2.12 obviously doesn't. :-(
Recent releases have tarballs again, check ftp://ftp.gnu.org/pub/gnu/glibc/
Re: How good is a release without tarball?
> Now, seriously, when will we see the
> next tarball release? Just
> a random CVS tag really is not good
> enough for such a vital
> system component. It's time to do the
> extra homework of
> putting a gold master release tarball
> up. Not only for comfort,
> but also for reference.
Agreed. This whole CVS only is crap. It should be a proper signed tarball that people can actually verify.
How good is a release without tarball?
Now, seriously, when will we see the next tarball release? Just
a random CVS tag really is not good enough for such a vital
system component. It's time to do the extra homework of
putting a gold master release tarball up. Not only for comfort,
but also for reference.
Re: Check Failing
> % I've had no trouble installing all
> % nessary versions of everything (egcs,
> % binutils, etc) and the compliled and
> % installed fine. Glibc 2.1 complies
> % but when I "make check" it stopes at
> % glob test in the postix directory.
> % know what would cause this?
> Yes, you probably have an earlier
> version of bash (pre 2.0.5? I think is
> the latest version). If you upgrade to
> the latest bash, globbing is fixed.
Mhh i used the latest bash while compiling glibc-2.3.3-20031202 and have the same trouble
Re: Upgrading from 2.1.3 to 2.3?
> Anyone got any links on upgrading from
> glibc 2.1.3 to 2.3? I want to upgrade my
> Red Hat Linux 6.2 system to 2.3 as the
> 2.1.3 it comes with is very old.
it's year late, but i just upgraded my hat 6.2 to glibc
2.3.2. i had to upgrade binutils to the most recent,
and had gcc 3.2.2. it went smoothly. the only
software i've found that quit working was a pre-
2000 version of wine that i need, and i've been
getting compilation errors about undefined reference
to 'errno.h'. thats all i know so far.
GLIBC 2.3.2 Compiling Weirdness
Ok. after 3-4 days of trying to figure out what is going on, I finally brought it down to this:
GLIBC 2.3.2 will not build from a build directory that looks like this: /build
But it will build without any problem in /setup/build
Anybody has a clue why? This one gets the weirdest compilation problem award...
GLIBC 2.3.1 does not have this problem.
Re: Compiling GLIBC 2.3.2 with GCC 3.3
Ok as it turns out, I successfully built GLIBC 2.3.2 with GCC 3.3 on my other system.
I'm using binutils-18.104.22.168 on the system where it does not compile, and binutils-2.13.1 on the one where it does. I tried replacing simply ld with 2.13.1, but without any success... What else could cause this?
Compiling GLIBC 2.3.2 with GCC 3.3
I've been searching the web and IRC for help with this without any luck... For some reason no one else seems to have this problem. I believe all the packages on this system are the latest and greatest and generally proved to work well so far compiling each other one of them.
First I got the sscanf issues which was using the old protoypes for variable arguments ... I changed the prototype to include the "..." at the end of the argument list and the compilation got further. But now I keep getting undefined errors on stuff like 'strcmp', 'errno', '__libc_open', '__libc_missing_32bit_uids', 'memcpy', etc. The undefined references are from '<build dir>/elf/librtld.os, with a line apparently trying to build "ld.so". Anybody could bring some light to this really annoying problem? I have a built version of this very same version of GLIBC built on my other system, but built with GCC 3.2.2. Unfortunately I upgraded it to GCC 3.3 as well and don't wanna go through the hassle of downgrading the compiler. Thanks for any help.
Re: Compilation problems
The same with my system. I have s slack 9.0 box with an rpm mod to go above tgz, rpm segfaults. Evem when using glibc-2.3.1 to compile glibc-2.3.2, I get some iconv internal error(s). Weird.
Will wait until 2.3.3 is out.
An open, cross-platform journaling program.
A scientific plotting package.