tags: NumPy, BIDS
# 2019-4-03 BIDS/NumPy Development Meeting
- Join via Zoom at [https://berkeley.zoom.us/j/400054438](https://berkeley.zoom.us/j/400054438), or dial-in through a [local phone number](https://zoom.us/u/adQDmEc1wI).
- [Trello workboard](https://trello.com/b/Azg4fYZH/numpy-at-bids)
- [Previous meetings](https://github.com/BIDS-numpy/docs/tree/master/status_meetings)
- [NumPy at BIDS Calendar](https://calendar.google.com/calendar?cid=YmVya2VsZXkuZWR1X2lla2dwaWdtMjMyamJobGRzZmIyYzJqODFjQGdyb3VwLmNhbGVuZGFyLmdvb2dsZS5jb20)
**Present:** Tyler, Matti, Ralf, Chuck, Hameer
## Follow-up from last meeting/discussions
## Community Suggested Topics
- Google Season of Docs
- It is important that we have *high quality* ideas: quality over volume; chances of success not great
## Work status
Kevin did some refactoring which has been merged to the PR. Since I couldn't get the docs to build cleanly I went back to the bashtage/randomgen repo, which [travis](https://travis-ci.org/bashtage/randomgen/jobs/514664628#L2784) builds with 130 warnings about footnotes. But it does not have the toctree warnings I get. Still investigating.
The deadline for new applicants has ended. We now have to decide between the candidates, who did some work resolving documentation issues
Still working on `@property` for cdef functions
- Code Modernization
- [[open PR]](https://github.com/scipy/scipy/pull/9989): replace the usage of the Qhull Cython buffer interface to NumPy arrays with typed memoryviews
- various reviewer comments about performance---it is not as simple as just replacing the buffers and everything is faster
- test coverage incomplete & not shown for Cython (or any other source files) for SciPy at the moment
- Testing / DevOps
- [[merged PR]](https://github.com/numpy/numpy/pull/13217): publish Azure CI test summaries even on failure
- now properly fixed [in newly-merged PR](https://github.com/numpy/numpy/pull/13253)
- [[open fork PR]](https://github.com/tylerjereddy/numpy/pull/6): use POWER8 OpenBLAS instead of ATLAS
- initial `OpenBLAS` [build script](https://github.com/tylerjereddy/openblas-static-gcc/blob/master/power8/build_openblas.sh) succeeds on ppc64le GCC compile farm machine, but not with static linking---I've contacted them about possibility of including `libgfortran.a` built with `-fPIC`
- currently, the ppc64le Travis CI runs with my prelim OpenBLAS POWER8 binaries error with `/usr/bin/ld: skipping incompatible /usr/lib/libopenblas.so when searching for -lopenblas`
- looks like I'm getting x86_64 arch on my fork for some reason---I think the reason might be related to [Travis `com` / app vs `org`](https://github.com/numpy/numpy/issues/12688#issuecomment-479613995)?
- still trying to get a solid reproduction environment for `Pinv` `generalized_sq_cases` -- I've been doing debug builds on private version of repo with Travis debug mode:
- I can reproduce this way from the debug / interactive mode:
- however, dissection of this stage into more manual / interactive components has so far completely failed to generate the test failure, even with `pytest-repeat` via: `python3.6 ../numpy/runtests.py -vv -n --mode=full -- -k "Pinv and generalized_sq_cases and not Matrix" --count=100 -x || true`
- also, 50-minute cap as usual
- pocketfft tests somehow changing a flag?--Ralf suggests just skipping the test and moving on...
- [[merged PR]](https://github.com/MacPython/numpy-wheels/pull/45): all NumPy Linux wheel testing builds were missing gfortran in container that runs final tests, so had > 550 tests skipping (32 & 64-bit); now down to ~75 skips
Please remember to archive this file and commit it to github.com/bids-numpy/docs