When dealing with the C/C++ layer and switching back and forth between
upstream master and your branch, I would suggest running "git clean -fxd"
and then do a complete rebuild. We have recently fixed a number of memory
leaks that may or may not have merged well with your code as it involved
the CXX layer.
On Tue, Jul 22, 2014 at 5:41 AM, Stefan H. <shmuell...@gmail.com> wrote:
> Hi! Last year, I have been working on making matplotlib compatible with
> PyPy
> here:
>
> https://github.com/shmuller/matplotlib.git
>
> I succeeded to get the svg and pdf backends working, but then had to drop
> the ball. My changes involve mostly the C++ parts in src, in particular
> replacing some PyCXX (which didn't work with PyPy) with the plain Python C
> API. I also needed some minor fixes in .py files.
>
> I understand that it is desirable to remove the PyCXX layer altogether from
> matplotlib, and I have come a good deal along this route. I would also be
> happy to continue working in this direction.
>
> I have a bit of time now, and so I merged my branch with upstream/master.
> The few conflicts were easy to resolve. However, then I tried to run the
> test suites (python tests.py), both on the current master and my branch, to
> see if I broke anything.
>
> Unfortunately, the test suites randomly segfault on my computer (Macbook
> Air
> 2011), at the latest when I try to run them 10 times in a row. Not only on
> my branch, but also on upstream/master. I ran the Apple Hardware Test,
> which
> didn't report any issues, so my Mac seems to be fine. I went back to the
> state of my branch before merging, which i have been using in production
> for
> a year now without any issues, and noticed that even there the test suite
> segfaulted if I try to run it 10 times in a loop. I haven't noticed this
> before (and didn't have any issues in production use of matplotlib for a
> year). The crashes occur randomly and not always at one specific test.
>
> - Did anyone have this issue before?
>
> - Could someone pull my branch and run the test suite on a different
> computer, so that I can exclude that my hardware is at fault after all?
>
> - Could someone please guide me to get my work pulled into the main
> repository?
>
> Many thanks,
> Stefan
>
>
>
> --
> View this message in context:
> http://matplotlib.1069221.n5.nabble.com/Trying-to-prepare-a-pull-request-but-tests-py-keeps-crashing-on-master-tp43680.html
> Sent from the matplotlib - devel mailing list archive at Nabble.com.
>
>
> ------------------------------------------------------------------------------
> Want fast and easy access to all the code in your enterprise? Index and
> search up to 200,000 lines of code with a free copy of Black Duck
> Code Sight - the same software that powers the world's largest code
> search on Ohloh, the Black Duck Open Hub! Try it now.
> http://p.sf.net/sfu/bds
> _______________________________________________
> Matplotlib-devel mailing list
> Matplotlib-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/matplotlib-devel
>
------------------------------------------------------------------------------
Want fast and easy access to all the code in your enterprise? Index and
search up to 200,000 lines of code with a free copy of Black Duck
Code Sight - the same software that powers the world's largest code
search on Ohloh, the Black Duck Open Hub! Try it now.
http://p.sf.net/sfu/bds
_______________________________________________
Matplotlib-devel mailing list
Matplotlib-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/matplotlib-devel
