med-file: specifically use python@3.10 with cmake

[ipatch]

• Have you ensured that your commits follow the commit style guide?

brew style freecad/freecad/[NAME_OF_FORMULA_FILE] 

output from running above command should output something similiar to the below

1 file inspected, no offenses detected

• Have you ensured your commit passed audit checks, ie.

brew audit freecad/freecad/[NAME_OF_FORMULA_FILE] --online --new-formula

---

Not all PRs require passing these checks ie. adding [no ci] in the commit message will prevent the CI from running but PRs that change formula files generally should run through the CI checks that way new bottles are built and uploaded to the repository thus not having to build all formula from source but rather installing from a bottle (significantly faster 🐰 ... 🐢)

For more information about this template file learn more

[chennes]

Just a confirmation that this resolves the medfile issue on my M2.

[ipatch]

current CI error with big sur, leaving this here for future me 👴, posterity, spiders 🕷, and the AI overlords (skynet™)

==> Running Formulae#formula!(freecad/freecad/med-file@4.1.1)
==> brew audit --strict --only=gcc_dependency freecad/freecad/med-file@4.1.1
==> FAILED
Full audit --strict --only=gcc_dependency freecad/freecad/med-file@4.1.1 output
  Warning: freecad/freecad/med-file@4.1.1 should not have a Linux-only GCC dependency!
  homebrew/core
    * pypi_formula_mappings.json references
      formulae or casks that are not found in the homebrew/core tap.
      Invalid formulae or casks: shub
  Error: 1 problem in 1 tap detected.
==> SKIPPED freecad/freecad/med-file@4.1.1

update

it seems manually running brew update resolves the above audit exception however further errors require troubleshooting.

[ipatch]

current build error on big sur, grabbed from gha log (logs will delete from server, thus the copy paste)

 cd /tmp/med-fileA4.1.1-20230624-24288-1blqjyi/med-4.1.1_SRC/build/src/2.3.6/ci && /usr/local/Homebrew/Library/Homebrew/shims/mac/super/clang -DH5_USE_16_API -DMED3_USESTATIC -DMED_API_23 -I/tmp/med-fileA4.1.1-20230624-24288-1blqjyi/med-4.1.1_SRC/include/2.3.6 -I/tmp/med-fileA4.1.1-20230624-24288-1blqjyi/med-4.1.1_SRC/build/include/2.3.6 -I/usr/local/Cellar/hdf5/1.14.1/include -I/usr/local/opt/libaec/include -O3 -DNDEBUG -isysroot /Library/Developer/CommandLineTools/SDKs/MacOSX11.sdk -fPIC -MD -MT src/2.3.6/ci/CMakeFiles/_ci236_static.dir/MED23v30axesInfo.c.o -MF CMakeFiles/_ci236_static.dir/MED23v30axesInfo.c.o.d -o CMakeFiles/_ci236_static.dir/MED23v30axesInfo.c.o -c /tmp/med-fileA4.1.1-20230624-24288-1blqjyi/med-4.1.1_SRC/src/2.3.6/ci/MED23v30axesInfo.c
  [ 34%] Built target _ci236_static
  make: *** [all] Error 2

---

helpful error message

/tmp/med-fileA4.1.1-20230625-71487-ujbv2w/med-4.1.1_SRC/src/hdfi/_MEDfileCreate.c:163:2: error: "Don't forget to change the compatibility version of the library !"
#error "Don't forget to change the compatibility version of the library !"
 ^
1 error generated.

[ipatch]

false positive with catalina

Warning: freecad/freecad/med-file@4.1.1 has dependencies without compatible bottles:
==> SKIPPED freecad/freecad/med-file@4.1.1
  cmake
  python@3.11
  gcc
  hdf5

update

seemed to have resolved no bottles for deps on macos catalina by creating a new github repo for macos catalina. 👍 and setting the tip of the master branch to a commit in the homebrew-core git repo where catalina bottles are still present.

[ipatch]

issue / mojave

mojave build errors seem to be related to python@3.11 in how python is looking for a gettext image when the ci is running through its checks, however neither big sur or catalina reference gettext as dep for python@3.11.

possible fix is to manually login to the account that manages the /usr/local/ for building bottles on the mojave vm and manually install gettext, however not sure if brew test-bot will nuke the prior installations of the formula thus have set several env vars in hopes of preventing test-bot from removing prior installs of formula 🤞

---

so even with the gettext formula issue related to python@3.11 the mojave runner still managed to build a bottle and leave the artifact in the root of the git repo ie. $(brew --repo freecad/freecad)

not sure if the bottle artifacts will be uploaded when this PR is assigned a pr-pull label, but can always go back submit another PR where this formula adds gettext for mojave only.