[PATCH] D41830: [libc++] Fix PR#35780 - make std::experimental::filesystem::remove and remove_all return false or 0 if the file doesn't exist

[Marshall Clow via Phabricator via cfe-commits]

mclow.lists added inline comments.


================
Comment at: test/std/experimental/filesystem/fs.op.funcs/fs.op.remove/remove.pass.cpp:67
         non_empty_dir,
-        file_in_bad_dir,
+        // file_in_bad_dir, // produces "St13exception_ptruncaught_exceptions not yet implemented"
     };
----------------
mclow.lists wrote:
> TyanNN wrote:
> > mclow.lists wrote:
> > > TyanNN wrote:
> > > > This error is quite strange: it is not inherited from std::exception, one can only really find our the error with std::current_exception, and it happens when you use the path in fs::remove and fs::remove_all. Any idea about this? @mclow.lists
> > > I'm not seeing this error on my system (nor are any of the bots reporting it AFAIK).
> > > 
> > > That looks like "`uncaught_exceptions` is not yet implemented", and that comes from the ABI library.
> > > 
> > > What are you using for an ABI library? Is it libc++abi?
> > > 
> > Yes, I use libc++abi from trunk that I've built in llvm tree with libc++. I don't have a system wide installation of libc++abi so that must be it. Does uncommenting the line really work on your system? For the tests report an uncaught exception and fail.
> It's definitely not commented out on my system.
> (or on the trunk)
> 
I see the message "uncaught_exceptions not yet implemented" in src/support/runtime/exception_fallback.ipp - but only in an #ifdef block for `__EMSCRIPTEN__`


https://reviews.llvm.org/D41830