Public SourceMage Discussion List

[Arwed von Merkatz <v.merkatz AT gmx.net>]

On Sun, Oct 31, 2004 at 05:26:46PM -0600, Robin Cook wrote:
> 
> On Sun, 2004-10-31 at 09:59 +0100, Benoit PAPILLAULT wrote:
> > Eric Sandall a écrit :
> > > Quoting Robin Cook <rcook AT wyrms.net>:
> > > 
> > >>The mysql spell was not detecting the NPTL in glibc for some reason so I
> > >>modified glibc to use config_query and build_api 2. Then modified mysql
> > >>to grep the glibc.p file for NPTL="y".
> > 
> > This kind of fix is wrong from my point of view. glibc.p is something 
> > that is not documented and which is "internal" to the sorcery api. What 
> > was the exact reason that mysql does not detect ntpl? What happens if 
> > you have compiled glibc without using sorcery?
> 
> If something was done outside of sorcery then I can not be held
> responsible to determine what was done.
> 
> Then it needs something else to test against to see if glibc was
> compiled with NPTL.  The test that was there before was not working so
> this was the only thing I could think of to get it working.

*sigh*
The glibc people broke the way we detect NPTL glibc in the spells that
need to detect it. We used to do this:
if  /lib/libc.so.6  |  grep  -q  NPTL;  then
  fix spell for NPTL
fi

With the latest glibc snapshot in devel this is broken, as it doesn't
contain the string NPTL anymore:
arwed@Otherland:~$ /lib/libc.so.6 
GNU C Library stable release version 2.3.3, by Roland McGrath et al.
Copyright (C) 2004 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.
There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A
PARTICULAR PURPOSE.
Compiled by GNU CC version 3.4.2.
Compiled on a Linux 2.6.5 system on 2004-10-12.
Available extensions:
        GNU libio by Per Bothner
        crypt add-on version 2.1 by Michael Glad and others
        Native POSIX Threads Library by Ulrich Drepper et al
        BIND-8.2.3-T5B
        NIS(YP)/NIS+ NSS modules 0.19 by Thorsten Kukuk
Thread-local storage support included.
For bug reporting instructions, please see:
<http://www.gnu.org/software/libc/bugs.html>.

In my opinion the best solution to this is to reverse your change to the
mysql spell and patch glibc to contain the string NPTL in that output.
I'm just glad we didn't integrate that glibc into test yet, it'll break
every spell that has to detect NPTL.
The other solution would be to grep for the new string instead, fixing
it in all spells. Though that would mean we would have to grep for both
strings, as people don't necessarily have the latest glibc snapshot
installed.
Opinions, Suggestions?

> And it is not that mysql is not detecting NPTL but that it is looking
> specifically for LINUX_THREADS and if it doesn't see it it fails on
> configure.  The detection of NPTL allows the applying of a workaround to
> get it past configure.  See http://bugs.mysql.com/bug.php?id=2173
> 
> If there was a sorcery mechanism to test for the config_query I would
> use that.
> 
> CuZnDragon
> Robin Cook
> 



> _______________________________________________
> SM-Discuss mailing list
> SM-Discuss AT lists.ibiblio.org
> http://lists.ibiblio.org/mailman/listinfo/sm-discuss


-- 
Arwed v. Merkatz                              Source Mage GNU/Linux developer
                                                    http://www.sourcemage.org