[galabovaa]

Description

For scipy build:

scipy/scipy#26440 (comment)

METIS interface move to a single HiGHS-owned header, extern/HighsMetis.h, outside extern/metis so it survives SciPy's trimming. metis/metis.h and HighsExtrasExternalDeps.h both include it.

Tested (Windows, HIPO=ON, openblas[threads]): MSVC and clang-cl both build cleanly. With MSVC, HiPO results and runtimes are unchanged.

valgrind failure is unrelated: ParallelSchedulerIsPerThread uses a 1s timeout in waitFor(), which is too short under valgrind's slowdown. Valgrind found no memory errors, and other branches pass. Suggest raising the timeout to around 30s.

Valgrind runs all threads one at a time on a single core. While thread A spins in waitFor it competes with thread B, which still has to start its scheduler threads and do the parallel sum. A busy CI runner could exceed 1s.
waitFor returns as soon as the flag is set, so in a passing run the timeout never adds any time. It only matters if the threads genuinely deadlock, and then the only effect is that the test fails after 30s instead of 3s.

Checklist

• I have read the contributing guidelines
• This PR targets the latest branch
• Tests are passing
• Documentation was updated where relevant
• This PR is not primarily AI-generated (per the AI contributions policy in CONTRIBUTING.md)

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 75.32%. Comparing base (1001c16) to head (95395c8).
⚠️ Report is 91 commits behind head on latest.

Additional details and impacted files

@@            Coverage Diff             @@
##           latest    #3480      +/-   ##
==========================================
+ Coverage   75.24%   75.32%   +0.07%     
==========================================
  Files         460      460              
  Lines      112567   112769     +202     
  Branches    17836    17856      +20     
==========================================
+ Hits        84697    84938     +241     
+ Misses      27588    27543      -45     
- Partials      282      288       +6     

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[filikat]

Looks fine to me.
The only issue is that we have a duplicate header and we need to make sure that they always match. Maybe we can write our own metis header and store it outside of the metis folder? The current header is already a modified version of the original.

[galabovaa]

@filikat something like this?