Hi, I've done a couple of upgrades to major packages locally (and rebuilt everything) to see if something might be relevant for 2.3, namely: coreutils 6.3: chgrp & friends seem to break in a fakeroot environment, causing any usage of them in ports to have no effect. This is likely a fakeroot issue, but IMO big enough to hold back the upgrade until a fix/workaround is available. db 4.5.20: caused python to fail compiling its bsddb module because some functions that have been marked as deprecated for a long time are actually gone now - I added a couple of #ifdef's to fix this, seems to work fine. glibc 2.5: I've tested it with both our current kernel headers (whatever their origin is is only known to danm, I guess...) and the 2.6.18 headers from kernel.org. strace seems to not build any longer (with both header sets), but that's pretty much it! gnutls 1.4.4 (as well as all other gcrypt projects): nothing except seahorse (gnome gpg front-end) doesn't crash as much anymore, probably because we're about the only distro shipping 1.0.x (well, together with debian stable :P) python 2.5: everything compiles, all small apps work, but some of the big ones (e.g. quodlibet, a media player that uses the gstreamer and gnome python bindings - 'nuff said) don't. I think we're better off waiting a month or two for wider adoption. util-linux 2.13-pre7: older versions abuse the kernel headers, which the kernel.org-provided ones don't allow - so this upgrade is needed. Also, they switched to autoconf, so the config patching isn't needed any longer. However the build breaks with --disable-nls - haven't looked into fixing that yet.
On Sun, Oct 01, 2006 at 09:06:39 +0200, Mark Rosenstand wrote:
Hi, [...] glibc 2.5: I've tested it with both our current kernel headers (whatever their origin is is only known to danm, I guess...) and the 2.6.18 headers from kernel.org. strace seems to not build any longer (with both header sets), but that's pretty much it! Not sure what your error was, but I can reproduce a build failure here as well (use of undefined symbols PT_CR_IPSR and PT_CR_IIP); strace uses symbols which are only defined in asm-ia64/ptrace_offsets.h (as of 2.6.18).
Related change: http://sources.redhat.com/ml/libc-alpha/2006-09/msg00027.html I guess we're still a bit away from using unpatched kernel headers for glibc. Trivial patch to make strace work with 2.5 attached. Johannes -- Johannes Winkelmann mailto:jw@smts.ch Zurich, Switzerland http://jw.smts.ch
On Sun, Oct 01, 2006 at 09:06:39AM +0200, Mark Rosenstand wrote: [...]
glibc 2.5: I've tested it with both our current kernel headers (whatever their origin is is only known to danm, I guess...)
They are borrowed from SuSE. Jürgen -- Juergen Daubert | mailto:jue@jue.li Korb, Germany | http://jue.li/crux
Hello, friends! Juergen Daubert wrote:
On Sun, Oct 01, 2006 at 09:06:39AM +0200, Mark Rosenstand wrote: [...]
glibc 2.5: I've tested it with both our current kernel headers (whatever their origin is is only known to danm, I guess...)
They are borrowed from SuSE.
I've tried the more or less new make headers_install and didn't run into any problems over here. You might want to read: http://lkml.org/lkml/2006/7/2/24 http://lwn.net/Articles/181094/ Best greets, Clemens Koller _______________________________ R&D Imaging Devices Anagramm GmbH Rupert-Mayer-Str. 45/1 81379 Muenchen Germany http://www.anagramm.de Phone: +49-89-741518-50 Fax: +49-89-741518-19
Dear Clemens, On Mon, Oct 02, 2006 at 12:51:02 +0200, Clemens Koller wrote:
Hello, friends!
Juergen Daubert wrote:
On Sun, Oct 01, 2006 at 09:06:39AM +0200, Mark Rosenstand wrote: [...]
glibc 2.5: I've tested it with both our current kernel headers (whatever their origin is is only known to danm, I guess...)
They are borrowed from SuSE.
I've tried the more or less new make headers_install and didn't run into any problems over here. This target wasn't available back when crux 2.2 was released as it was only introduced in 2.6.18, this is why the SuSE headers were chosen earlier.
HTH, Johannes -- Johannes Winkelmann mailto:jw@smts.ch Zurich, Switzerland http://jw.smts.ch
Hi, On Mon, Oct 02, 2006 at 12:51:02 +0200, Clemens Koller wrote:
Hello, friends!
Juergen Daubert wrote:
On Sun, Oct 01, 2006 at 09:06:39AM +0200, Mark Rosenstand wrote: [...]
glibc 2.5: I've tested it with both our current kernel headers (whatever their origin is is only known to danm, I guess...)
They are borrowed from SuSE.
I've tried the more or less new make headers_install and didn't run into any problems over here.
Forgot to say: if you use glibc 2.5 and the headers from headers_install, you _will_ run into compilation problems with both strace and util-linux in their current versions from core. Johannes -- Johannes Winkelmann mailto:jw@smts.ch Zurich, Switzerland http://jw.smts.ch
Hello, Johannes!
glibc 2.5: I've tested it with both our current kernel headers (whatever their origin is is only known to danm, I guess...)
They are borrowed from SuSE.
I've tried the more or less new make headers_install and didn't run into any problems over here.
Forgot to say: if you use glibc 2.5 and the headers from headers_install, you _will_ run into compilation problems with both strace and util-linux in their current versions from core.
I didn't test those yet. Can you give me a brief update about the current status here? What are the problems? Who is working on fixes? What are the suggested solutions? Thanks, Clemens Koller _______________________________ R&D Imaging Devices Anagramm GmbH Rupert-Mayer-Str. 45/1 81379 Muenchen Germany http://www.anagramm.de Phone: +49-89-741518-50 Fax: +49-89-741518-19
On Mon, Oct 02, 2006 at 15:07:25 +0200, Clemens Koller wrote:
Hello, Johannes!
glibc 2.5: I've tested it with both our current kernel headers (whatever their origin is is only known to danm, I guess...)
They are borrowed from SuSE.
I've tried the more or less new make headers_install and didn't run into any problems over here.
Forgot to say: if you use glibc 2.5 and the headers from headers_install, you _will_ run into compilation problems with both strace and util-linux in their current versions from core.
I didn't test those yet. ...
Can you give me a brief update about the current status here? glibc 2.5 was released last friday (i.e. 3 days ago); 2.6.18 was released 13 days ago.
What are the problems? Who is working on fixes? Some applications make assumptions about the kernel headers which aren't met anymore. Once wider adaption of the new glibc and kernel header starts the fixes will make it to the appropriate places.
I've have updated ports in my httpup, but unless you know what you're doing (and judging from your questions you don't) I'd strongly advice you not to use them.
What are the suggested solutions? Use the current ports, they work fine.
HTH, Johannes -- Johannes Winkelmann mailto:jw@smts.ch Zurich, Switzerland http://jw.smts.ch
On Mon, 2006-10-02 at 15:07 +0200, Clemens Koller wrote:
Hello, Johannes!
glibc 2.5: I've tested it with both our current kernel headers (whatever their origin is is only known to danm, I guess...)
They are borrowed from SuSE.
I've tried the more or less new make headers_install and didn't run into any problems over here.
Forgot to say: if you use glibc 2.5 and the headers from headers_install, you _will_ run into compilation problems with both strace and util-linux in their current versions from core.
I didn't test those yet. Can you give me a brief update about the current status here? What are the problems? Who is working on fixes? What are the suggested solutions?
Did you even read my initial message?
Hello, Mark! Mark Rosenstand wrote:
On Mon, 2006-10-02 at 15:07 +0200, Clemens Koller wrote:
Hello, Johannes!
glibc 2.5: I've tested it with both our current kernel headers (whatever their origin is is only known to danm, I guess...)
[...]
Did you even read my initial message?
Yup, I did. I expected to get some details about the problems in the form of messages of the build errors (around glibc 2.5) to get an idea about which things are broken there. But ... not today. Best greets, Clemens Koller _______________________________ R&D Imaging Devices Anagramm GmbH Rupert-Mayer-Str. 45/1 81379 Muenchen Germany http://www.anagramm.de Phone: +49-89-741518-50 Fax: +49-89-741518-19
participants (4)
-
Clemens Koller -
Johannes Winkelmann -
Juergen Daubert -
Mark Rosenstand