The associated patch is not (yet) intended as a contribution to 0AD. It's a trimmed down version of the test case introducted by D1511, which caused a Jenkins test error on macos (but not on VS build).
The idea of the patch is to have the little bit of code which utilizes UniqueRange execute without the "overhead" of ArchiveReader_Zip and other code found in archive_zip.cpp. If the test case of this patch still crashes, it indicates that there is a problem in the allocator code.
Currently, no review required as the code is not yet intended for production. However, should the test indicate an error in the allocator code, it has to be extended into a bugfix with review and everything.