On Sun, Feb 05, 2006 at 10:41:54AM +0100, Ralf Stephan wrote:
Regarding the right way of handling the stdint/inttypes case, Jon, I don't like fiddling with configure if it's just a glibc issue where we can use __GLIBC_MINOR__ etc. for discrimination of that case, but to prove it's glibc, either Jim has to upgrade or I'd have to reproduce the error somehow...
http://lists.gnu.org/archive/html/autoconf/2001-12/msg0034.html
So we know Jim's system now, and they were the only ones to include a buggy development version of glibc-2.2.4. Now if we had a macro that identifies Mandrake 8.1...
As I write this, that URL seems to be dead ("...not found on this server"), but I'll take Ralf's remark to imply that he thinks I'm on a Mandrake 8.1 system. I'm actually on what was once a SuSE 7.3 system, before I upgraded Gcc, binutils, X, and most everything else except glibc, as it seems a bit more challenging upgrade, and up until now, hasn't been an issue. I'll keep trying that URL, as it sounds interesting, and I'll consider embarking on the glibc journey, but that may take a while.
After spending way too many hours working my way through SuSE 9.2 source code only to discover they'd hacked my SCSI driver beyond use, I decided my time would be better spent upgrading what I need, rather than debugging an unresponsive vendor's product.
This is the first gettext-related issue I've encountered. I had to set an environment variable, LIBS=-liconv, to get fontconfig to configure the link process properly, but nothing has had any compile issues with gettext.
Regarding glibc-2.2.4, here's what mine says to --version: Compiled by GNU CC version 2.95.3 20010315 (SuSE). Compiled on a Linux 2.4.9 system on 2001-09-20. Available extensions: GNU libio by Per Bothner crypt add-on version 2.1 by Michael Glad and others Berkeley DB glibc 2.1 compat library by Thorsten Kukuk linuxthreads-0.9 by Xavier Leroy NoVersion patch for broken glibc 2.0 binaries BIND-8.2.3-T5B libthread_db work sponsored by Alpha Processor Inc NIS(YP)/NIS+ NSS modules 0.19 by Thorsten Kukuk
Sounds like the test should be more glibc-specific rather than looking for Mandrake.
Cheers,
Jim