Fedora Weekly News Issue 100

Thomas Chung tchung at fedoraproject.org
Mon Aug 13 08:33:33 UTC 2007

= Fedora Weekly News Issue 100 =

Welcome to Fedora Weekly News Issue 100 for the week of August 6th,
2007. http://fedoraproject.org/wiki/FWN/Issue100

In this week, we have announcements on Virtual FudCon8, Announcing
Fedora 8 Test 1.

In Ask Fedora, we have a few good questions on Intel IP2200 Wireless
In Fedora 7, Distribution Upgrades And Peripherals, Yum Reverse
Dependency Removal.

In Daily Package, we have a few good reviews on Qcad - Simple 2D CAD
program, Gscan2pdf - Frontend for scanning utilities, Xephyr - New
nested X server and Really Slick Screensavers

To celebrate our 100th issue, one lucky winner will receive "Fedora 7
Bible" by Christopher Negus. See Extras Extras section for more

To join or give us your feedback, please visit

   1. Announcements
         1. Virtual FudCon8
         2. Announcing Fedora 8 Test 1 (7.90)
   2. Ask Fedora
         1. Intel IP2200 Wireless In Fedora 7
         2. Distribution Upgrades And Peripherals
         3. Yum Reverse Dependency Removal
   3. Planet Fedora
         1. Fedora 8 Artwork: We Need Your Help!
         2. Smolt passed 100k mark
         3. FOSS.IN/2007 Announced
         4. To Infinity and Beyond
         5. OSCON Wrap-Up
   4. Daily Package
         1. New: Fedora Daily Package Weekly Video Summary
         2. Qcad - Simple 2D CAD program
         3. Gscan2pdf - Frontend for scanning utilities
         4. Wednesday Why: Fast User Switching
         5. Xephyr - New nested X server
         6. Really Slick Screensavers
   5. Marketing
         1. Red Hat And Its Fedora 8 Friends
         2. LinuxLinks Fedora 7 Review
         3. Fedora 7: Community remix
         4. Fedora 7 Chosen as LiveContent CD Platform
   6. Developments
         1. Xfs Going Away. Minor Bug In KDE Control Center
         2. RPM Roadmap Winds On
         3. Kmods: Does Fedora Intentionally Break 3rd Party Firewire
Drivers? No.
         4. New "finddebug-info.sh". Don't Run "ld" Directly In %build
         5. Enabling Compiz By Default
         6. Changes TO Cvsadmin Requests And ACLs
         7. Reducing Power Usage Of Fedora
         8. Disabling Atime
         9. ATIXpress200M And R500 Support For Fedora8
        10. Layering An IDS on Linux: Only Use Abort() If You Intend A Core Dump
   7. Maintainers
         1. packages.fedoraproject.org
   8. Infrastructure
         1. Meetings and Voice Conferencing
         2. Using Asterisk
   9. Advisories and Updates
  10. Fedora 7 Security Advisories
  11. Fedora Core 6 Security Advisories
  12. Events and Meetings
         1. Fedora Board Meeting Minutes 2007-MM-DD
         2. Fedora Ambassadors Meeting 2007-08-09
         3. Fedora Documentation Steering Committee 2007-MM-DD
         4. Fedora Engineering Steering Committee Meeting 2007-08-09
         5. Fedora Extra Packages for Enterprise Linux Meeting (Log) 2007-08-08
         6. Fedora Infrastructure Meeting (Log) 2007-08-09
         7. Fedora Packaging Committee Meeting 2007-08-07
         8. Fedora Release Engineering Meeting 2007-08-06
         9. Fedora Translation Project Meeting (Log) 2007-MM-DD
  13. Extras Extras
         1. Fedora 7 Book for FWN 100th Issue

== Announcements ==

In this section, we cover announcements from various projects.

Contributing Writer: ThomasChung

=== Virtual FudCon8 ===

JeffSpaleta announces in fedora-announce-list[1],

"Announcing Fedora 8's Online FUDCon[2].

Come join us over the next few days for our first online community
conference where Fedora developers will be discussing features and
projects that will be impacting Fedora 8 and beyond."

[1] https://www.redhat.com/archives/fedora-announce-list/2007-August/msg00002.html

[2] http://fedoraproject.org/wiki/FUDCon/FUDConF8

=== Announcing Fedora 8 Test 1 (7.90) ===

JesseKeating announces in fedora-announce-list[1],

"Fedora 8 Test one[2] has been loosed upon the world today.  Included in
this release is a "Fedora" installable 'choose your own adventure'
style set of isos and trees for i386, x86_64, and ppc(64).  Also
included are Live images of both the Fedora Desktop and the Fedora KDE
desktop.  These are available for both i686 and x86_64 (x86_64 is DVD
size only).  Remember these can be used on USB media via the
livecd-iso-to-disk utility available in the livecd-tools package."

[1] https://www.redhat.com/archives/fedora-announce-list/2007-August/msg00001.html

[2] http://fedoraproject.org/wiki/Releases/8/Schedule

== Ask Fedora ==

In this section, we answer general questions from Fedora community.
Send your questions to askfedora at fedoraproject.org and Fedora News
Team will bring you answers from the Fedora Developers and
Contributors to selected number of questions every week as part of our
weekly news report. Please indicate if you do not wish your name
and/or email address to be published.


Contributing Writer: RahulSundaram

=== Intel IP2200 Wireless In Fedora 7 ===

'' adrian.mowrey at gmail.com: I have a laptop (Inspiron B130) from Dell,
and it has a wireless card (Intel(R) PRO/Wireless 2200BG Network
Connection). I also have a wireless connection established through a
linksys router. How can I connect my laptop to the Internet by
wireless? If you could explain to me whatever I have to do so that I
can connect to the Internet by wireless, it will be very helpful for
me as a beginner to Linux. Thank you.


Adrian Mowre''

The kernel driver and firmware to support this card is included by
default in Fedora 7. We recommend the Live images for desktop use
which supports seamless wireless networking via Network Manager [1]
which is enabled by default. If you have used the DVD image to install
Fedora 7, you have to enable it manually. Network Manager is a
background service that has both GNOME (nm-applet) and KDE
(knetworkmanager) frontends.

Set the main service to automatically start on boot:

su -c '/sbin/chkconfig --level 345 NetworkManager on'

Set the dispatcher service to automatically start on boot:

su -c '/sbin/chkconfig --level 345 NetworkManagerDispatcher on'

Start the background services:

su -c '/sbin/service NetworkManager start ; /sbin/service
NetworkManagerDispatcher start'

In the next release, the plan [2] is to enable Network Manager by
default and use it system wide.

[1] http://www.gnome.org/projects/NetworkManager/

[2] http://fedoraproject.org/wiki/Releases/8/FeatureList

=== Distribution Upgrades And Peripherals ===

''William Brown <wbrn314 at cox.net>: Why is it when I upgrade to a new
version of the same distro some of my peripherals won't function? I
was using Fedora 6 and upgrade to Fedora 7. I thought an upgrade was
to fix bugs and make things more user friendly. OK, I have to agree
the GUI looks better but the overall
functionaly seems to be the same. If we want to introduce more people
to the Linux community and the Open Software way of computing then the
developers have to concentrate on making the distros more user
friendly instead of more complicated???''

While the upgrades are intended to bring in more functionality and bug
fixes, there tends to be some amount of regressions in newer releases
too unfortunately. If you are having any issues with this release, do
report those in Red Hat Bugzilla. In Fedora 7, the majority of work
was on the infrastructure side merging in Fedora Core and Fedora
Extras, open build system and distribution composing and customization
tools, mirror manager etc but we did bring in a lot of new features
for end users too[1]. Note that Fedora updates lifecycle [2] has been
extended with Fedora 7 onwards to allow the flexibility for end users
to skip alternative releases if necessary and if you believe that the
advantages in the latest release are not compelling enough for you,
you can take advantage of this flexibility while continuing to receive
updates for the older version of Fedora.

We request the community to participate early in the testing process
by using the development branch of Fedora called 'rawhide' or one of
the three test releases, send feedback and report bugs[3]. Thank you
for your support.

[1] http://docs.fedoraproject.org/release-notes/f7/en_US/sn-OverView.html#sn-New-in-Fedora

[2] http://fedoraproject.org/wiki/LifeCycle

[3] http://fedoraproject.org/wiki/BugsAndFeatureRequests

=== Yum Reverse Dependency Removal ===

'' Roberto Vanto <roberto-vanto at tiscali.it>: In my opinion yum should
support reverse dependencies handling. Something like  this:
http://planet.sabayonlinux.org/?p=65 ''

Yum supports reverse dependencies just fine.  Try, yum remove
<packagename> and yum will automatically list all the packages that
depend on it and prompt for removing all of them. One of the requested
enhancements to this is to remove packages that has been installed as
dependencies for the package you are removing if they are only needed
for that particular package and this feature is under development as a
yum plugin [1]. Meanwhile package-cleanup utility which is part of
yum-utils package in Fedora is quite handy for a number of similar

[1] https://lists.dulug.duke.edu/pipermail/yum/2007-June/009933.html

== Planet Fedora ==

In this section, we cover a highlight of Planet Fedora - an
aggregation of blogs from world wide Fedora contributors.


Contributing Writers: ThomasChung

=== Fedora 8 Artwork: We Need Your Help! ===

MairinDuffy reports in her blog[1],

"For the theme artwork[2] (wallpaper, login, banners, etc.) for Fedora
8, the Fedora Art Team is following a '3-round' process for creating
the artwork. We started by accepting theme concept proposals from the
entire Fedora community. Round 1 involved coming up with
concepts/ideas and rough sketches to demonstrate the concept; Round 2
involved coming up with some more polished artwork towards the concept
proposed, including a requirement of at least one wallpaper design and
3 supplementary designs. Round 2 just ended this past Monday, and the
two themes left standing are: (See the blog)"

[1] http://mihmo.livejournal.com/42902.html

[2] http://fedoraproject.org/wiki/Artwork

=== Smolt passed 100k mark ===

RolandWolters reports in his blog[1],

"Smolt passed the 100k mark, a month and ten days after the 50k mark.
While almost all the machines are Fedora machines (since the project
is at the moment fully included only in Fedora) there are also some
dozen OpenSuse machines."

[1] http://liquidat.wordpress.com/2007/08/07/fedora-8-test-1-released-smolt-passed-100k-mark/

[2] http://smolt.fedoraproject.org/

=== FOSS.IN/2007 Announced ===

JamesMorris reports in his blog[1],

"The 2007 FOSS.IN conference has been announced[2], and will take
place in Bengalaru from December 4-8."

"It looks like they're tightening their focus on community development
and depth of talks, as discussed by Andrew Cowie."

[1] http://james-morris.livejournal.com/22258.html

[2] http://foss.in/2007/info/Announcement

=== To Infinity and Beyond ===

JackAboutboul reports out in his blog[1],

"Today, I'm at LinuxWorld in San Francisco and I've just wrapped up a
set of press interviews along with our very good friends from the
Creative Commons. Whats the skinny you ask? I'll tell you, but only
because I like you. Today we have announced[2] (link to press release)
the first step in what is a long term partnership between Fedora and
he Creative Commons. That first step is the CC LiveContent
Distribution (available off the Fedora torrent site), built on Fedora,
available to all."

[1] http://feeds.feedburner.com/~r/MadRhetoric/~3/141378247/to-infinity-and-beyond.html

[2] http://www.redhat.com/about/news/prarchive/2007/creativecommons.html

=== OSCON Wrap-Up ===

JackAboutboul reports out in his blog[1],

"I cannot begin to tell you how many people were so extremely
enthusiastic about our little PS3[2] demo. People were walking up to
us left and right asking us, "Hey is that Fedora on the PS3? That's
wicked Cool!" I know, thats why I thought we should be showing that
off. We even had a couple of IBM developers walk by us and tell us,
"All of the cool innovation happens in Fedora anyway." Thats a direct

[1] http://feeds.feedburner.com/~r/MadRhetoric/~3/141343423/oscon-wrap-up.html

[2] http://fedoraproject.org/wiki/PlayStation

== Daily Package ==

In this section, we recap the packages that have been highlighted as a
Fedora Daily Package.

[1] http://dailypackage.fedorabook.com/

Contributing Writer: ChrisTyler

=== New: Fedora Daily Package Weekly Video Summary ===

We've introduced a weekly screencast[1] of the packages featured as
Fedora Daily Packages. The video is available from Google Video (flash
and mp4/iPod Video formats) as well as OggTheora format. We look
forward to your comments on this experiment.

[1] http://dailypackage.fedorabook.com/index.php?/archives/118-Fedora-Daily-Package-Weekly-Video-Summary.html

=== Qcad - Simple 2D CAD program ===

''Productive Mondays'' highlight a timesaving tool. This Monday[1] we
covered QCad[2]:

"Qcad is a two-dimensional drafting program, available in proprietary
and GPL versions. Fedora includes the GPL version of QCAD, which does
not have scripting or polyline support."

[1] http://dailypackage.fedorabook.com/index.php?/archives/113-Productive-Monday-Qcad-Simple-2D-CAD-program.html

[2] http://www.ribbonsoft.com/qcad.html

=== Gscan2pdf - Frontend for scanning utilities ===

''Artsy Tuesdays'' highlight a graphics, video, or sound application.
This Tuesday[1] Gscan2pdf[2] was featured:

"Gscan2pdf is a one-stop GUI tool for the production of high-quality
PDFs from scanned images, providing a graphical front end for
xscanimage, unpaper, gocr or tesseract, and other tools."

[1] http://dailypackage.fedorabook.com/index.php?/archives/114-Artsy-Tuesday-Gscan2pdf-Frontend-for-scanning-utilities.html

[2] http://gscan2pdf.sourceforge.net/

=== Wednesday Why: Fast User Switching ===

The ''Wednesday Why'' article[1] took a look at the Fedora's Fast User
Switching configuration:

"Fast User Switching enables more than one user to log in to a local
graphical user interface on a computer equipped with a single seat
(monitor/mouse/keyboard) and permits quick switching between logged-in

[1] http://dailypackage.fedorabook.com/index.php?/archives/115-Wednesday-Why-Fast-User-Switching.html

=== Xephyr - New nested X server ===

''GUI Thursdays'' highlight software that enables, provides, enhances,
or effectively uses a GUI interface. This Thursday[1], Xephyr[2] was
discussed and compared to Xnest[3]:

"Xephyr is an alternative to the Xnest nested X server mentioned last
week. Both of these programs provide a nested X display -- drawing
into a window on a parent X server instead of drawing directly onto a
hardware screen -- but they accomplish this in very different ways."

[1] http://dailypackage.fedorabook.com/index.php?/archives/116-GUI-Thursday-Xephyr-New-nested-X-server.html

[2] http://x.org/

[3] http://dailypackage.fedorabook.com/index.php?/archives/110-GUI-Thursday-Xnest-A-nested-X-server.html

=== Really Slick Screensavers ===

''Friday Fun'' highlights fun, interesting, and amusing programs. This
Friday[1], we took a look at the Really Slick Screensaver collection[2]:

"The Really Slick Screensavers (RSS) -- OpenGL screensavers originally
developed on Windows -- have been ported to work with X and integrated
into the GNOME and KDE screensaver systems."

[1] http://dailypackage.fedorabook.com/index.php?/archives/117-Friday-Fun-Really-Slick-Screensavers.html

[2] http://rss-glx.sourceforge.net/

== Marketing ==

In this section, we cover Fedora Marketing Project.


Contributing Writer: ThomasChung

=== Red Hat And Its Fedora 8 Friends ===

RahulSundaram reports in fedora-marketing-list[1],

"When it comes to expanding the core feature set of Fedora, though,
Fedora 8 has a lot to offer. JackAboutboul explained that a new dbus
service launch service will aim to cut down boot time by minimizing
the number of services that start when a machine powers up.

On the identity side, Fedora 8 will include something called freeIPA
(Identity, Policy, Audit), which is intended to be an easy way for
system administrators to install, setup and administer centralized
identity management and authentication."

[1] https://www.redhat.com/archives/fedora-marketing-list/2007-August/msg00121.html

=== LinuxLinks Fedora 7 Review ===

ThomasChung reports in fedora-marketing-list[1],

"Fedora 7 has seen the word "Core" trimmed from it's name. This
reflects just one of several changes internally to the project. With
the first ever inclusion of tools to create your own version of
Fedora, Fedora 7 really reaches out to the community, quite literally.
If you don't like it, then by all means change it to be the way you
want. The Fedora Project has given you the power. To date this is the
most successful release of Fedora in my opinion."

[1] https://www.redhat.com/archives/fedora-marketing-list/2007-August/msg00114.html

=== Fedora 7: Community remix ===

RahulSundaram reports in fedora-marketing-list[1],

"One of the popular features of Fedora 7 is the ability to remix
Fedora and build your own custom version. Now don't get me
wrong–building a new distribution is nice. But what about those who
wish to create their own Fedora-based project? How do you grow a
complete community in an enterprise environment or in the general
public? Thanks to some of the lesser-known features, anyone can use
the exact same tools that make Fedora, well… Fedora."

[1] https://www.redhat.com/archives/fedora-marketing-list/2007-August/msg00113.html

=== Fedora 7 Chosen as LiveContent CD Platform ===

RahulSundaram reports in fedora-marketing-list[1],

"The Fedora 7 Linux distribution has been chosen as the platform for
the Creative Commons LiveContent[2] CD, an initiative to showcase
free, open-source software and Creative Commons-licensed multimedia

"Fedora 7, which is the result of community-based open-source
collaboration, has a new build capacity that allows for the creation
of custom distributions and individual appliances."

[1] https://www.redhat.com/archives/fedora-marketing-list/2007-August/msg00105.html

[2] http://wiki.creativecommons.org/Livecontent

== Developments ==

In this section, we cover the problems/solutions,
people/personalities, and ups/downs of the endless discussions on
Fedora Developments.


Contributing Writer: OisinFeeley

=== Xfs Going Away. Minor Bug In KDE Control Center ===

A report[1] from ArthurPemberton detailed a failure of Xorg to start
after a shutdown.  Arthur had figured out that it was something to do
with the addition of the line ''FontPath "/usr/local/share/fonts"'' to
his Xorg.conf.  He further suspected that this line had been added by
KDE's control center when he tried to add some TTF fonts.

[1] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00291.html

HansdeGoede thanked[2] Arthur for the clarity of the report and
requested that he file a Bugzilla entry so that this problem could be
elevated from mere discussion on @fedora-devel into an actual
trackable, non-losable bug report.

[2] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00293.html

RahulSundaram agreed[3] with Arthur that Xfs (the X font-server) was deprecated
and would be removed in Fedora8 eliminating this type of problem.
Kelly wasn't impressed[4] with this due to past experience with Xorg
being flaky when Xfs was
disabled.  RahulSundaram doubted this was true with more recent
desktops which use fontconfig (e.g. Fedora7) and Kelly admitted[5]
that the problems s/he experienced had been with Debian two years ago
and never occurred with Fedora.  Later[6] Kelly posted some unclear
details which seemed to indicate a problem with Fedora7 in which Xorg
will not start if Xfs is disabled. RahulSundaram seriously doubted[7]
this on the basis of personal experience and asked Kelly to verify.

[3] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00294.html

[4] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00298.html

[5] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00300.html

[6] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00303.html

[7] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00304.html

An interesting addendum[8] was provided by MichaelSchwendt who thought
that Rahul's blanket disavowal of problems like this in Fedora7 wasn't
justified. Michael showed that both "emacs" and "xterm" would complain
about missing fonts if xfs were stopped. In response to
ArthurPemberton's suggestion that Michael add the font paths to his
Xorg.conf Michael wondered why he should have to perform manual edits
if Fedora7 was supposed to "run without xfs out of the box".  Michael
also provided a link[8] to a Bugzilla entry he had opened which
suggests that the Fedora "compiled-in" font path deviates from that
documented in the manpage.

[8] https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=251203

AndyGreen identified the use of "/usr/local" in the FontPath as
suspicious and unusual for a package-mediated modification, which
Arthur agreed with and reminded[9] that he had already fingered "KDE's
font manager" as the possible culprit. Rahul noted that this provided
the component against which Arthur could file a bug, which was duly

[9] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00316.html

[10] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00440.html

=== RPM Roadmap Winds On ===

Ideas for improvements to RPM kept on rolling in (see FWN [1] and FWN
[2] for earlier coverage).  MatthiasClasen raised[3] the idea of
support for language packs or else some way to install a limited set
of locales.  This would make LiveCDs much more useful.

[1] http://fedoraproject.org/wiki/FWN/Issue98#head-90b3708648e89def933055ad00d6cb887b9cde0d

[2] http://fedoraproject.org/wiki/FWN/Issue99#head-197a66d4dbef91d613deea97f40b26097fe794f6

[3] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00306.html

Panu thought[4] that the problem with this was that metadata bloat
would occur and that although the mechanisms to install a limited set
of locales already existed they weren't useful because the only way to
get any skipped languages back was to reinstall completely.

[4] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00364.html

A potential spanner in the works was identified[5] by JonathanDieter
in the form of deltarpms. This will only work if all files are present
from the original rpm (e.g. none excluded on the basis of locale) and
yum would need instead to download the whole updated rpm instead of
the smaller drpm.

[5] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00372.html

Matthias was undeterred[6] by Jonathan and JeremyKatz's worries and
had confidence that the problem of verifying RPM integrity was
solvable. ToshioKuratomi's solution of GPG-signing the rpm metadata
made BillNottingham wince[7] when he considered the metadata bloat in
which it would result.

[6] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00380.html

[7] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00391.html

A request for "scriptlets" was submitted[8] by TillMaas along with
examples of how they might be useful for installing, for example,
.desktop files.

[8] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00522.html

NicolasMailhot added[9] two items to the wish-list: 1) true
strong-dependency versioning; 2) noarch sub-packages for doc and
noarch resources.

[9] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00779.html

=== Kmods: Does Fedora Intentionally Break 3rd Party Firewire Drivers? No. ===

The discussion of kmods is another long-running thread (see FWN#99
"Kmods Clarified"[1]) and TillMaas provided a response[2] to
ThorstenLeemhuis 's request for a specfile which
uses DKMS in a mock-buildable SRPM which compiles modules shippable as
kmod packages and also produces a DKMS file for dynamic kernel-module
generation. Phew!

[1] http://fedoraproject.org/wiki/FWN/Issue99#head-98170522fb95c1c25efc05a5f2f6367656a7f174

[2] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00791.html

A longish thread centered around the negative experience which
LesMikesell had with Firewire ceasing to function mid-version due to
kernel updates.  Les was fairly persistent in explaining his problem
and in return RahulSundaram pointed to the problems of offering
support for 3rd-party drivers given Fedora's development methodology.
EmmanuelSeyman provided an excellent deconstruction[2a] of the
argument that Fedora should be more like RHEL. RahulSundaram also
noted that Les was seeking to change way in which the upstream kernel
worked and that discussing it on @fedora-devel was pointless.
ChristopherBrown didn't like[2b] Rahul's intervention and stated that
he'd opened the topic (of kqemu originally) because @fedora-devel had
a "better ambiance" and had a greater chance of generating an answer.

[2a] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00389.html

[2b] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00325.html

RahulSundaram noted[3] problem with Firewire had affected all
distributions, not just Fedora, and that Red Hat had rewritten the
firewire stack to solve it and it was now merged upstream.
RahulSundaram also disputed Les's apparent suggestion that Fedora was
intentionally breaking third-party proprietary modules.

[3] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00414.html

In a disclaimer[4] from Les, the specter of Fedora intentionally
changing the ABI in order to "encourage users to switch to RHEL" was
raised and rejected, with the caveat that there should be a statement
that although third-party drivers may work at the start of a version
there is no guarantee that they will work with any updates. Les also
suggested that the Centosplus kernel was a good model and that Fedora
should make "untested changes" opt-in. Rahul's reply[5] emphasized
that all changes are opt-in anyway, that they are not untested, that
the Centosplus kernel relies on a huge amount of back-porting which
the Fedora Project can't do due to resource constraints. RahulSundaram
also pointed out that it is completely clear that software outside of
the Fedora Project repositories (for example any proprietary
third-party modules) cannot be supported.

[4] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00423.html

[5] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00424.html

A stronger statement[6] from AlanCox averred that the default
expectation with Linux was that any such third-party drivers might
break, and that RHEL was an anomalous case which resulted from great
pain and hard work to preserve the kernel ABI. In addition Alan
redirected the finger of blame to point at the packager of the
third-party driver as rpm should have flagged the kernel-update as
violating a requires. Les thought[7] that Alan's assumptions and those
of a "typical new fedora user" were not congruent and that the
distinction should be made clear on the Fedora Project pages. A short
discussion between VilleSkyttä and Alan on the subject of when support
had been introduced for requiring a specific kernel version
occurred[8] with Ville apparently demonstrating that it wasn't for as
long as Alan thought.

[6] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00426.html

[7] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00432.html

[8] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00436.html

JefSpaleta[9] and RahulSundaram[10] both thought that Les should run
the updates-testing kernel and report any firewire bugs which he

[9] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00425.html

[10] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00434.html

A response[11] from ParagN to DavidWoodhouse's question about the
"spca5xx" and "gspca" kmods provided some information about them.
David's post also observed that the idea that Fedora would carry
out-of-tree modules so long as they had either "a hope of going
upstream" or "good reason to continue [being] carried forward" was
unsurprising because this was exactly what Fedora had been doing for

[11] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00318.html

Parag's response didn't provide the information which David was
looking for, namely were the modules good enough to be merged upstream
and what needed to be fixed to convince Linus to take them.  Parag
responded[12] that the author wasn't interested in persuading upstream
and questioned whether it was now a hard and fast rule that although a
package was popular and functional it wouldn't be carried in Fedora if
its author were unwilling to upstream it.

[12] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00437.html

=== New "finddebug-info.sh". Don't Run "ld" Directly In %build ===

A certain amount of surprise was expressed[1] by JesseKeating at the
unannounced replacement of 'find-debuginfo.sh' in the latest rpm in
the Koji build-system. Jesse noted the failures of some builds due to
this change and his intention to revert it until there was supporting
documentation to help package maintainers adapt. Jesse pointed to
older information in the wiki about the purpose of these changes,
which is to introduce a unique buildID into stripped objects and their
corresponding '.debug' files.

[1] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00457.html

RolandMcGrath apologized[2] and explained that crossed-wires had
resulted in him being unavailable to shepherd in the change for which
he was entirely responsible.  Roland emphasized that the script should
be a simple drop-in replacement that only affected some corner cases
and that he was available to help anyone with problems.

[2] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00472.html

After JesseKeating pointed to mock failing to build Roland
explained[3] that this was due to a .so was being built improperly and
that the new BuildID support had enabled this to be caught.
JasonTibbitts was sure[4] that this .so issue was going to occur for
other packages and that it should be documented by someone that had a
thorough understanding.  Roland agreed and wondered where that should
be.  A detailed explanation[5] by Roland followed explaining why
''ld'' should not be run directly and instead it should be called
through ''gcc''. TomCallaway was happy[6] because this would fix some
sparc64 issues.

[3] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00475.html

[4] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00477.html

[5] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00481.html

[6] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00488.html

JefSpaleta wondered[7] if local mock testing would help and Roland
responded[8] that while BuildID would work exactly the same he had
already taken care of obvious special cases (although he'd missed a
problem with noarch packages) and he could quickly take care of any
problems which occurred in Fedora development and testing, i.e.

[7] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00479.html

[8] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00484.html

Further discussion with Jef, who favored using local mock builds
before sending to the build server, failed to convince Roland that his
current approach of introducing changes into rawhide and fixing the
resulting breakage would result in mobs bearing pitchforks and
torches. Roland questioned the need for a wordy documentation
preferring to post to @fedora-maintainers.  In response to his
question "Do we really need a dissertation on the subject?"
JesseKeating answered[9] "Nope.  But send a note to
'fedora-devel-announce redhat com' alerting people that this change
will go in, some builds may break when running the debug script, and
that they should contact you for help."

[9] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00512.html

HansdeGoede was hopeful[10] that Roland had also fixed some ugly
cross-building behavior in the old debuginfo.sh, but unfortunately it
appeared not[11] although Roland declared his willingness to help out.

[10] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00486.html

[11] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00491.html

=== Enabling Compiz By Default ===

RahulSundaram wanted to know[1] if the default enabling of Compiz in
GNOME had been considered.

[1] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00559.html

"Dragoran" thought it would be better to wait until it was possible to
make OpenGL apps play nicely with AIGLX and posted[2] a link to an
informative blog entry by KristianHoegsberg explaining the issues.

[2] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00562.html

A highly negative reaction[3] from AlanCox questioned the maintenance
of Compiz and its problems with Java menus.  Alan also pointed to the
detrimental effects on power usage occasioned by its use.

[3] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00565.html

"Dragoran" replied that it was not Compiz which was broken, but Java, and
provided a simple environment variable fix and wondered what Alan
meant about Compiz being poorly maintained. Alan responded with "the
user view point:  I DON'T CARE"  and argued[4] that compatibility
worked because components didn't enforce strict standards compliance
and hence Fedora needed to either: 1)fix Java; 2)set "Dragoran's"
environment fix as default. "Dragoran" was in favor of #1, thought
that #2 was a hack and pointed[5] to the number of commits in the git
tree as evidence of the active maintenance of Compiz.

[4] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00580.html

[5] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00585.html

A bugzilla was filed by "Dr.Diesel" pointing out[6] that "vino"
(remote desktop utility using Xv) was broken with Compiz.

[6] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00598.html

More skepticism came[7] from JoshBoyer who wanted to know if Compiz
could do basic window-manager tasks such as making a window present
simultaneously on all desktops. "Dragoran" confirmed[8] that it could
and later after JesseKeating had suggested enabling only subtle
features such as drop-shading and fading windows Josh and MikeChambers
agreed[9] that this might be useful as long as the basic WM features
were working.

[7] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00591.html

[8] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00593.html

[9] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00660.html

BillNottingham was another naysayer, advancing[10] the breaking of
other GL applications and Xv applications as reasons against enabling
Compiz as the default. JonNettleton suggested choosing a compositing
window manager which doesn't need OpenGL, but DaveAirlie answered[11]
that this didn't solve Bill's two problems. Dave argued that Compiz
wasn't broken, instead the problem is that drivers and GL are broken
when compositing is enabled and that KristianHoegsberg (krh) and
Dave's work on the Intel drivers was fixing this.  The actual
compositing manager doesn't make any difference due to regressions at
the driver level.

[10] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00682.html

[11] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00742.html

The meat of the problem was exposed further in a thread started when
BernardoInnocenti posted[12] that Ubuntu's "Gutsy Gibbon" was going to
use Compiz-fusion (the merged Beryl/Compiz) by default and was fixing
bugs to this end.  Bernardo thought that AlanCox was wrong to believe
that all compositing effects would be turned off and cited the high
user-satisfaction with OSX's "Quartz".  He also queried Alan's claims
about power inefficiency. DaveAirlie responded[13] that Ubuntu were
not fixing Compiz's technical problems and were benefitting
asymetrically from Red Hat's work. RudolfKastl was keen on encouraging
Dave's work and in response to his queries about getting a working
"nouveau" driver (the FL/OSS nvidia driver) and a driver for r500
(ATI) into rawhide DaveAirlie responded that work on 3D redirection
and on speeding up EXA were the important prerequisites for all
compositing managers. (These projects are done with KristianHoegsberg
and CarlWorth[14] respectively). The chances of a nouveau or r500
driver with these features are very small as only Intel have provided
enough information to allow the necessary fixes to be written.

[12] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00589.html

[13] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00600.html

[14] http://cworth.org/blog/

JonNettleton pointed out[15] the advantages of 2D-compositing
window-managers, such as "xfwm4" which works well with any
EXA-supporting driver and in response KevinKofler drew attention to
KDE4's "Kwin" (which will probably be in Fedora9).

[15] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00606.html

AlanCox explained[16] his claims about power consumption to
IoannisNousias who had advanced the example of OSX.  Alan's response
was that Apple had taken advantage of independent power-management of
the 2D and 3D graphics cores to achieve acceptable battery life
overall. Ioannis didn't accept the argument and cited[17] examples of
comparable battery longevity between an OSX system and an
Xorg+metacity system, indicating that compositing wasn't a big drain.
Ioannis further noted that the concern about power use seemed a little
hypocritical given other Fedora choices.

[16] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00665.html

[17] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00724.html

A slight defence was mounted by KevinKofler who noted[18] that much
work on reducing spurious wakeups was being undertaken (see also
"Reducing Fedora Power Use" in this same issue of FWN#100), especially
under the aegis of the OLPC project. Kevin also mentioned KWin again
which led Ioannis to wonder[19] why compositing was accepted in KWin
and Metacity but not Compiz.

[18] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00800.html

[19] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00822.html

DaveAirlie responded[20] that many applications would break for many
specific drivers and that Intel hardware would probably be an
exception once he and Kristian had made some more progress.  After
this any compositing manager could be chosen and that although Compiz
might be overkill it could be acceptable if it had the same hotkeys as
Metacity. Later Dave confirmed[21] to RahulSundaram that much work
which he and Kristian had in hand needed to be upstreamed.

[20] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00825.html
[21] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00827.html

=== Changes TO Cvsadmin Requests And ACLs ===

ToshioKuratomi posted[1] an announcement of changes which will be
implemented over a couple of weeks. The changes concern the transition
from flat files to a packageDB.  Toshio's post provided essential
information for packagers and an invitation to test the new
functionality provided.

[1] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00546.html

One of the first queries came from ChrisWeyl who wanted to know[2] if
the PackageDB would be updating the old owners.list or if the data
would be otherwise exported in some other format. Toshio responded[3]
that the data was available in JSON format, that owners.list was dead
and gone and that there was a ticket open to create a simply greppable
plain-text list.  RalfCorsepius was irritated[4] that his scripts were
now broken as a result of this.

[2] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00549.html

[3] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00569.html

[4] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00760.html

MartinSourada asked[5] whether "Cvsextras Commits" would now be set to
"yes" if he had already removed "pkg.acl" in order to enable anyone in
cvsextras to access his packages. According[6] to Toshio this depended
on exactly when he had taken this action, with the cutoff point being
the importation of the packagedb.  Toshio suggested that Martin check
this at https://admin.fedoraproject.org/pkgdb/packages/name/[PACKAGENAME]

[5] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00552.html

MichaelSchwendt sought[6] the status of the replacement of owner.list
downloads+parsers and Toshio asked for further specification of the
problem and provided links[7] which list: packages owned by a user;
packages which has the user as a comaintainer or on cclist; packages
which can set bugzilla; packages which can set cvs ACLs.

[6] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00624.html

[7] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00697.html

In response to Toshio Michael clarified that he wanted a Python module
and Toshio opened[8] a ticket for this and Michael further
specified[9] that anonymous access to a package's src.rpm %name,
primary maintainer, co-maintainers and cclist retrieved in a single
flat-file or db would be simplest.

[8] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00753.html

[9] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00765.html

RalfCorsepius also described[10] similar requirements.

[10] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00761.html

=== Reducing Power Usage Of Fedora ===

BillNottingham asked whether we would like longer battery life and
lower server power bills.  He stated[1] that unnecessary CPU wake-ups
were one of the biggest current problems and asked for our help in
tracking down misbehaving applications by installing "powertop" and
reporting the results. (See FWN#88 "PowerTOP Release Opens Up New
Directions In Power Saving"[1a].)

[1] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00544.html

[1a] http://fedoraproject.org/wiki/FWN/Issue88#head-f0b77fa701a85115b677b05bf3693fb5a594340b

A report from JonathanUnderwood about his LCD screen switching on/off
while running powertop was confirmed by ArjanvandeVen to be abnormal
and further details[2] were provided in a bugzilla report. TomLondon
also had this problem.

[2] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00554.html

Trying to run powertop on F7 with the "-d" flag didn't work[3] for
"nodata" who also wondered what to make of three Fedora-specific
features which powertop suggests changing. In response AlanCox
agreed[4] that disabling HAL polling of the cdrom was a good idea as
was enabling AC97 powersave mode by executing ''echo 1 >
/sys/module/snd_ac97_codec/parameters/power_save''. BillNottingham
cleared up the confusion[5] about the "-d" option with the information
that it was only in the F8 version.

[3] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00557.html

[4] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00564.html

[5] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00611.html

RahulSundaram posted[6] his results from several weeks of powertop
which included the suggestion of ''echo 5 >
/proc/sys/vm/laptop_mode''.  "Dragoran" wasn't sure how useful this
was and ArjanvandeVen responded[7] that laptop_mode did not spin down
the disks, but rather clustered IO which was useful not merely with a
battery but in large data centers.

[6] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00558.html

[7] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00579.html

Integration with Smolt was suggested[8] by ChristopherBrown.

[8] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00637.html

RalfCorsepius wondered whether the kernel's failure to boot without
ACPI disabled would class it as a misbehaving application and met
JoshBoyer's request for a bugzilla entry with one he had filed in late
2006 and the information that this problem had started sometime in
FedoraCore5.  "Nodata" wanted[9] to start a tracker bug for "bugs that
have been hanging around for ages, but aren't yet fixed"

[9] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00686.html

DebarshiRay added results which showed a python process taking second
place (at 17% of wakeups) and further interest from BillNottingham
prompted[10] Debarshi to list some candidates.  ColinWalters pointed
to a 10Hz loop in all PyGTK2 programs in Fedora as the source of the
problem[11] and ArjanvandeVen thought there was no need for this on an
NPTL[12] system.

[10] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00701.html

[11] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00704.html

[12] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00757.html

=== Disabling Atime ===

A link to the Linux kernel mailing list (lkml) was posted [1] by
RahulSundaram to draw attention to the idea of disabling "atime" as a
mounting option for filesystems in order to improve performance.
Initial proposals from IngoMolnar were to use "noatime" to simply
avoid the performance penalty imposed by heavy I/O with "atime".
Based on discussion with Linus, AlanCox and others Ingo rapidly
produced a modified version of this proposal which adds "relatime" as
an option. "relatime" would simply not update the atime of the file
unless there had been a modification in the interval since the last

[1] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00561.html

The main practical issues seem to be that mounting filesystems with
"noatime" instead of "atime" on non-CPU-bound systems may provide
serious performance increases.  The downside is that applications
which rely upon examining the atime of inodes (e.g. tmpwatch and mutt
and possibly HSM[1a]) will be broken. "atime" support is also
necessary for strict compliance with POSIX.

[1a] Hierarchical Storage Management is also known as Tiered Storage
and the practice of storing infrequently accessed files on
slower/cheaper storage media (somewhat analogous to memory caches on
CPUs): http://en.wikipedia.org/wiki/Hierarchical_storage_management

Reactions were mixed.  "DrDiesel" thought[2] it would be worth testing
and LesMikesell also was willing[3] to forgo two uses of atime (as an
aid to verifying whether changes were being picked up).  A brief
discussion between Les and JohnReiser ended with a handy tip[4] from

[2] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00566.html

[3] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00572.html

[4] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00617.html

FlorinAndrei and "Dragoran" noted[5] positive experiences on a wide
variety of systems as did AlexandruCiobanu[6].

[5] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00571.html

[6] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00576.html

An alternative suggestion (also on the lkml thread) is to use "relatime".  This
was proposed for Fedora by RudolfKastl and seconded by AlanCox.
Rudolf was prompted by RahulSundaram to expand on his reasons for
preferring "relatime" and explained[7] that merely switching off
flawed features and working around slows down fixes. Cautious
agreement was expressed[8] by EricSandeen who wanted more hard data
provided by the advocates of change.

[7] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00599.html

[8] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00685.html

RalfCorsepius was against the idea of disabling atime by default
because it is a fundamental feature of UNIX filesystems.  JamesHubbard
expressed[9] disagreement based on the evidence that NFS optimization
documents explicitly suggest disabling it and suggesting that most
applications and users would benefit from its absence and the minority
that needed could turn it on. Ralf's response[10] was skeptical and
provided some figures which showed a much slimmer gain from "noatime"
than reported by others.  Ralf also pointed out that James might be
considering only single-user desktop systems at the expense of other

[9] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00654.html

[10] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00667.html

In response[11] to the figures provided by Ralf it was suggested by
JoshBoyer that the task (a heavy compiler run) wasn't that useful and
that it would be better to do a "find / -exec cat {} > /dev/null \;"
both with and without atime.  DougMcClendon was happy to find some
actual numbers[12] and thought that Ralf was underplaying the
significance of the possible 5% advantage.  Doug also noted the
benefit to laptop drives which wouldn't need to be spun up to update
the atime each time a cached file is read.

[11] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00678.html

[12] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00692.html

MartinEbourne reacted[13] to Ralf's post by questioning its rhetoric
and restating that the idea was to provide a sensible default for the
majority of users.  Martin had been running "atime" disabled systems
for over five years with no problems (with the caveat that he didn't
use Mutt) and was surprised it had taken so long for the rest of the
world to come around to this idea. A robust defense[14] of his
position was mounted by Ralf who thought that Martin's logic was a
slippery slope which could lead to the removal of many
useful/essential features which impose some performance loss.

[13] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00744.html

[14] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00764.html

The issue of breaking "tmpwatch" was raised by MattMiller and after
RahulSundaram noted the existence of a patch MilsolavTrmac
wondered[15] what the rush was and suggested it would be easiest to
wait for upstream (lkml) to sort out a decision. Rahul pointed out[16]
that the lkml discussion had been redirected to @fedora-devel and that
this was an old discussion and it would be good to make a decision.

[15] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00782.html

[16] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00786.html

LesMikesell suggested[17] a possible strategy for introducing the
change gradually via an /etc/sysconfig file which contains mount
options merge-able with /etc/fstab specified options.  There would be
no default options provided unless "noatime" were selected when
prompted during install. Further discussion of this with BenjaminLewis
(who thought that targetting the defaults to the simple home user case
made more sense) led Les to outline the benefits in terms of following
RedHat-style administration tools and preserving legacy defaults.

[17] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00783.html

=== ATIXpress200M And R500 Support For Fedora8 ===

The 2007-07-08 Rawhide Report[1] excited ZoltanBoszormenyi because[2] it
contained the promise of 3D support for the ATI Xpress 200M.  Zoltan noted
that the current Mesa (6.5.2) locked up with this driver and that Mesa-7.0.1
should be used instead. AlanCox asked Zoltan to file this as a security bug
because it was a local DoS which he promptly did[3].

[1] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00482.html

[2] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00494.html

[3] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00626.html

DaveAirlie promised[4] to look at getting the new Mesa into the tree so that 3D
support could be offered on the chipset and PeterGordon added[5] that this
might also fix a Gnash problem with YouTube.

[4] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00496.html

[5] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00498.html

ManuelWolfshant wondered[6] whether the experimental, Free ATI R500
driver (which he had packaged succesfully for a friend) might be
included in the xorg-drv-ati package. AdamJackson responded[7] that he
would rather include it separately from -ati, but would like to see it
in Fedora8.

[6] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00507.html

[7] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00536.html

=== Layering An IDS on Linux: Only Use Abort() If You Intend A Core Dump ===

SteveGrubb outlined[1] his plans to create a host-based IDS/IPS[2] system in
the Fedora9 cycle. Steve was concerned that his feature in the 2.6.22 kernel
which enables detection of buffer overflows may cause some programs to falsely
trigger the IPS in the future if they call abort().

[1] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00272.html

[2] http://en.wikipedia.org/wiki/Host-based_intrusion_detection_system

MiloslavTrmac thought[3] that Steve's request flew in the face of tradition
and POSIX and that only repeated abort()s should be taken as indications of a
DoS. Steve argued[4] that Miloslav's objections only applied in a
non-production, debugging scenario.  Steve later expanded[5] on the problem in
the face of Miroslav's continued objections and SimoSorce's suggestion that
SteveGrubb outlined[1] his plans to create a host-based IDS/IPS[2] system in
the Fedora9 cycle. Steve was concerned that his feature in the 2.6.22 kernel
which enables detection of buffer overflows may cause some programs to falsely
trigger the IPS in the future if they call abort().

[1] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00272.html

[2] http://en.wikipedia.org/wiki/Host-based_intrusion_detection_system

MiloslavTrmac thought[3] that Steve's request flew in the face of tradition
and POSIX and that only repeated abort()s should be taken as indications of a
DoS. Steve argued[4] that Miloslav's objections only applied in a
non-production, debugging scenario.  Steve later expanded[5] on the problem in
the face of Miroslav's continued objections and SimoSorce's suggestion that
as abort() was used so extensively it might be better [6] to give up on trying
to stop abort() usage and to build SELinux-like application profiles.

[3] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00276.html

[4] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00278.html

[5] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00282.html

[6] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00310.html

AlanCox and ArjanvandeVen cooked up[7] the idea that although single SIGABRT
terminations might be ignorable it would also be nice to dump core somewhere
secure and then communicate the backtrace to the vendor.  SteveGrubb (SteveG)
confirmed that this would be easy to add as a plugin to the audit event
dispatcher (audispd) [8].

[7] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00305.html

[8] https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00309.html

== Maintainers ==

In this section, we cover Fedora Maintainers, the group of people who
maintain the software packages in Fedora.


Contributing Writer: MichaelLarabel

=== packages.fedoraproject.org ===

Ever visit packages.ubuntu.com or similar web-based deb package
viewers? It sure is nice and has led JohnPye to question why we don't
have a packages.fedoraproject.org[1]. There is the Fedora repobrowser
but right now it's really not a prominent resource. PatriceDumas had
replied with word that there is such an effort taking place with the
new repoview version.

[1] https://www.redhat.com/archives/fedora-maintainers/2007-August/msg00151.html

== Infrastructure ==

In this section, we cover the Fedora Infrastructure Project.


Contributing Writer: JasonMatthewTaylor

=== Meetings and Voice Conferencing ===

JeffOllie and some others have been working on setting up a PBX
service for difference groups as a way to enhance and perhaps enable
others who don't have access to IRC a way to attend meetings.[1] The
Infrastructure group as always is open to any suggestions for

[1] https://www.redhat.com/archives/fedora-infrastructure-list/2007-August/msg00061.html

=== Using Asterisk ===

MikeMcGrath discusses the pros and cons of current meeting
communication and the addition of asterisk.[1]

[1] https://www.redhat.com/archives/fedora-infrastructure-list/2007-August/msg00075.html

== Advisories and Updates ==

In this section, we cover Security Advisories and Package Updates from


Contributing Writer: ThomasChung

== Fedora 7 Security Advisories ==

 * mediawiki-1.9.3-34.0.2.fc7  -
 * openssl-0.9.8b-14.fc7  -
 * moodle-1.8.2-1.fc7  -
 * balsa-2.3.17-2.fc7  -
 * dovecot-1.0.3-14.fc7  -
 * cups-1.2.12-4.fc7  -
 * tetex-3.0-40.1.fc7  -

== Fedora Core 6 Security Advisories ==

 * xorg-x11-xinit-1.0.2-16.fc6 -
 * vixie-cron-4.1-69.fc6 -
 * kernel- -

== Events and Meetings ==

In this section, we cover event reports and meeting summaries from
various projects.

Contributing Writer: ThomasChung

=== Fedora Board Meeting Minutes 2007-MM-DD ===

 * No Report

=== Fedora Ambassadors Meeting 2007-08-09 ===

 * https://www.redhat.com/archives/fedora-ambassadors-list/2007-August/msg00016.html

=== Fedora Documentation Steering Committee 2007-MM-DD ===

 * No Report

=== Fedora Engineering Steering Committee Meeting 2007-08-09 ===

 * https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00556.html

=== Fedora Extra Packages for Enterprise Linux Meeting (Log) 2007-08-08 ===

 * https://www.redhat.com/archives/epel-devel-list/2007-August/msg00062.html

=== Fedora Infrastructure Meeting (Log) 2007-08-09 ===

 * https://www.redhat.com/archives/fedora-infrastructure-list/2007-August/msg00081.html

=== Fedora Packaging Committee Meeting 2007-08-07 ===

 * https://www.redhat.com/archives/fedora-maintainers/2007-August/msg00128.html

=== Fedora Release Engineering Meeting 2007-08-06 ===

 * https://www.redhat.com/archives/fedora-devel-list/2007-August/msg00405.html

=== Fedora Translation Project Meeting (Log) 2007-MM-DD ===

 * No Report

== Extras Extras ==

In this section, we cover any noticeable extras news from various
Linux Projects.

Contributing Writer: ThomasChung

=== Fedora 7 Book for FWN 100th Issue ===

For the latest guidelines and rules, please visit

Thomas Chung

More information about the fedora-announce-list mailing list