I've been lurking this list for quite some time now, waiting for a good
reason to post...
Seems that I found one.
Let me first point out that this is not a rant; I understand that the
Fedora Core is supplied "as is" and that the FC development team isn't
working *for me*. In short, they don't owe me anything.
Plus, being a Linux developer myself, I can appreciate the *perceived*
(in my eyes) lack of interest in dealing with (what-seems-to-be-in-my-
eyes) bugs, when the FC5 development is at full swing.
I'm not trying to start a flame war; I am trying to understand the
direction in which the Fedora Core foundation is heading, and make my
own decisions as a result. (Roll back machines to FC3, keep others at
FC2/3, switch to another distro, etc)
Three months ago FC4 was released; as expected, being bleeding edge FC4
was buggy as hell (compared to FC2/FC3). However, living on the bleeding
edge comes at a price, and I'm willing to pay it.
Like any good user, I did my best to report what-ever bugs I saw:
Some of them were fix promptly:
Missing KDE screensavers:
LVM2 boot problems:
But others were not:
GDB segfaults when debugging libraries:
Thus far, nothing new. FC4 has known bugs that are being (slowly?)
However, here comes my problem:
FC5 is 6 months (at least) away.
The lvm2 problem was only fixed in -updates; it still plagues new
installations. (I saw a couple of threads about it in fedoraforum.) The
fix did not go downstream to a new ISO images.
FC4 users cannot use their Palm and a full fix is no where to be seen.
GDB is effectively dead when debugging libraries.
And python python-gtksourceview, while fixed in rawhide, will not making
it (at least to my knowledge) into FC4.
My question is simple:
Is it the view of the FC foundation, that the FC4 bug-fixing is taking
second seat to the FC5 development?
Is it acceptable, again, in the FC foundation's eyes, that up until the
release of the FC5, people will not be able to sync with their Palm or
have dead installations on their hands (lvm problem)
I may be wrong here, isn't the lvm problem big enough to require ISO
Isn't the Palm problem serious enough to warrant a switch to an older
pilotd (and gnome-pilot)?
Am I the only to feel that these problems are critical?
Again let me stress, that being free-riding user (Sadly enough, I've yet
to pitch in and find some why to contribute to FC) I'm in no position to
rant about the stability of FC4.
However, being someone with vested interest in the FC project, I'm very
interested in the view of the FC project about the above.
I know how to generate multiple rpms from a single .spec, but I don't
know how to generate
a mix of .i386 and .noarch rpms. Is this possible?
If so, could someone point me at an example, or provide a quick howto?
Hi, Jeremy, hi, Peter:
I would really appreciate if someone added a capability to Anaconda to
install from /dev/uba1 and similar. This way, we do not need to reload
usb-storage, which contributes to various problems.
I did not request to turn ub on in kernel yet, because it interferes
with usb-storage and this makes some users unhappy. But I am working
on solutions to that, and the results are encouraging. I may need
Anaconda support in FC5 on a very short notice if we wait until that
Do you want me to file an RFE in Bugzilla?
any change to someone package mysql-administrator and
mysql-query-browser for fedora eg in extras. it seems mysql ab no longer
create rpms for fedora and redhat and it'd be useful to package these
Levente "Si vis pacem para bellum!"
I've installed yum-2.4.0-3 yesterday and now I can't update any package
--> Processing Dependency: policycoreutils >= 1.25.9-1 for package:
--> Processing Dependency: selinux-policy-targeted = 1.25.4-11 for
Traceback (most recent call last):
File "/usr/bin/yum", line 27, in ?
File "/usr/share/yum-cli/yummain.py", line 133, in main
(result, resultmsgs) = base.buildTransaction()
File "/usr/lib/python2.4/site-packages/yum/__init__.py", line 339, in
(rescode, restring) = self.resolveDeps()
File "/usr/lib/python2.4/site-packages/yum/depsolve.py", line 251, in
(checkdep, missing, conflict, errormsgs) = self._processReq(dep)
File "/usr/lib/python2.4/site-packages/yum/depsolve.py", line 382, in
File "/usr/lib/python2.4/site-packages/yum/depsolve.py", line 489, in
for pkg in pkgs:
TypeError: iteration over non-sequence
Radek Vokál <radekvokal(a)gmail.com>
I wanted to bring up something that has bothered multimedia users and
developers for a long time about Fedora - the ownership of devices.
Currently, in Fedora, permissions are set to 600. Ownership is set to
root, until a user logs in, and then it gets owned by the user.
This has multiple problems.
a) user switching assigns ownership to the first user that managed to
log in on the machine. A second login under a different account will
not be able to use sound or video devices
b) server programs like flumotion and icecast cannot be installed to
work as a service without manual intervention
A lot of other distributions solve this simply by creating a group for
these devices, "audio", "sound", "media", whatever. Then normal users
get added to this group, and rpms can as part of the install add their
newly-created user to this group, and everyone is happy.
The only problem is that apparently it is hard for an rpm to add a group
when a system is being upgraded between distros; ie, while anaconda
could add this new group just fine, people doing a dist upgrade, it is
claimed, will not get this group installed.
I would like to know if anyone else sees this as a problem (I have to
explain the situation to users that want to install and run flumotion
pretty much every time someone tries it). At this point I am very
tempted to do evil things from the flumotion rpm just so this problem
gets dealt with instead of being shuffled under the carpet :)
Dave/Dina : future TV today ! - http://www.davedina.org/
<-*- thomas (dot) apestaart (dot) org -*->
I can't go away with you on a rock climbing weekend
What if something's on TV and it's never shown again
Just as well I'm not invited I'm afraid of heights
I lied about being the outdoor type
<-*- thomas (at) apestaart (dot) org -*->
URGent, best radio on the net - 24/7 ! - http://urgent.fm/
ISTR someone was playing around with jigdo <http://atterer.net/jigdo/> and upgrading
the FC[3or4] test isos to the final isos with a massive download bandwidth saving.
Is that person still around?
It would certainly make FC4.x updates easier.
I'm going to have a look myself, but if there's anyone out there who knows more about
jigdo than I (not difficult :-) please speak up.
Cheers! (Relax...have a homebrew)
THEOREM: VI is perfect.
PROOF: VI in roman numerals is 6. The natural numbers < 6 which divide 6 are
1, 2, and 3. 1+2+3 = 6. So 6 is a perfect number. Therefore, VI is perfect.
-- Arthur Tateishi
After the recent discussion "FC4 state of affairs and FC5"
on fedora-devel about the possibility of making updated
installation iso images I gave it a try and actually
managed to put it together nicely.
(Can't say I liked the anaconda build utils, but ohwell)
So, if you feel like testing a potentially dangerous
unofficial release of updated FC4 install images I have
put them up on a server.
It's only i386, only cd, only torrent.
Please don't bug Fedora developers with FC4.1 specific bugs.