On 2024-08-10 17:57, Collin Funk wrote:
All seem to have Y2K:
Yes, and that's the correct value. Thanks to both you and Bruno for
reporting it. I installed the attached.From 8fa1fba5dab1a126c25e534867f653359039d177 Mon Sep 17 00:00:00 2001
From: Paul Eggert
Date: Sat, 10 Aug 2024 21:47:07 -07
Hi Paul, Bruno,
Bruno said:
> Today, one of the continuous integrations (on Debian 10, with gcc)
> fails:
My host system (Fedora 40 BTRFS) fails the same three tests but in
different places:
test-utimens.h:197: assertion 'st2.st_atime == BILLION' failed
FAIL test-fdutimensat (exit statu
Hi,
I've pushed this patch + tests adding htonl, htons, ntohl, ntohs to
arpa/inet.h. I've just named the module 'htonl' since POSIX documents
these functions under that page [1]. Doing a separate for each seems
excessive. It's all just reusing endian.h anyways.
Collin
[1] https://pubs.opengroup.
Patrice Dumas noticed [1] that gperf-generated files from gnulib modules
contain not only the source file name, but also the directory name of
that source file name. It occurs
- in the gperf command that is replicated at the top of the file,
- in the many #line statements.
This is problematic
