I have been trying to solve a bug for about a week now, but I can't
manage to understand what happens:
>From time to time (rather randomly), my CPU gets used at 100% (at least,
my laptop's led blinks a lot), my mouse works (I can move it but not
click) but not my keyboard. Then, about 30 seconds later, even the mouse
stops moving and I have to hard-reboot my laptop.
As I listen to music all day long, it always happens while playing music
(of course it could a coincidence). So, I filed a bug against rhythmbox:
GNOME bugzilla's bug #127814
Then, I noticed that it also happened with XMMS and MPG123, as well as
rhythmbox-gstreamer and rhythmbox-xine. I also tried to
activate/deactivate ESD. Same result. Then, I installed ALSA. The freeze
kept happening. Then I saw that Alan Cox advised to append apm=off,
nohlt, or acpi=off to grub.conf to fix a freeze-at-idle bug. I tried all
3 (in combinations or individually), and none helped.
Sometimes it freezes at idle, and sometimes while I'm using my laptop.
I tried letting top run in a visible terminal. But when it froze, i
couldn't see anything special in top.
What's very strange is that it seems to happen very randomly, but tends
to happen more often on the afternoon (around 4pm) after about 15 min of
I'm really totally lost... please help me solve this very critical bug.
PS: I happily used Redhat 8, then 9 on this very same laptop before and
never had this bug.
PPS: I usually run GNOME-2.4, with Evolution, Rhythmbox and Mozilla.
Julien Olivier <julo(a)altern.org>
For popular ftp sites there usually exist quite a number of mirrors. The
link on the webpage of the original developer/owner to his own ftp site.
Furthermore there is usually a page containing all the mirror sites.
Consquence of this is that ftp site of the original developer/owner gets
most of the traffic and quite some users experience a download which is
slower then would be needed (a mirror close by would probably be faster).
For Windooz there exist programs which you run on your machine which
automatically find the best mirror or download from multiple mirrors
E.g. Download Accelelator http://www.speedbit.com/DAP7/DefaultT.asp?
Disadvantage of this is that you need this extra program.
Isn't it possible to solve this just on the server side, so that for a
user it is just transparent where the file comes from.
I'll try to explain it further with an example.
Let's say redhat publishes a new file and put a link on their webpage
the user which checks which of the mirrors is geographically close by
and has still spare capacity and starts downloading from that file.
This has 3 advantages. The user receives the file earlier, Redhat saves
some bandwidth and its completely transparent
You could also enhance apt or yum with this.
Does this exist? Is it possible? If it possible/exists why does nobody
I was wondering if there were any sites that have CPAN modules already
in RPMs ready to install for say Fedora or RH? (even if its just using
simple cpan2rpm program)
If not would there be any interest in creating one?
I made a new partition on a 60GB drive that was the 4th primary
partition on that drive.
Added this line to /etc/fstab
/dev/hda4 /pub ext3
auto,user,dev,exec,rw,async 0 0
I checked /etc/rc.d/rc.sysinit and the command is there to use 'mount
-a' to mount the
dir at boot. But it fails to do so. mounting it manually does work and
the dir is rw
accessible to all users, as I wanted. The permissions are set correctly.
As root command line 'mount -a' also fails to mount the dir.
I added a line mount /pub to /etc/rc.d/rc.local and
the mount point is properly
brought up and dir is useable for all users.
But why doesn't 'mount -a' work at boot in the rc.sysinit script as I
assume it should?
Wrong assumption? Bug? dumb user?
-----BEGIN PGP SIGNED MESSAGE-----
When compiling wap11gui (http://wap11gui.sourceforge.net), I run into
the following error:
g++ -DHAVE_CONFIG_H -I. -I. -I.. -I/usr/lib/qt-3.1/include
- -I/usr/X11R6/include -DQT_THREAD_SUPPORT -D_REENTRANT
- -DDATADIR=\"/usr/local/share/wap11gui\" -Wnon-virtual-dtor
- -Wno-long-long -Wbad-function-cast -Wundef -Wall -pedantic -W
- -Wpointer-arith -Wmissing-prototypes -Wwrite-strings -ansi
- -D_XOPEN_SOURCE=500 -D_BSD_SOURCE -Wcast-align -Wconversion -O2
- -fno-exceptions -fno-check-new -fexceptions -DNET_SNMP=1 -c -o
macauthimp.o `test -f macauthimp.cpp || echo './'`macauthimp.cpp
cc1plus: warning: "-Wbad-function-cast" is valid for C/ObjC but not for C++
In file included from /usr/include/net-snmp/utilities.h:26,
~ from /usr/include/net-snmp/net-snmp-includes.h:64,
~ from wap11config.h:24,
~ from macauthimp.h:21,
~ from macauthimp.cpp:22:
/usr/include/net-snmp/library/getopt.h:8: error: declaration of `int
~ getopt(int, char* const*, const char*)' throws different exceptions
/usr/include/getopt.h:154: error: than previous declaration `int getopt(int,
~ char* const*, const char*) throw ()'
make: *** [macauthimp.o] Error 1
make: Leaving directory `/home/ori/projects/wap11gui/wap11gui'
make: *** [all-recursive] Error 1
The problem is that /usr/include/net-snmp/utilities.h defines getopt()
as an extern, instead of including <getopt.h>, as it probably should.
The proper getopt.h is included by unistd.h, which, in a QT app, is
included somewhere insode the bowels of QT. This means that every QT app
that tries to use net-snmp under Fedora won't compile. I can't think of
any way an application could work around this.
This appears to be a new problem - RH 9 and 8.0 didn't have this, and I
haven't seen this problem on any other platform I tried compiling on
(Mandrake, SuSe, FreeBSD and Mac OS X).
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.2 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
-----END PGP SIGNATURE-----
Ok, I have an HP netserver 5/133 LS2 I like and use. Under RH8.0 it ran fine
with the SMP kernel. Upgraded to Fedora Core 1 on it, and I no longer have
SMP. Is there something not working with SMP on i586 with NPTL or something?
I'm downloading the kernel SRPM now, and will be rebuilding the kernel for
i586 smp overnight, unless somebody knows something about this.
Director of Information Technology
Pisgah Astronomical Research Institute
1 PARI Drive
Rosman, NC 28772
New package expect
A program-script interaction and testing utility
Removed package tcltk
* Thu Nov 27 2003 Than Ngo <than(a)redhat.com> 3.8.1-3
- rebuild against python 2.3/ Qt 3.2.3
* Fri Nov 28 2003 Tim Waugh <twaugh(a)redhat.com> 2.05b-33
- Speed up UTF-8 command-line redrawing in the common case (bug #102353,
* Sat Nov 29 2003 Tim Waugh <twaugh(a)redhat.com> 3.0.0-12
- Undo over-zealous percent escaping in PostScript.xml
- Build requires libxml2-devel (bug #110589).
- Use relative, not absolute, symlink for CUPS filter.
* Fri Nov 28 2003 Than Ngo <than(a)redhat.com> 2:3.1.93-0.4
- fixed db4 issue
* Thu Nov 27 2003 Than Ngo <than(a)redhat.com> 2:3.1.93-0.3
- get rid of rpath
* Wed Nov 26 2003 Than Ngo <than(a)redhat.com> 2:3.1.93-0.2
- fixed type conflict on x86_64
* Thu Nov 13 2003 Than Ngo <than(a)redhat.com> 2:3.1.93-0.1
- KDE 3.2 Beta1
* Fri Nov 28 2003 Phil Knirsch <pknirsch(a)redhat.com> 2:1.5-6
- Bumped release and rebuilt.
* Fri Nov 28 2003 David Jee <djee(a)redhat.com> 7.4-3
- uncomment buildrequires tcl-devel
* Fri Nov 28 2003 David Jee <djee(a)redhat.com> 7.4-2
* Mon Nov 24 2003 David Jee <djee(a)redhat.com> 7.4-1
- initial Red Hat build
- move jars to /usr/share/java
- fix rpm-multilib patch to use sysconfig
* Fri Nov 21 2003 Lamar Owen <lowen(a)pari.edu> <lamar.owen(a)wgcr.org>
- Development JDBC jars in addition to the 7.3 jars; will replace the
- 7.3 jars once 7.4 official jars are released.
- Changed to use the bzip2 source to save a little size.
- Removed some commented out portions of the specfile.
- Removed the 7.3.4 PDF docs. Will replace with 7.4 PDF's once they
- are ready.
* Tue Nov 18 2003 Kaj J. Niemi <kajtzu(a)fi.basen.net> 7.4-0.1
- Fixed Patch #1 (now rpm-pgsql-7.4.patch)
- Fixed Patch #2 (now rpm-multilib-7.4.patch):
- Patch #4 is unnecessary (upstream)
- Fixed Patch #6 (now postgresql-7.4-src-tutorial.patch)
- Added Patch #8 (postgresql-7.4-com_err.patch) as com_err()
is provided by e2fsprogs and CPPFLAGS gets lost somewhere
inside configure (bad macro?)
- No 7.4 PDF docs available yet (Source #17)
- PyGreSQL is separated from the upstream distribution but
we include it as usual (Source #18)
- Default to compiling libpq and ECPG as fully thread-safe
- 7.4 Origin. See previous spec files for previous history. Adapted
- from Red Hat and PGDG's 7.3.4 RPM, directly descended from
- postgresql-7.3.4-2 as shipped in Fedora Core 1.
* Fri Nov 28 2003 Mihai Ibanescu <misa(a)redhat.com> 2.3.2-7
- rebuilt against newer tcl/tk
* Fri Nov 28 2003 Thomas Woerner <twoerner(a)redhat.com> 4.3-9
- removed rpath
* Fri Nov 28 2003 Thomas Woerner <twoerner(a)redhat.com> 4.1-18
- removed rpath
* Fri Nov 28 2003 Phil Knirsch <pknirsch(a)redhat.com> 2:1.2.3-1
- New bugfix release 1.2.3 came out today, updated again.
* Fri Nov 28 2003 Karsten Hopp <karsten(a)redhat.de> 1.9.1-2
- add patch from -stable CVS
* Fri Nov 28 2003 Karsten Hopp <karsten(a)redhat.de> 1.9.1-1
- update to 1.9.1
- remove obsolete patches
[second attempt, seems the first mail didn't make it to the list]
Full legal name:
Build and QA packages that I use myself that are currently missing and
may (or may not) be usable for others.
None appreciable ;-) Okay, some real info: After 10 years I changed
from OS/2 to Red Hat Linux in the beginning of 2000. Good in
bash-programming, some skills in java. Don't know if I ever find enough
time to really learn C or Python. Lot of interest and know-how in Linux
compatible hardware and how to make it work.
GPG KEYID and fingerprint
pub 1024D/55FB7C4E 2003-11-28 Thorsten Leemhuis <fedora(a)leemhuis.info>
Schl.-Fingerabdruck = BB70 9E3B 9941 B800 006A 54A9 D997 D8BA 55FB
sub 1024g/A587192E 2003-11-28 [verfällt: 2005-11-27]
Well, here it is... a preview anyway. For those of you on the edge of your
seat waiting for an AMD64 version of Fedora Core 1, we present a preview.
ISOs will not be provided for this release, but everything is there for
an install. It is generally version synced with Fedora Core 1 and the
current updates. Please test it out, and send any questions/complaints/bugs
to me if you would before bugzilla, as this is not an official release of
any caliber yet, just a preview. I will forward what is necessary to
* WARNING: This release is a preview, it is not an official Fedora
* Core 1 Release, this is not an official Fedora Core Test Release.
* This release may very well cause damage to your data, your system,
* your pets and loved ones, and most certainly your sleep schedule.
* There is no guarantee of any type on performance, stability, or
* your sanity. Use at your own risk.
Now that the formality is out of the way, the tree is available at:
Enjoy, and let me know. We will be working diligently in the background to
move closer to a stable Fedora Core system for AMD64.
- SATA: Some SATA controllers, notably Sil, do not work
- ACPI is not really functional on NForce 3. Please install and boot with
- Mozilla: both 32-bit and 64-bit mozilla are installed. This is to meet
the dependencies of other installed applications. A system to handle
user selection at boot time has not been implemented yet. Current work
around is to pull /usr/bin/mozilla from the 32-bit package as
/usr/bin/mozilla32 and 64-bit as mozilla64, then just link your
preference to /usr/bin/mozilla. The standard plugins do work in Mozilla
- No documentation (release notes, etc) has been updated, and
fedora-release is still synced with FC1 (IE wrong for preview)
- If you are looking to use this kernel on i686, etc. I have not updated
the config files for anything but x86_64 yet. The kernel should build
and work fine, but the SRPM will probably complain when 'make oldconfig'
asks for input.
I am sure I am leaving something out, it is late, and I am eternally short
on sleep, let me know what else you find.
Justin M. Forbes