[c++/python] Update minimum macOS version to 13.3

[XanthosXanthopoulos]

Issue and/or context:
#3154

Notes for Reviewer:
CI broke on a draft PR the above changes fixed it.

[johnkerl]

@XanthosXanthopoulos there are subtleties with MacOS with "10" and "11" which aren't as simple as what one might think.

Looping in @jdblischak for extra eyes.

Re

CI broke on a draft PR the above changes fixed it.

please share here the link to which draft PR, and which failed CI. (I'm curious to know why that draft PR failed but our main is green with C++ 20 for several days now ...)

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 85.74%. Comparing base (ca00d5b) to head (90fc44f).

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #3361      +/-   ##
==========================================
+ Coverage   85.64%   85.74%   +0.09%     
==========================================
  Files          57       57              
  Lines        6201     6201              
==========================================
+ Hits         5311     5317       +6     
+ Misses        890      884       -6     

Flag	Coverage Δ
python	85.74% <ø> (+0.09%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Components	Coverage Δ
python_api	85.74% <ø> (+0.09%)	⬆️
libtiledbsoma	∅ <ø> (∅)

---- 🚨 Try these New Features:

• Flaky Tests Detection - Detect and resolve failed and flaky tests

[XanthosXanthopoulos]

This failed #3329 with error 'to_chars' has been explicitly marked unavailable here caused by std::format

[jdblischak]

My only concern is from a conda-forge feedstock maintenance perspective:

• We can require a newer macOS SDK in the feedstock builds, but it should be done with caution:

While the default version 10.13 can be overridden using the following changes to the recipe, it should be done as a last resort.

• I checked, and there are already 50 conda-forge feedstocks that set MACOSX_SDK_VERSION to 13.3
• The nightly feedstock builds can currently successfully build the Python clients using C++20 and a macOS SDK version of 11 (Nightly feedstock build failed TileDB-Inc/tiledbsoma-feedstock#230 (comment)). And the code in main already uses std::format, so what is different about the use of std::format in [c++] SOMAColumn [WIP] #3329?
• If we decide to bump to 13.3, we need to update MACOSX_SDK_VERSION (and possibly c_stdlib_version, though we should first try without bumping it) in conda_build_config.yaml and rerendering

[johnkerl]

Excellent feedback @jdblischak -- @XanthosXanthopoulos I think we should hold off merging this PR until we do a health-check PR on tiledbsoma-feedstock using a recipe/meta.yml that points at this PR's commit on TileDB-SOMA. (You can do that if you know how; otherwise I'm happy to.)

Context is we're close to a release, and all releases go through Conda constraints, and there are tricky shoals to navigate on the Conda seas .....

[XanthosXanthopoulos]

I don't know how to do the conda feedstock.
Regarding the different behavior on the PR vs. the nightly, I haven't been able to find anything. Even on the same PR some macos workflows built successfully which is even more baffling.