kwant/tinyarray/-/merge_requests/11Circumvent an error in test_dot for Numpy 1.12.0.2017-02-13T16:48:04ZJörg BehrmannCircumvent an error in test_dot for Numpy 1.12.0.Numpy 1.12.0 updated iscomplexobj which always shows False for
tinyarrays, which makes assert_almost_equal fail for complex
array-valued arguments, as they will not be split up into real and
imaginary part. We circumvent this bug by test...Numpy 1.12.0 updated iscomplexobj which always shows False for
tinyarrays, which makes assert_almost_equal fail for complex
array-valued arguments, as they will not be split up into real and
imaginary part. We circumvent this bug by testing every entry
separately in this case. This bug will be fixed for Numpy 1.12.1.
Fixes #11Christoph GrothChristoph Grothhttps://gitlab.kwant-project.org/kwant/tinyarray/-/merge_requests/12correct hashing and object size calculations2018-09-25T20:14:53ZJoseph Westoncorrect hashing and object size calculationsThese fixes were only apparent on Windows with the Visual Studio compiler, where `long` is 32 bits, even on 64-bit platforms.These fixes were only apparent on Windows with the Visual Studio compiler, where `long` is 32 bits, even on 64-bit platforms.Christoph GrothChristoph Grothhttps://gitlab.kwant-project.org/kwant/tinyarray/-/merge_requests/13turn off tests that rely on hash implementation in python 3.82019-12-30T09:09:30ZAnton Akhmerovturn off tests that rely on hash implementation in python 3.8While properly addressing #16 would require implementing an alternative hash implementation, I would like to re-enable conda builds and indicate that this failure is about optional functionality.While properly addressing #16 would require implementing an alternative hash implementation, I would like to re-enable conda builds and indicate that this failure is about optional functionality.Christoph GrothChristoph Grothhttps://gitlab.kwant-project.org/kwant/tinyarray/-/merge_requests/14Python 3.10 Support2021-12-17T15:15:04ZStefano RiveraPython 3.10 SupportIn Python 3.10, NaNs no longer hash equal, instead they are hashed by
their object's address. (See: https://bugs.python.org/issue43475)
Replace that behaviour by passing the address of the object to the
hash() function.
This is all rat...In Python 3.10, NaNs no longer hash equal, instead they are hashed by
their object's address. (See: https://bugs.python.org/issue43475)
Replace that behaviour by passing the address of the object to the
hash() function.
This is all rather messy, I'm sure someone more familiar with C++
templates can do something a lot tidier...
Fixes: #20https://gitlab.kwant-project.org/kwant/tinyarray/-/merge_requests/15use Py_ssize_t everywhere2021-12-22T14:55:09ZAnton Akhmerovuse Py_ssize_t everywherePrompted by the conda build failure; `ssize_t` is not guaranteed to be defined on non-POSIX platformsPrompted by the conda build failure; `ssize_t` is not guaranteed to be defined on non-POSIX platforms