On 12/9/21 22:59, Jan Nieuwenhuizen wrote:
We are using the canonical form as an automatic include guard, to not
include the same file twice.
Gnulib's same-inode module is often a better way to attack that problem.
On 12/10/21 02:22, Vincent Vermilya wrote:
If DEST directory has a trailing slash in "mv -T --backup=numbered", mv
gets stuck in an infinite loop. This does not happen without the trailing
slash. strace shows renameat2 as the call that gets looped on.
Thanks for the bug report. This is due to a
On 12/10/21 04:06, Jan Nieuwenhuizen wrote:
(thanks to your input on my previous patch, we discussed and decided not
to do case folding in our client code either).
That patch still tries to case-fold one-letter drive specs, though - why
bother doing that if you don't case-fold the rest of the
Bruno Haible writes:
Hello Bruno,
Thanks for your quick and thougtful response.
TL;DR: Okay, well too bad. I guess we and everyone else has to patch
their client code, then? For the archives: a version 2 attached that
fixes two bugs.
>> * lib/canonicalize-lgpl.c (filesystem_name)[__MINGW32__]
Hello again!
So...it turns out that the bug report we got had nothing to do with the
file name casing after all, but was about '/' and '\\'; a somewhat
related issue though.
Currently, using CWD=~/.wine/drive_c/windows, calling
canonicalize_file_name on variations of "system.ini" gives three
triv
Marc Nieper-Wißkirchen wrote:
> I could reformulate my question to: "Is there a way to produce a (quiet)
> NaN that won't occur as a result of the C library functions?"
There are 2^51 possible quiet NaNs [1]. I would guess that the hardware and
libc functions produce only a small portion of this s
