Running with the brakes on ...
by Paul Michael Reilly
During the past few weeks, my Rawhide system (A31p Thinkpad Laptop, 1G
RAM, 2GHz CPU, running KDE desktop) has been running noticably slower
(starts out OK but gets worse over time), particularly with
Thunderbird, Firefox and other GUI tools but not so with Emacs, FWIW.
I've been scratching my head trying to figure out how to characterize
the slowdown in order to say something useful and file a bug report if
appropriate but until now about all I could say was "... boy this
system seems slower ... maybe if I do an update it will cure whatever
ails it ..." only to find out that no magic bullet was available from
yum yet. But a few minutes ago, I started playing with multiple X
sessions via the "Switch User" facility. I created a root session on
Ctrl-Alt-F8 running Gnome. Then I set about to enable sound to run
for multiple Users running both KDE and Gnome to run the Soundcard
Detection tool and configuring /etc/profile to set the permissions for
multiple simultaneous sound car users. I discovered, much to my
surprise that the F8 (GNOME) session works perfectly normally while
the F7 session (KDE). Which is not to suggest that I think this is a
desktop selection issue by any stretch. Now I think I have enough
information to solicit suggestions that are more meaningful than the
terse
"man oprofile"
suggestion that I received a few days ago towards getting to the root
of the issue. Even better, it would be good to know if anyone else
running a remotely similar environment (rawhide, KDE, thunderbird,
firefox) is experiencing similar performance slowdowns with the GUI
tools.
When running top, I see that X is consuming 40% of memory which is not
surprising since I am running two X sessions with 3200x1200 (dual
head, radeon, open source) along with long running firefox,
thunderbird and VNC (also 3200x1200) apps). But when I bring up the
Soundcard Detection tool under KDE top shows Xorg is consistently
grabbing 93% of the CPU even when all I am doing is typing this
message. That would certainly explain a lot of the lag in response to
mouse clicks from the app. :-) Switching to Gnome and running top
there shows varying, but high (60%ish) CPU use with the Soundcard
Detection tool still running in both sessions. But with Gnome,
response to mousee clicks is fine, even with the high X CPU use.
So I'm still scratching my head but something about the KDE
environment is causing a performance problem for my system.
-pmr
17 years, 4 months
yum update can commit suicide
by Florin Andrei
Yesterday I was trying to update my FC6 laptop. Quite a few updates,
because this system was not used that much recently.
I did not use the updates applet, because I was on a restrictive
network: first establish a VPN connection, then you can use a proxy
through VPN to browse the Internet.
Since this is not my typical location, the proxy was not set in the
desktop configuration.
So I had to run yum manually in a terminal window. First do an "export
http_proxy=..." then run yum.
Halfway through the update, X was killed by some script in one of the
update packages. Of course, that also killed yum, mid-transaction. :-(
Init restarted X, of course, but now I was staring at the gdm login screen.
This is pretty rude, not to mention that killing yum in the middle of a
transaction can leave the system in a messy state.
I'm not sure what needs to be fixed (don't kill X, or make yum or rpm
smarter and don't kill X if the session is being run from within X, or
something like that), I don't even know which package caused that, but
my take is that this is something that needs to be fixed.
--
Florin Andrei
http://florin.myip.org/
17 years, 4 months
Static linking considered harmful
by Jakub Jelinek
Hi!
We should more proactively discourage static linking in FC7+, for
reasons for that see
http://people.redhat.com/drepper/no_static_linking.html
Removing libc.a would be most effective, but I'm afraid we still need
a handful of statically linked binaries for boot time initialization and
system recovery utilities.
So, I think it would be best if we could analyze what in FC7/FE7
is linked statically, why, if it really needs to be linked that way
and what *.a libraries does it link in and kill all other *.a libraries
(unless the library is only in *.a form, examples libbfd.a,
libc_nonshared.a, libpthread_nonshared.a, libsupc++.a, libgfortranbegin.a)
and kill all other libraries.
E.g. ideally we'd drop libpthread.a, librt.a, libstdc++.a, libgfortran.a,
libboost*.a, all GUI libs that have also *.so libs, etc.
Thoughts?
Jakub
17 years, 4 months
Blowfish encryption for local passwords
by Thomas M Steenholdt
Hi guys,
Is there a reason why fedora does not support blowfish (at least through
the various included tools, system-config-authentication etc.) for
password encryption?
From my understanding, Blowfish provides encryption far superior to
even MD5 and there should no license problems.
Even though MD5 might seem hard-enough-to-crack, why would we stop there?
Also, it seems like supporting blowfish would not be very hard to
implement in fedora, so why don't we?
(and the unavoidable:) Other linux distros have Blowfish encryption for
passwords ;-)
Thanks
/Thomas
17 years, 4 months
tcp_wrappers
by Thomas Woerner
Hello,
tcp_wrappers has two new sub packages: tcp_wrappers-devel and
tcp_wrappers-libs.
All build requirements for tcp_wrappers have to get changed to
tcp_wrappers-devel.
Thanks,
Thomas
--
Thomas Woerner
Software Engineer Phone: +49-711-96437-310
Red Hat GmbH Fax : +49-711-96437-111
Hauptstaetterstr. 58 Email: Thomas Woerner <twoerner(a)redhat.com>
D-70178 Stuttgart Web : http://www.redhat.de/
17 years, 4 months
rawhide report: 20061204 changes
by Build System
Updated Packages:
binutils-2.17.50.0.8-1
----------------------
* Sun Dec 03 2006 Jakub Jelinek <jakub(a)redhat.com> 2.17.50.0.8-1
- update to 2.17.50.0.8
- initialize frch_cfi_data (BZ#3607)
db4-4.5.20-4.fc7
----------------
* Mon Dec 04 2006 Jindrich Novy <jnovy(a)redhat.com> 4.5.20-4
- apply upstream patches for 4.5.20
(Java API <-> core API related fixes)
Broken deps for i386
----------------------------------------------------------
perl-RPM-Specfile - 1.51-1.fc7.noarch requires perl(YAML)
Broken deps for ppc64
----------------------------------------------------------
perl-RPM-Specfile - 1.51-1.fc7.noarch requires perl(YAML)
Broken deps for x86_64
----------------------------------------------------------
perl-RPM-Specfile - 1.51-1.fc7.noarch requires perl(YAML)
Broken deps for s390
----------------------------------------------------------
perl-RPM-Specfile - 1.51-1.fc7.noarch requires perl(YAML)
systemtap - 0.5.10-1.fc7.s390 requires kernel >= 0:2.6.9-11
systemtap-runtime - 0.5.10-1.fc7.s390 requires kernel >= 0:2.6.9-11
Broken deps for ppc
----------------------------------------------------------
perl-RPM-Specfile - 1.51-1.fc7.noarch requires perl(YAML)
Broken deps for ia64
----------------------------------------------------------
perl-RPM-Specfile - 1.51-1.fc7.noarch requires perl(YAML)
Broken deps for s390x
----------------------------------------------------------
perl-RPM-Specfile - 1.51-1.fc7.noarch requires perl(YAML)
17 years, 4 months
Hibernate on FC-6 (partly) fails
by Gianluca Sforna
Sorry for the cross posting (I can't really decide which list was better...)
I'm seeing this regression (compared to FC-5, where it used to work
fine) on the hibernate function with this laptop, an ASUS M6Ne
(Centrino 1.7Ghz, 512Mb RAM, ATI mobility radeon 9700, _not_ using any
proprietary kernel modules)
Basically, the hibernate procedure go as far as the "shutting down
hda" message, then fails to turn off power.
I am forced to turn it off manually. Otherwise the operation seems working.
Anyone else with similar symptoms?
Moreover, I saw from time to time a screen of kernel error messages (
a stack trace ?) on the consule during wake up operation, but I am not
sure if this was generated during hibernation.
I will try to find a pattern for this and eventually post agina more details.
Cheers
Gianluca
17 years, 4 months
rawhide report: 20061203 changes
by Build System
Updated Packages:
openoffice.org-1:2.1.0-6.1
--------------------------
* Fri Dec 01 2006 Caolan McNamara <caolanm(a)redhat.com> - 1:2.1.0-6.1
- next version
- Resolves: rhbz#217813 openoffice.org-2.1.0.ooo72129.vcl.fontglyphindex.patch
- relocate dictionaries
perl-Carp-Clan-5.8-1.fc7
------------------------
* Sat Dec 02 2006 Robin Norwood <rnorwood(a)redhat.com> - 5.8-1
- New version
perl-DBD-MySQL-3.0008-1.fc7
---------------------------
* Sat Dec 02 2006 Robin Norwood <rnorwood(a)redhat.com> - 3.0008-1
- New version from CPAN: 3.0008
perl-DBI-1.53-1.fc7
-------------------
* Sat Dec 02 2006 Robin Norwood <rnorwood(a)redhat.com> - 1.53-1
- Upgrade to latest CPAN version: 1.53
perl-Devel-Symdump-2.0604-1.fc7
-------------------------------
* Sat Dec 02 2006 Robin Norwood <rnorwood(a)redhat.com - 2.0604-1
- Upgrade to latest CPAN version: 2.0604
perl-IO-Socket-SSL-1.02-1.fc7
-----------------------------
* Sat Dec 02 2006 Robin Norwood <rnorwood(a)redhat.com> - 1.02-1
- Upgrade to latest CPAN version: 1.02
perl-PDL-2.4.3-1.fc7
--------------------
* Sat Dec 02 2006 Robin Norwood <rnorwood(a)redhat.com> - 2.4.3-1
- Latest version from CPAN: 2.4.3
perl-RPM-Specfile-1.51-1.fc7
----------------------------
* Sat Dec 02 2006 Robin Norwood <rnorwood(a)redhat.com> - 1.51-1
- Build latest version from CPAN: 1.51
perl-XML-Simple-2.16-1..fc7
---------------------------
* Sat Dec 02 2006 Robin Norwood <rnorwood(a)redhat.com> - 2.16-1
- Upgrade to latest CPAN version: 2.16
Broken deps for ppc64
----------------------------------------------------------
perl-RPM-Specfile - 1.51-1.fc7.noarch requires perl(YAML)
Broken deps for i386
----------------------------------------------------------
perl-RPM-Specfile - 1.51-1.fc7.noarch requires perl(YAML)
Broken deps for s390
----------------------------------------------------------
perl-RPM-Specfile - 1.51-1.fc7.noarch requires perl(YAML)
systemtap - 0.5.10-1.fc7.s390 requires kernel >= 0:2.6.9-11
systemtap-runtime - 0.5.10-1.fc7.s390 requires kernel >= 0:2.6.9-11
Broken deps for ia64
----------------------------------------------------------
perl-RPM-Specfile - 1.51-1.fc7.noarch requires perl(YAML)
Broken deps for x86_64
----------------------------------------------------------
perl-RPM-Specfile - 1.51-1.fc7.noarch requires perl(YAML)
Broken deps for ppc
----------------------------------------------------------
perl-RPM-Specfile - 1.51-1.fc7.noarch requires perl(YAML)
Broken deps for s390x
----------------------------------------------------------
perl-RPM-Specfile - 1.51-1.fc7.noarch requires perl(YAML)
17 years, 4 months