Re: Libc6 2.1.3-18 broken ?
Ben Collins <[EMAIL PROTECTED]> writes: > On Wed, Apr 25, 2001 at 04:12:16PM -0700, Philippe Troin wrote: > > > > Uh ? Setresuid has been here for ages... > > Check the /usr/include/bits/syscall.h, it's not there. > > > Plus, glibc contains the prototype for setresuid, it's just a missing > > symbol in the library... > > The prototype is always there, the syscall is not always there. Check > /usr/include/linux/ and /usr/include/asm-sparc/, it's not defined > there, so glibc doesn't define it. I stand corrected, libc is not broken. Any idea why linux does not implement setresuid on sparc ? Phil.
Re: Libc6 2.1.3-18 broken ?
On Wed, Apr 25, 2001 at 04:12:16PM -0700, Philippe Troin wrote: > > Uh ? Setresuid has been here for ages... Check the /usr/include/bits/syscall.h, it's not there. > Plus, glibc contains the prototype for setresuid, it's just a missing > symbol in the library... The prototype is always there, the syscall is not always there. Check /usr/include/linux/ and /usr/include/asm-sparc/, it's not defined there, so glibc doesn't define it. > > > I'm currently rebuilding libc6 to see if it's yet an other compilation > > > "chroot" issue. > > > > '"chroot" issue'? The buildd does not suffer from a chroot issue, that > > was a security team issue. 2.1.3-18 was compiled by me on a machine > > running stable. > > I was not sure about that... There are so many broken packages in > debian-sparc (gnupg uninstallable, wdm uninstallable) that I thought > it might have been another compilation-on-unstable uploaded to stable > or the like... > > How do I deal with a package that requires setresuid ? Which package? -- ---===-=-==-=---==-=-- / Ben Collins -- ...on that fantastic voyage... -- Debian GNU/Linux \ ` [EMAIL PROTECTED] -- [EMAIL PROTECTED] -- [EMAIL PROTECTED] ' `---=--===-=-=-=-===-==---=--=---'
Re: Libc6 2.1.3-18 broken ?
Ben Collins <[EMAIL PROTECTED]> writes:
> On Wed, Apr 25, 2001 at 10:47:28AM -0700, Philippe Troin wrote:
> > I have problems with libc6 2.1.3-18: it's missing some symbols...
> >
> > For example setresuid is missing, the following program compile but
> > does not link:
> >
> > #include
> > int main()
> > {
> > return setresuid(0,0,0,0);
> > }
> >
> > The same program links correctly on i386...
>
> The kernel headers (2.2.19pre17) used to compile libc6 2.1.3-18 did not
> define this syscall, so it was not included in libc6 on sparc.
Uh ? Setresuid has been here for ages...
Plus, glibc contains the prototype for setresuid, it's just a missing
symbol in the library...
> > I'm currently rebuilding libc6 to see if it's yet an other compilation
> > "chroot" issue.
>
> '"chroot" issue'? The buildd does not suffer from a chroot issue, that
> was a security team issue. 2.1.3-18 was compiled by me on a machine
> running stable.
I was not sure about that... There are so many broken packages in
debian-sparc (gnupg uninstallable, wdm uninstallable) that I thought
it might have been another compilation-on-unstable uploaded to stable
or the like...
How do I deal with a package that requires setresuid ?
Phil.
Re: Libc6 2.1.3-18 broken ?
On Wed, Apr 25, 2001 at 10:47:28AM -0700, Philippe Troin wrote:
> I have problems with libc6 2.1.3-18: it's missing some symbols...
>
> For example setresuid is missing, the following program compile but
> does not link:
>
> #include
> int main()
> {
> return setresuid(0,0,0,0);
> }
>
> The same program links correctly on i386...
The kernel headers (2.2.19pre17) used to compile libc6 2.1.3-18 did not
define this syscall, so it was not included in libc6 on sparc.
> I'm currently rebuilding libc6 to see if it's yet an other compilation
> "chroot" issue.
'"chroot" issue'? The buildd does not suffer from a chroot issue, that
was a security team issue. 2.1.3-18 was compiled by me on a machine
running stable.
Ben
--
---===-=-==-=---==-=--
/ Ben Collins -- ...on that fantastic voyage... -- Debian GNU/Linux \
` [EMAIL PROTECTED] -- [EMAIL PROTECTED] -- [EMAIL PROTECTED] '
`---=--===-=-=-=-===-==---=--=---'
Libc6 2.1.3-18 broken ?
I have problems with libc6 2.1.3-18: it's missing some symbols...
For example setresuid is missing, the following program compile but
does not link:
#include
int main()
{
return setresuid(0,0,0,0);
}
The same program links correctly on i386...
I'm currently rebuilding libc6 to see if it's yet an other compilation
"chroot" issue.
Phil.
Re: Network booting issues
Hi again, On Tue, 24 Apr 2001, Andrew Sharp wrote: > time, even though I have two disks, because a 10/100 network card + > cheap x86 NFS server is actually a faster disk combo than the local > disks. I compile kernels on it all the time. I also run netscape. I fully agree here. Once I built new kernels on my root on nfs enabled clients as well. Without any problem. > > Andreas Jaehnigen <[EMAIL PROTECTED]> writes: > > > > > elftoaout -o vmlinuz-2.2.14.net vmlinuz-2.2.14 > > > I don't know why that would be necessary, but I load my kernel off > the disk, then use an NFS root. However, I've got a completely Heh, dont you want to have DISKLESS clients? ;) But seriously.. Do you mean floppy disk? Since Suns are capable to boot off from network, the floppy disk solution isnt very charming.. Of course, you must have running an arp server (where the clients get its ip adresses from once linux starts up) and a tftp server. (From there the clients get the a.out boot image from). Andi
Re: Please build sketch and python-mysqldb for unstable
On Wed, Apr 25, 2001 at 02:25:42PM +0200, Gregor Hoffleit wrote: > Hi, > > perhaps a kind maintainer with access to a woody/sid SPARC system could try > to build sketch 0.6.9-3 and python-mysqldb 0.3.5-2, and upload them ? > They're stuck in unstable since a few weeks, and according to the build > daemons log, it not even has been tried to build. > > If it doesn't build, could you please mail me the problems. > > vore.debian.org doesn't have the necessary packages on it, and I don't want > to bug the maintainer team. I just reinserted them into the build queue. If they aren't installed today, then check back, and there will be a log on vore. -- ---===-=-==-=---==-=-- / Ben Collins -- ...on that fantastic voyage... -- Debian GNU/Linux \ ` [EMAIL PROTECTED] -- [EMAIL PROTECTED] -- [EMAIL PROTECTED] ' `---=--===-=-=-=-===-==---=--=---'
Please build sketch and python-mysqldb for unstable
Hi, perhaps a kind maintainer with access to a woody/sid SPARC system could try to build sketch 0.6.9-3 and python-mysqldb 0.3.5-2, and upload them ? They're stuck in unstable since a few weeks, and according to the build daemons log, it not even has been tried to build. If it doesn't build, could you please mail me the problems. vore.debian.org doesn't have the necessary packages on it, and I don't want to bug the maintainer team. Thanks again, Gregor
Re: Mozilla on sparc
Janine Restis writes: > On Tue, 24 Apr 2001, Eduardo Trapani wrote: > > > > > Mozilla is slow, but same as netscape under solaris > > > > > > My Mozilla is not just slow. Image you click on something (not > > > necessarily in the web, but also the task buttons). Then you _hear_ > > > something is happening, but... then you have to wait about 30-60 seconds > > > to _see_ what it's doing. That's not really usable... > > > > That is what is happening to me too. > > > > I telnet to the Sparc box, start the program, let it load the start page > > and then move to another virtual desktop. When I get back to the > > original virtual desktop the mozilla window takes minutes to refresh. > > Same thing with Netscape. Sometimes it takes a while to happen, visit a > > couple of pages with images and you will be able to reproduce it. Well, I have a clone of SS20. The same thing happens: mozilla is too slow to be useable. It takes many second, a minute pehaps, just to refresh a screen. In the past, this machine has been used for a lot of computational work. It runs netscape up to version 3.x reasonably fast. Netscape is also very slow, by still faster than mozilla. Anyhow, netscape now segfaults, perhaps after some libc6 upgrade, just notice it because I don't launch unless it is absolutely necessary. (There is also some posts on the sparc list on this subject. -- PHAM Dinh Tuan | e-mail: [EMAIL PROTECTED] Laboratoire de Modelisation et Calcul | Tel: +33 4 76 51 44 23 BP 53, 38041 Grenoble cedex 9 (France) | Fax: +33 4 76 63 12 63 ---
Re: Mozilla on sparc
> Joshua Uziel writes: >> Hmm, this gives me a bus error that same as the other version. >> It seems to be a glibc issue, Netscape compiled it against 2.0 >> and I have 2.2 on my Woody system. What version of Debian are >> you running? Perhaps something can be done for compatibility >> with the old libraries (see /usr/doc/libc6/FAQ.gz). I'll play >> with that tonight. > I've got potato on the machine... or in other words: > ii libc6 2.1.3-15 GNU C Library: Shared libraries and Timezone > (Yep, I've not upgraded either.) So I've tried extracting some files from potato and setting LD_LIBRARY_PATH for Netscape to see them first. All the necessary libraries seem to be in libc6_2.1.3-18 and xlib6g_3.3.6-11potato32 (except libstdc++.so.2.8, it's an older compatibility library, so it shouldn't depend on glibc 2.2). When I run netscape with these libaries in LD_LIBRARY_PATH, I still get this error: /usr/local/netscape/netscape: error while loading shared libraries: /home/kels/glibc2.1.3/lib/libc.so.6: symbol _dl_debug_impcalls, version GLIBC_2.0 not defined in file ld-linux.so.2 with link time reference I think this is due to the fact that it is still finding /lib/ld-linux.so.2 (ld-2.2.2.so) instead of the potato version (ld-2.1.3.so). The clue given by ldd: % ldd /usr/local/netscape/netscape libXt.so.6 => /usr/X11R6/lib/libXt.so.6 (0x5002c000) libSM.so.6 => /usr/X11R6/lib/libSM.so.6 (0x50092000) libICE.so.6 => /usr/X11R6/lib/libICE.so.6 (0x500ab000) libXmu.so.6 => /usr/X11R6/lib/libXmu.so.6 (0x500d1000) libXpm.so.4 => /usr/X11R6/lib/libXpm.so.4 (0x500f7000) libXext.so.6 => /usr/X11R6/lib/libXext.so.6 (0x50116000) libX11.so.6 => /usr/X11R6/lib/libX11.so.6 (0x50135000) libdl.so.2 => /lib/libdl.so.2 (0x50231000) libstdc++.so.2.8 => /usr/lib/libstdc++.so.2.8 (0x50244000) libm.so.6 => /lib/libm.so.6 (0x50298000) libc.so.6 => /lib/libc.so.6 (0x5032e000) /lib/ld-linux.so.2 => /lib/ld-linux.so.2 (0x5000) So /lib/ld-linux.so.2 is the only one that has the full path specified. I'm not sure how to get around that without breaking other things. Probably something clever could be done with chroot, but my cursory attempt went nowhere. Probably it would be easier to downgrade to potato altogether. Any other insights? Kelsey -- Kelsey Jordahl
Re: Mozilla on sparc
On Tue, 24 Apr 2001, Eduardo Trapani wrote: > Both the stable Mozilla and the one at > pandora.debian.org/~robots101/mozilla run incredibly slow on my Sparc > Station 4m. Other X applications seem to run OK, but Mozilla just > crawls so that it seems it is not working. Somehow that doesn't surprise me. It's sluggish and bloated even on my 750MHz Athlon with 256M RAM and a Matrox G400. Mozilla is a pig. Try Konqeror from the KDE project. I haven't tried much X stuff on my LX (mouse doesn't want to work), but it's certainly much quicker on my Athlon workstation. -- 8<8<8<8<8<8<8< Ian Tester *8)# \7\LINUX: because geeks will find a way [EMAIL PROTECTED] \7\ http://www.zipworld.com.au/~imroy
