can not create archives on FC2, and something about file-roller
by Zhiyuan Liu
I can not create archives on FC2.
I right-click on a folder in views of nautilus, then the context-sensitive menu is popuped, then
I choose the "Create Archive" .But nothing happens. The dialog box does not appeard,on which the type and file name of the archive can be inputed.
This bug also occurs in my FC3-test1 system.
This might be concerned with the context-sensitive menu activation in bonobo system, because
this bug will disappear when I delete following lines in the file "component/File_Roller_Component.server.in.in" of file-roller's souce code
<oaf_attribute name="bonobo:supported_mime_types" type="stringv">
<item value="application/x-tar"/>
<item value="application/x-compressed-tar"/>
<item value="application/x-bzip-compressed-tar"/>
<item value="application/x-lzop-compressed-tar"/>
<item value="application/zip"/>
<item value="application/x-arj"/>
<item value="application/x-zip"/>
<item value="application/x-lha"/>
<item value="application/x-rar"/>
<item value="application/x-rar-compressed"/>
<item value="application/x-gzip"/>
<item value="application/x-bzip"/>
<item value="application/x-compress"/>
<item value="application/x-lzop"/>
<item value="application/x-zoo"/>
<item value="application/x-jar"/>
</oaf_attribute>
<oaf_attribute name="bonobo:supported_mime_types" type="stringv">
<item value="*/*"/>
</oaf_attribute>
these lines control the context-sensitive menu activation of file-roller.
so the bug may be caused by bonobo/bobonoui or file-roller itself.
anyone can help?
-------------
Zhiyuan Liu
-------------
19 years, 6 months
Current rawhide kernel?
by Louis Garcia
Rawhide has kernel-2.6.8-1.541, the same as what came with FC3T2. Has
their not been and new development kernels since then?
Somehow I am running kernel-2.6.8-1.590, I got this from rawhide maybe a
week ago. Then the kernel version went back to .541. Should I be
testing .541 or .590?
--Thanks
19 years, 6 months
Hacking Anaconda
by Mark Wormgoor
Hi,
I have a question; how can I quickly hack Anaconda?
I burnt boot.iso to a cd and copied the Fedora3 files to a local NFS
mount. But, if I hack netstg2.img, I get an error during install that
it is not the correct directory for my boot.iso.
Is there a way to circumvent this check, or do I have to recreate and
burn the iso every time? There must be a quicker way to hack Anaconda?
Kind regards,
Mark
--
***************************************************************
* |\ /| | /| / Mark Wormgoor *
* | \ / | | / | / mailto:mark@wormgoor.com *
* | \/ |ark |/ |/ormgoor http://www.wormgoor.com/mark/ *
***************************************************************
19 years, 6 months
sqlite?
by Neal Becker
Hardly might seem that we need yet another database, but it looks like
sqlite is rapidly becoming very popular. Being 'lite', it really serves a
different need than others. I think it would be a good addition to fedora.
19 years, 6 months
rp-pppoe/initscripts/broken PPPoE annoyance
by Pekka Pietikäinen
Hiya
I'm wondering if I'm the only one annoyed by rp-pppoe (#134923)/udev
(#131114)/whatnot bug of the day/provider flakyness that makes the system
get stuck on boot while trying to get PPPoE up. (sysrq kill-everything or
booting in interactive mode and skipping networking fortunately gets around
this so whatever was the problem can be repaired)
What's especially evil is that this happens before syslogd is up ->
no useful debug output on the console.
Not sure what the proper solution would be or whether this even is a bug
(the bugs that have triggered this behaviour for me are fixed, sure).
A simple workaround is just using ONBOOT=no and putting "ifup
<interface> &" in rc.local. Other options would be to have an option for
ifuping interfaces in the background (some distros apparently do this),
another to change adsl-start to try a few times and if it still fails
either give up or continue trying in the backgruond.
Comments?
--
Pekka Pietikainen
19 years, 6 months
Re: Cooperative Bug Isolation project on Fedora Core 2 & 1
by Marius Andreiana
On Tue, 2004-10-05 at 23:03 -0700, Ben Liblit wrote:
> Marius Andreiana asked:
> > Is it possible to get debugging packages for software in FC3t2?
>
> Unfortunately, we're not planning to post FC3t2 packages right now.
> That is due in part to limited manpower and in part to our need for a
> large user base.
>
> To clarify that second requirement, we don't look at individual
> failures. Rather, we hunt for bugs by mining statistical *trends* out
> of large numbers of runs. I'm not convinced that there are enough FC3t2
> users to produce the kind of numbers we would need. I do understand
> that this is where the most unstable code lives, but it's not where the
> most users live, and our techniques really depend on having lots of users.
>
> Now, if the Fedora engineering team wanted make this part of the
> standard Fedora build process, so that *every* test release user would
> be getting our instrumented code, that would be a whole other story....
is this possible? at least for a some desktop packages.
(this is posted to fedora-devel)
--
Marius Andreiana
Galuna - Solutii Linux in Romania
http://www.galuna.ro
19 years, 6 months
Re: Linux@DUKE Nosrc-RPM project.
by Damian Christey
Greetings and thanks for your great work,
I have also been packaging proprietary software (Matlab to start with)
to run on Fedora for our Math Department and am interested in
collaborating with others who are doing similar things.
There are a few things I did slightly differently.
1. I build one monolithic Matlab package rather than packaging each
toolkit separately. Not sure if this is good or bad, I was just trying
to make building packages as simple as possible.
2. I used the "expect" program to make a script that automates the
Matlab installer. This adds a build dependency on expect >= 5.0, which
is in the base Fedora distribution as well as previous Redhat versions
back to 6.0, but may not be installed by default.
3. I tried to make the whole process as generic as possible, so it could
be easily modified to build packages for other proprietary software.
I've made a nosrc rpm available here:
http://www.math.wvu.edu/~damian/fedora/nosrc-rpm/matlab-7.0.1-2.nosrc.rpm for testing purposes.
To build packages:
1. Download the Matlab installer, all the toolboxes you want to package,
and your license.dat file from the Mathworks' website.
2. Place all of the downloaded files in a folder named "matlab" in your
build directory, for example: /usr/src/redhat/BUILD/matlab/
3. Make sure expect is installed:
# rpm -q expect
if not:
# yum install expect
4. Build it!
# rpmbuild --rebuild matlab-7.0.1-2.src.rpm
--
Damian Christey <damian(a)math.wvu.edu>
System Administration, WVU Department of Mathematics
19 years, 6 months
Telnet doesn't work in FC3test2
by RAT
Does anyone experienced the same problem as I did. I was using telnet
for connecting to some machine but now in FC3t2 I always get this
telnet kosh
Trying 130.#.#.#...
telnet: connect to address 130.#.#.#: Connection timed out
telnet: Unable to connect to remote host: Connection timed out
Where's the problem? Is it somehow related to SELinux?
(ssh to another machime still works)
RAT
19 years, 6 months
Linux@DUKE Nosrc-RPM project.
by Konstantin Ryabitsev
Hello, everyone:
A while ago we were talking about packaging proprietary software. Since
I've had to do a lot of it for Duke, I've cleaned up our specfiles and
extra sources and released them as .nosrc.rpm packages so others can use
them and not re-invent the wheel. They are released as part of
Linux@DUKE Nosrc-RPM project and can be found here:
http://linux.duke.edu/projects/nosrc-rpm/
Everyone is invited to use them for making their own packages. I've put
a lot of work into trying to make them sane, but if someone has any
comments, please feel free to email me.
(also sent to univ-linux mailing list).
Regards,
--
Konstantin ("Icon") Ryabitsev
Duke University Physics Sysadmin
19 years, 6 months
SRPM's and determining spec file
by Alan Milligan
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Hi,
Not sure if this is the right place to discuss the behaviour of RPM, but
I'm sure somebody can correct me!
I have a requirement to determine the spec file from a SRPM so that my
rpmbuild can be guaranteed to be passed a valid file.
I'd hoped to just find in the BASENAMES, however, it seems to be
filtered out when using rpm -q.
So then I thought I'd just use the RPM2 perl module. This indeed works,
excepting that it fails constantly due to bad signatures etc in packages:
error reading package at
/usr/lib/perl5/site_perl/5.8.3/i386-linux-thread-multi/RPM2.pm line 85.
To my mind (i) rpm -q --qf"%{BASENAMES}" should not filter the .spec
file; and (ii) the RPM2 perl module should return as much of the header
as it can construct, with undefs etc in bad fields. Even throwing an
exception or returning an error code would at least put the decision in
user space.
I am happy to provide a patch for either if someone would indicate a
preference.
Cheers, Alan
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
iD8DBQFBYv5DCfroLk4EZpkRAnyxAKDOUdnARVP5u+lpVI3uF7L98H2iiACfaeP3
WjZHpzRxqlW8pLxS0MtqQW8=
=5ltK
-----END PGP SIGNATURE-----
19 years, 6 months