[PATCH 1/2] posix: User scratch_buffer on fnmatch

[Paul Eggert]

On 1/14/21 3:44 AM, Adhemerval Zanella wrote:
> I think we still should fix BZ#14185 with the suggested with the strategy of
> falling back to non wide mode in case of encoding error.

For glibc 2.33 that's likely the best we can do.

> The full fix will take some time, since it is pretty much a rewrite of the
> fnmatch_loop.c.

Yes. Should be doable for 2.34 though, I'd hope.