debuginfo
by David Kewley
Some debuginfo rpms have snuck into updates-released. Sneaky
sneakers...
Speaking of debuginfo, is the idea of keeping them *out* of repos that
if you want to debug, then you need to rebuild & install your own-built
debuginfo package?
David
18 years, 9 months
FC4 and G5
by Zoltan Kota
Hi,
Is there anybody using FC4 final on an iMac G5?
I have problems that I didn't have with test3 (as much as I can say after
a short test period).
I could install FC4, but within some
minutes of using, the system doesn't respond and I get the following
messages on the console:
ata1: command 0x35 timeout, stat 0x50 host_stat 0x4
hda: lost interrupt
mpic_enable_irq timeout
ata1: command 0x35 timeout, stat 0x50 host_stat 0x4
Moreover, pushing the TAB button (tab completion in shell at the console)
results in a scrolling screen and the above error at once.
What's this?
Zoltan
18 years, 9 months
Enable Default CONFIG_IP6_NF_QUEUE?
by Jeff Weston
Userspace queueuing of IPv6 packets (CONFIG_IP6_NF_QUEUE kernel option) has been significantly improved in the 2.6.12 kernel.
Userspace queueing is already enabled by default for IPv4 packets, but it does not seem to be enabled for IPv6 packets even in the
FC4 release kernel.
Would it be possible for future Fedora updates to enable CONFIG_IP6_NF_QUEUE (as a module) by default? I doubt that enabling it as
a module would cause any problems, and it would be extremely useful in cases where IPv6 packets need to be modified before being
sent out an interface. We would be using it in the immediate future for multicast forwarding/flooding in Mobile Ad-Hoc Networks
(MANETs), though it has potential applications in Mobile IPv6 and elsewhere.
Thanks for your consideration.
-Jeff Weston
18 years, 9 months
Re: FC4 kernel performance
by Malita, Florin
On Tue, 2005-06-21 at 11:54 -0400, Colin Walters wrote:
> Ah, but is Unixbench a relevant benchmark for a desktop?
Probably not as relevant as for servers, but I'm sure the kernel does
make a difference. I would assume I/O performance affects application
startup time a whole lot more than process creation.
On a server though, that 62% (taken with a grain of salt until I get the
2.6.12-rc6 results) is really worrying me.
> I'm skeptical that (apart from preemption) having e.g. a "desktop"
> kernel would be useful.
How about the kernel debug options - shouldn't that be a special-purpose
kernel?
18 years, 9 months
Re: FC4 kernel performance
by Malita, Florin
On Tue, 2005-06-21 at 11:47 -0400, Dave Jones wrote:
> On Tue, Jun 21, 2005 at 11:43:58AM -0400, Malita, Florin wrote:
> > I've recently benchmarked the FC4 kernel (2.6.11-1.1369_FC4)
> against
> > vanilla 2.6.11.12 using UnixBench.
>
> Which is an apples to oranges comparison. 2.6.11-1.1369_FC4 is
> actually based on 2.6.12rc6. I'll be interested in seeing results
> rerun against this kernel.
Will do.
Is the FC tree public? I was tinkering with the idea of
automated/nightly kernel performance tests.
18 years, 9 months
A map of Fedora packages and dependencies
by Tarjei Knapstad
[Disclaimer: This is probably going to end up as a blurb about
dependencies - bear with me :)]
Firstly, I was just perusing through the FC4 package description list
and was reminded of something I've thought of many times:
Would it be possible to map out the package dependencies in Fedora by
doing rpm queries and forming a directed graph of the results? (by using
graphviz or something like that)
I think such a package dependency map would be quite useful to get an
overview of the distribution, both for developers and users. "End nodes"
in the graph could help identify candidates that could be moved to
extras for instance (you might even tag vertices with the package size).
>From the users perspective it would get a lot easier to get an overview
of the impact of removing a package, or if you want to rebuild a
homebrew package for some reason ("I'd really like to just rebuild this
package myself to the bleeding edge, but what else would it impact and
maybe even break, if anything?").
Secondly I was thinking of a "DependencyLint" kind of tool - a small
Python script maybe that would go through every installed RPM on a
system and check if the requirements in the .spec files are correct
and/or sufficient. Here I'm not quite sure about what the rules are for
Fedora however - should absolutely everything that's needed to build a
package be in BuildRequires:? Should every package that has C code
require glibc? Or is it a goal to be as lax as possible to prevent
(hopefully) unnecessary build problems?
In any case I was thinking of maybe using a combination of RPM querying
and ldd (and possibly other tools). An example (from a FC1 box):
1. Get a list of all the RPM's on the system.
2. For each package, list the files and grep the bin/ and lib/ stuff.
For instance 'rpm -ql mutt|grep bin' lists /usr/bin/mutt as one of it's
files.
3. Check which dynamic libraries these binaries and libraries depend on
using 'ldd'. 'ldd /usr/bin/mutt' lists /usr/lib/libz.so.1 among others.
4. For each of these, check which package provides them.
'rpm -q --whatprovides /usr/lib/libz.so.1' returns 'zlib-1.2.0.7-2'
5. Finally check that the original package depends on zlib. (This step
I'm not sure what's the best way to query.) 'rpm -q --requires mutt'
does not list zlib and 'rpm -e --test zlib' does not list mutt so we've
actually found a missing dependency here.
These are just my first thoughts to maybe encourage some discussion. I'm
sure this could be done in a better way and could probably also be
expanded to do other stuff (I'm not very RPM-savvy). If you run a live
testinstall of rawhide for instance such a tool could be run
automatically after each update to detect any packaging problems wrt.
dependencies.
Thoughts?
Regards,
--
Tarjei
18 years, 9 months
rawhide report: 20050617 changes
by Build System
New package adaptx
AdaptX
New package bsh
Lightweight Scripting for Java
New package carol
CAROL: Common Architecture for RMI ObjectWeb Layer
New package castor
An open source data binding framework for Java
New package gif89encoder
Java class library for encoding GIFs
New package gnu.regexp
Java NFA regular expression engine implementation
New package howl-logger
High speed ObjectWeb logger
New package hsqldb
Hsqldb Database Engine
New package jakarta-commons-cli
Jakarta Commons CLI, a Command Line Interface for Java
New package jgroups
Toolkit for reliable multicast communication.
New package jonathan-core
Distributed Object Platform (DOP) written entirely in Java
New package jonathan-jeremie
Distributed Object Platform (DOP) written entirely in Java
New package joram
JORAM: Java (TM) Open Reliable Asynchronous Messaging
New package jotm
JOTM : A Java Open Transaction Manager
New package monolog
API for monitoring and logging
New package nanoxml
NanoXML is a small XML parser for Java
New package objectweb-anttask
ObjectWeb Ant task
New package oldkilim
A generic configuration framework for Java
New package p6spy
Database statement interceptor for Java
New package wsdl4j
Web Services Description Language Toolkit for Java
Updated Packages:
NetworkManager-0.4-31.cvs20050616
---------------------------------
* Thu Jun 16 2005 Dan Williams <dcbw(a)redhat.com> - 0.4-31.cvs20050616
- Update to latest CVS
o Clean up wording in Wireless Network Discovery menu
o Robert Love's applet beautify patch
alsa-utils-1.0.9rf-2
--------------------
* Thu Jun 16 2005 Martin Stransky <stransky(a)redhat.com> 1.0.9-2
- New upstream version
* Mon May 30 2005 Martin Stransky <stransky(a)redhat.com> 1.0.9-1
- New upstream version.
- moved alsacard utility from alsa-lib to alsa-tools
audit-0.9.7-1
-------------
* Thu Jun 16 2005 Steve Grubb <sgrubb(a)redhat.com> 0.9.7-1
- fixed bug in send_user_message which errored on pam logins
- Change nanosleeps over to select loops
- Change the 'e' option to auditctl -p to 'x'
* Thu Jun 16 2005 Steve Grubb <sgrubb(a)redhat.com> 0.9.6-1
- fix bug in incremental flush where is wrongly reported an error
- ausearch should not do uid check for -if option
- adjust ipc interpretation to not use ipc.h
coreutils-5.2.1-51
------------------
* Fri Jun 17 2005 Tim Waugh <twaugh(a)redhat.com> 5.2.1-51
- Use upstream hostid fix.
* Thu Jun 16 2005 Tim Waugh <twaugh(a)redhat.com> 5.2.1-50
- Don't display the sign-extended part of the host id (bug #160078).
* Tue May 31 2005 Dan Walsh <dwalsh(a)redhat.com> 5.2.1-49
- Eliminate bogus "can not preserve context" message when moving files.
cryptsetup-luks-1.0-2
---------------------
* Thu Jun 16 2005 Bill Nottingham <notting(a)redhat.com> 1.0-2
- add patch for 32/64 bit compatibility (#160445, <redhat(a)paukstadt.de>)
cups-1:1.1.23-16
----------------
* Thu Jun 16 2005 Tim Waugh <twaugh(a)redhat.com> 1:1.1.23-16
- Make DeletePrinterFromClass faster (bug #160620).
eclipse-1:3.1.0_fc-0.M7.9
-------------------------
* Tue Jun 14 2005 Andrew Overholt <overholt(a)redhat.com> 3.1.0_fc-0.M7.9
- Add tomcat5 patch and symlinks.
* Thu May 26 2005 Andrew Overholt <overholt(a)redhat.com> 3.1.0_fc-0.M7.8
- Fix ant jar removal (gbenson).
* Wed May 25 2005 Andrew Overholt <overholt(a)redhat.com> 3.1.0_fc-0.M7.7
- Fix ecj symlink in /usr/share/java (rh#158734).
gcc-4.0.0-12
------------
* Thu Jun 16 2005 Jakub Jelinek <jakub(a)redhat.com> 4.0.0-12
- update from CVS
- PRs fortran/22038, libfortran/20930, libfortran/21950, rtl-opt/21528,
target/20301, target/21889, tree-opt/19768, tree-optimization/21171,
tree-optimization/21847, tree-optimization/22043
- further fixes for Fortran FORALL, also use less temporary memory for masks
- make libltdl aware of */lib64 paths (#156005)
- cast of vector to integral type fix (PR middle-end/21850)
- libstdc++.so symbol versioning fixes (Benjamin Kosnik)
iiimf-1:12.2-5
--------------
* Fri Jun 10 2005 Akira TAGOH <tagoh(a)redhat.com> - 1:12.2-5
- Fixed the compiler warnings. (#159392)
- xiiimp-xft.patch: updated to fix use of uninitialized variable.
- EIMIL-fix-uninitialized-value.patch: likewise.
* Mon Jun 06 2005 Akira TAGOH <tagoh(a)redhat.com>
- iiimgcf-fix-memory-leak-r2660.patch: fixed a memory leak.
- xiiimp-fix-infinite-loop-property-notify-event-r2661.patch: fixed
the infinite loop of PropertyNotify event that it causes eating more CPUs
and memory at iiimd. (#158254)
* Thu Jun 02 2005 Akira TAGOH <tagoh(a)redhat.com>
- added a simple wrapper for iiimd to avoid iiimd running as root
when it needs to be run at the terminal for debugging purpose.
- ensure the directory owner of /var/run/iiim is iiimd.
kdemultimedia-6:3.4.1-0.fc4.1
-----------------------------
* Thu Jun 16 2005 Than Ngo <than(a)redhat.com> 3.4.1-0.fc4.1
- 3.4.1
* Tue May 03 2005 Than Ngo <than(a)redhat.com> 6:3.4.0-3
- fix kde-multimedia-music.menu
kdesdk-3.4.1-0.fc4.1
--------------------
* Thu Jun 16 2005 Than Ngo <than(a)redhat.com> 3.4.1-0.fc4.1
- 3.4.1
mod_auth_mysql-1:2.9.0-1
------------------------
* Thu Jun 16 2005 Joe Orton <jorton(a)redhat.com> 1:2.9.0-1
- update to 2.9.0 (#160239)
netatalk-4:2.0.3-1
------------------
* Thu Jun 16 2005 Jason Vas Dias <jvdias(a)redhat.com>
- Upgrade to upstream version 2.0.3
- fix bug 160486: use netatalk's initscript
redhat-rpm-config-8.0.36-1
--------------------------
* Thu Jun 16 2005 Elliot Lee <sopwith(a)redhat.com> - 8.0.36-1
- Fix the fix
* Wed Apr 06 2005 Elliot Lee <sopwith(a)redhat.com> - 8.0.35-1
- Fix #129025 (enable python byte compilation)
ruby-1.8.2-8
------------
* Thu Jun 16 2005 Akira TAGOH <tagoh(a)redhat.com> - 1.8.2-8
- ruby-1.8.2-tcltk-multilib.patch: applied to get tcltklib.so built. (#160194)
selinux-policy-strict-1.23.18-11
--------------------------------
* Thu Jun 16 2005 Dan Walsh <dwalsh(a)redhat.com> 1.23.18-11
- Fix NetworkManager dhcpd communications
- Fix hotplug
* Thu Jun 16 2005 Dan Walsh <dwalsh(a)redhat.com> 1.23.18-9
- Update Ivan trusted/untrusted patch
- add texrel_shlib_t to targeted
selinux-policy-targeted-1.23.18-11
----------------------------------
* Thu Jun 16 2005 Dan Walsh <dwalsh(a)redhat.com> 1.23.18-11
- Fix NetworkManager dhcpd communications
- Fix hotplug
* Thu Jun 16 2005 Dan Walsh <dwalsh(a)redhat.com> 1.23.18-9
- Update Ivan trusted/untrusted patch
- add texrel_shlib_t to targeted
* Wed Jun 15 2005 Dan Walsh <dwalsh(a)redhat.com> 1.23.18-7
- Fixed for new cups domain hplip
shared-mime-info-0.16-4
-----------------------
* Fri Jun 17 2005 David Zeuthen <davidz(a)redhat.com> - 0.16-4
- Add MIME-types for .pcf Cisco VPN settings files (fdo #3560)
system-config-printer-0.6.134-1
-------------------------------
* Thu Jun 16 2005 Tim Waugh <twaugh(a)redhat.com>
- Updated file manifest to include glob patterns for python byte compilation.
- 0.6.134:
- No code changes. Rebuilt for python byte compilation.
* Thu Jun 16 2005 Tim Waugh <twaugh(a)redhat.com> 0.6.133-1
- 0.6.133:
- No code changes. Rebuilt for python byte compilation.
util-linux-2.12p-10
-------------------
* Thu Jun 16 2005 Karel Zak <kzak(a)redhat.com> 2.12p-10
- fix #157656 – CRM 546998: Possible bug in vipw, changes permissions of /etc/shadow and /etc/gshadow
- fix #159339 - util-linux updates for new audit system (pam_loginuid.so added to util-linux-selinux.pamd)
- fix #159418 - sfdisk unusable - crashes immediately on invocation
- fix #157674 – sync option on VFAT mount destroys flash drives
- fix .spec file /usr/sbin/{hwclock,clock} symlinks
xorg-x11-6.8.2-38
-----------------
* Thu Jun 09 2005 Mike A. Harris <mharris(a)redhat.com> 6.8.2-38
- Removed unused legacy with_new_savage_driver macro and conditional
spec file code.
- Added xorg-x11-6.8.2-ati-radeon-7000-disable-dri.patch to disable DRI on
Radeon 7000/VE hardware to test patch in rawhide prior to inclusion in
RHEL4U2. (#150174)
* Mon Jun 06 2005 Mike A. Harris <mharris(a)redhat.com>
- Removed with_libs_data macro as it is no longer useful.
- Updated "Obsoletes: xorg-x11-libs-data" line to remove versioning
* Mon May 30 2005 Mike A. Harris <mharris(a)redhat.com> 6.8.2-37
- Implemented xorg-x11-6.8.2-redhat-kt.patch new kernel tainting diagnostics
patch to aide in troubleshooting reported issues.
- Removed older redhat-custom patch as the kt patch above replaces it now.
- s/XFree86CustomVersion/XorgCustomVersion/ in host.def
- Build for FC5 development.
18 years, 9 months
stateless-linux problems.
by Carl-Johan Kjellander
Hi. I'm playing around with stateless-linux which really suits my Company's
needs, and it works ok but we've encountered some problems. All machines,
both server and diskless clients run FC3.
1. The screen resolution on the diskless machines is 640x480 and not
1400x1050 as we set on the protosystem. Why is that and how can we change
it?
2. When logging into Gnome we get the old xkb error:
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=120858
How can we avoid that?
3. SELinux spews tons of messages at startup, not really a bug but
annoying.
4. When trying to reboot the diskless clients the machines get stuck
in an infinite loop of RPC-errors, 111 if I remember correctly but
I have to recheck.
And where should I file bugs in the stateless-linux scripts?
/cjk
--
begin 644 carljohan_at_kjellander_dot_com.gif
Y1TE&.#=A(0`F`(```````/___RP`````(0`F```"@XR/!\N<#U.;+MI`<[U(>\!UGQ9BGT%>'D2I
Y*=NX,2@OUF2&<827ILW;^822C>\7!!Z1,!K'B5(6H<SH-"E*TJ3%*/>QI6:7"A>Y?):D2^*U@NCV
R<MOQ=]V(B6>LZYD-_T1U<@3W]A4(^$-W4]A#V")W6#.R"$;IR'@).46BN7$9>5D``#L`
18 years, 9 months