Update mimalloc to 3.5.1

[kleisauke]

No description provided.

[sbc100]

No new or remove local patches?

[sbc100]

This directory would be a great candidate for using git subtree ? I'm hoping to move some of our third_party deps over soon.

[kleisauke]

No new or remove local patches?

Indeed, no extra patches were made and all existing patches are reapplied as-is (see e.g. commit d70353c were these patches were removed and commit 4b56f61 were they were reapplied).

I only needed to modify tools/system_libs.py (to add the newly added prim-tls.c file) and rebase the codesize expectations for other.test_malloc_size_mimalloc_O2.

This directory would be a great candidate for using git subtree ? I'm hoping to move some of our third_party deps over soon.

I think this would be a good candidate (the design doc in PR #27593 looks interesting too). Hopefully we can just point it at the upstream sources without any patches. For example, the emscripten/prim.c patch should probably be fairly easy to upstream and the alloc-override.c patch seems doable as well (the only exception is those weak decls, which are only needed for the test_wrap_malloc_mimalloc test, afaik).

[kleisauke]

the only exception is those weak decls, which are only needed for the test_wrap_malloc_mimalloc test, afaik

I just confirmed this locally. After removing those weak decls, test_wrap_malloc_mimalloc fails with:

wasm-ld: error: duplicate symbol: free
>>> defined in /tmp/emscripten_temp_88xc3jzz/test_wrap_malloc.o
>>> defined in /home/kleisauke/emscripten/cache/sysroot/lib/wasm32-emscripten/libmimalloc.a(alloc.o)

wasm-ld: error: duplicate symbol: malloc
>>> defined in /tmp/emscripten_temp_88xc3jzz/test_wrap_malloc.o
>>> defined in /home/kleisauke/emscripten/cache/sysroot/lib/wasm32-emscripten/libmimalloc.a(alloc.o)

wasm-ld: error: duplicate symbol: realloc
>>> defined in /tmp/emscripten_temp_88xc3jzz/test_wrap_malloc.o
>>> defined in /home/kleisauke/emscripten/cache/sysroot/lib/wasm32-emscripten/libmimalloc.a(alloc.o)