fix(meson): force-include Cap'n Proto MSVC interface guard

[HaoZeke]

Summary

MSVC windows.h defines interface, which breaks Cap'n Proto templates. The feedstock was putting /FI msvc_capnp_guard.h in CXXFLAGS; Meson 1.12 feeds those flags into the cl.exe sanity compile, a drive-letter include becomes /FID:/..., and setup dies with "Compiler cl.exe cannot compile programs."

add_project_arguments('/FI', ...) after project() applies the force-include only to the real compile. Same guard as conda-forge/eon-feedstock#30 (0001-win-msvc-capnp-fi-after-project.patch). After the next tarball the feedstock can drop that patch.

The remaining win-64 activation work (clear GNU LDFLAGS so cl does not see -Wl,... / D8021; clang_rt / flang_rt LIBPATH) stays on the feedstock. That is compiler-activation, not eOn source.

Test plan

• eon 3.2.0 conda-forge/eon-feedstock#30 win_64 py3.10-3.14 SUCCESS with this source change as a recipe patch
• GHA Windows job on this branch

[codecov-commenter]

⚠️ Please install the to ensure uploads and comments are reliably processed by Codecov.

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!

[github-actions]

Benchmark Results

Note

All benchmarks unchanged

	Count
⚪ Unchanged	8

8 unchanged benchmark(s)

Benchmark	Before	After	Ratio
bench_eonclient.TimeMinimizationLJCluster.peakmem_minimization_lbfgs	39M	39M	~1x
bench_eonclient.TimeMinimizationLJCluster.time_minimization_lbfgs	23.7±0ms	21.8±0ms	~0.92x
bench_eonclient.TimeNEBMorsePt.peakmem_neb	38.9M	38.9M	~1x
bench_eonclient.TimeNEBMorsePt.time_neb	206±0ms	207±0ms	~1.01x
bench_eonclient.TimePointMorsePt.peakmem_point_evaluation	39M	38.9M	~1x
bench_eonclient.TimePointMorsePt.time_point_evaluation	10.4±0ms	10.4±0ms	~1x
bench_eonclient.TimeSaddleSearchMorseDimer.peakmem_saddle_search_dimer	39M	39M	~1x
bench_eonclient.TimeSaddleSearchMorseDimer.time_saddle_search_dimer	59.6±0ms	60.1±0ms	~1.01x

Details

• Base: be4ed4f1
• Head: 9ce90d81
• Runner: ubuntu-22.04

Raw asv-spyglass output

All benchmarks:

| Change   | Before   | After    |   Ratio | Benchmark (Parameter)                                                  |
|----------|----------|----------|---------|------------------------------------------------------------------------|
|          | 39M      | 39M      |    1    | bench_eonclient.TimeMinimizationLJCluster.peakmem_minimization_lbfgs   |
|          | 23.7±0ms | 21.8±0ms |    0.92 | bench_eonclient.TimeMinimizationLJCluster.time_minimization_lbfgs      |
|          | 38.9M    | 38.9M    |    1    | bench_eonclient.TimeNEBMorsePt.peakmem_neb                             |
|          | 206±0ms  | 207±0ms  |    1.01 | bench_eonclient.TimeNEBMorsePt.time_neb                                |
|          | 39M      | 38.9M    |    1    | bench_eonclient.TimePointMorsePt.peakmem_point_evaluation              |
|          | 10.4±0ms | 10.4±0ms |    1    | bench_eonclient.TimePointMorsePt.time_point_evaluation                 |
|          | 39M      | 39M      |    1    | bench_eonclient.TimeSaddleSearchMorseDimer.peakmem_saddle_search_dimer |
|          | 59.6±0ms | 60.1±0ms |    1.01 | bench_eonclient.TimeSaddleSearchMorseDimer.time_saddle_search_dimer    |