Paul Eggert wrote:
> Thanks for reporting those problems. I installed the attached, which
> fixed things for me on GNU/Linux and AIX. I think it should also put a
> dent into the problems you reported on MS-Windows
Thanks. I confirm it fixes the compilation warning and the test failures on AIX.
Thanks for reporting those problems. I installed the attached, which
fixed things for me on GNU/Linux and AIX. I think it should also put a
dent into the problems you reported on MS-Windows, though I can't easily
test this.
From f34fd6a906b6dafe4a5fba13a622b12358510bef Mon Sep 17 00:00:00 2001
Hi Paul,
On 2022-06-12 you wrote:
> Although I don't have MS-Windows I stared at
> the code a bit and I think I see what might be the problem. I found some
> other potential issues too (basically, I didn't switch from
> fstatat/lstat to readlinkat/readlink often enough so the code is still
> p
On 6/12/22 08:03, Bruno Haible wrote:
Two tests now fail, that succeeded before yesterday's patch.
Thanks for reporting that. Although I don't have MS-Windows I stared at
the code a bit and I think I see what might be the problem. I found some
other potential issues too (basically, I didn't
Paul Eggert wrote:
> Also, now that I'm
> thinking about it, the Gnulib code didn't work if fstatat fails with
> EOVERFLOW and should be using readlinkat instead. Please try the first
> attached patch, which I've installed in Gnulib.
The patch works fine on
- Linux
- FreeBSD (both with and
On 6/8/22 13:27, Lance Fredrickson wrote:
Would be nice to see a fix upstream before more projects update gnulib
and the issue becomes broader.
It sounds like a source-code configuration issue, as your platform's
headers in /usr/include correspond to a kernel newer than what you're
running (w
