The three .h files in your tests are --in effect-- huge data blobs,
which is something that should usually be avoided in source code repositories
as far as is possible. I haven't had a chance to look in detail at how you are using them,
but here a couple of things you may want to investigate:
1. Can the tests be accomplished with smaller sample files?
2. Better yet, can the buffers be --at least partially-- generated programmatic at test time? (as I did with `setup_pe_with_sections` in my tests)
BTW, what was your source for those three buffers?
--
https://gitlab.winehq.org/wine/wine/-/merge_requests/3870#note_46566