Windows: MSVC: Allocate wave_slice_dq array using mem_alloc_dq()

[jhunkeler]

Resolves N/A

Closes #6466, closes #6905

Related: #5578, spacetelescope/crds#790 (comment)

This replaces two variable length array declarations that are preventing Microsoft CL from succeeding. (Refer to #6466 above.)

I tested locally with:

• Windows 10 Pro
• MSVC 2022
• Mambaforge

$ pytest -rsv jwst/cube_build/
============================= test session starts ==============================
platform linux -- Python 3.10.9, pytest-7.2.2, pluggy-1.0.0
crds_context: jwst_1064.pmap
rootdir: /home/jhunk/Downloads/omg/jwst, configfile: setup.cfg
plugins: asdf-2.14.3, requests-mock-1.10.0, openfiles-0.5.0, doctestplus-0.12.1, cov-4.0.0, ci-watson-0.6.1, jwst-1.9.5.dev37+g73135ae5b.d20230313
collected 16 items                                                             

jwst/cube_build/tests/test_configuration.py .....                        [ 31%]
jwst/cube_build/tests/test_cube_build_step.py .                          [ 37%]
jwst/cube_build/tests/test_miri_cubepars.py ...                          [ 56%]
jwst/cube_build/tests/test_nirspec_cubepars.py .                         [ 62%]
jwst/cube_build/tests/test_wcs.py ......                                 [100%]

============================== 16 passed in 3.48s ==============================

PS - I also noticed drizzle-1.13.7 fails to build on Windows, however at the time of this writing drizzle/master and drizzle/1.13.7 are the same... yet pip install git+https://github.com/spacetelescope/drizzle actually succeeds. Very strange. Maybe the archive is missing something?

Checklist for maintainers

• added entry in CHANGES.rst within the relevant release section
• updated or added relevant tests
• updated relevant documentation
• added relevant milestone
• added relevant label(s)
• ran regression tests, post a link to the Jenkins job below.
  How to run regression tests on a PR
• Make sure the JIRA ticket is resolved properly

[pllim]

CI is also failing on master, so the failures here probably unrelated.

[codecov]

Codecov Report

Patch and project coverage have no change.

Comparison is base (141ecbc) 77.56% compared to head (cc5555d) 77.56%.

Additional details and impacted files

@@           Coverage Diff           @@
##           master    #7491   +/-   ##
=======================================
  Coverage   77.56%   77.56%           
=======================================
  Files         452      452           
  Lines       36211    36211           
=======================================
  Hits        28086    28086           
  Misses       8125     8125           

Flag	Coverage Δ		*Carryforward flag
nightly	77.57% <ø> (ø)		Carriedforward from 141ecbc

*This pull request uses carry forward flags. Click here to find out more.

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

☔ View full report in Codecov by Sentry.
📢 Do you have feedback about the report comment? Let us know in this issue.

[jemorrison]

I have a similar fix to some other cube_build updates on a branch I am working on and I have tested it quite a lot. This fix cleans up the remaining memory leaks.

[mcara]

I do not think the old code could have resulted in a memory leak but allocating large arrays on stack is not a good idea in general and allocating variable size arrays is not supported by all compilers (not a standard C feature depending on the standard version). For MS, in particular see https://devblogs.microsoft.com/cppblog/c11-and-c17-standard-support-arriving-in-msvc/#variable-length-arrays

[jemorrison]

Running on a Mac M1 computer - the memory is smaller now than before the change to the dq arrays.

[pllim]

Just merge this already! 😸

[jemorrison]

I have a pr with exactly the same updates + other stuff. So if you merge this one I will rebase my code to grab you merge. It would helpful to merge this one soon because I want to submit my pr today and I can get it clean up with including this pr before I do that.

[pllim]

I don't have the power. Otherwise, I would have done it already. 😸

[hbushouse]

Just curious, what does MSVC stand for?

[mcara]

https://en.m.wikipedia.org/wiki/Microsoft_Visual_C

[mcara]

Comment could be improved

[pllim]

Thanks!

[pllim]

@zacharyburnett , are you sure jwst won't work on Windows? Though hard to tell without actually running CI on it.

xref spacetelescope/hstcal#517

[zacharyburnett]

@zacharyburnett , are you sure jwst won't work on Windows? Though hard to tell without actually running CI on it.

xref spacetelescope/hstcal#517

Oh nice, I didn't know about this! Shows what I know haha 😅

I can add some CI tests to the weekly cron to see how stable Windows support is

EDIT: made a PR #7562