Okay that was /s to some extent but I gotta rant, I'm totally convinced that there's still new software today that completely trip over themselves when files or paths have non-ASCII characters, or sometimes even a space. Incompetence didn't go anywhere.
Unicode in filenames can be a bad idea, since there are more than one way to achieve what looks like the same character. So matching patterns could fail if you think it's one way, but it's actually another representation in unicode.
Speaking of which, it blew my mind when I discovered that .EXEs are just zip files compressed archives. Same goes for .DLLs, and a lot of other common Windows file extensions as well. (.DOC too, for example IIRC). They all open in your favorite archiver software (I like NanaZip; which is a fork of 7-Zip with a modern UI).
I don't think that's true for .exe or .dll files, but it's definitely true for .docx files and other Office files ending with x. Some .exe's are self-extracting archives or have other files embedded in them, so maybe that's what you've been seeing.
Just because they open in 7-Zip or whatever doesn't mean they are just a zip file. There are several kinds of archives. EXEs are a special case as well. They aren't archives at all. Rather they can contain archives or extra content along with being an executable. One reason is self extracting archives. Here an archive is packaged with an extraction program as an exe all in one. The other case is exes that have extra resources like images, videos, graphics textures, etc. Either way it's an executable plus some extra stuff, not a zip archive. DLLs I am not sure about, but I suspect something similar is happening here.
Next time you should research stuff before posting it on Lemmy. Things are sometimes more complicated than they appear.
docx you are correct about though. Specifically it's a zip file that contains XML files and resources.