All spins have succeeded in composing against rawhide today (20090424):
http://www.kanarip.com/revisor/
Here's how this list of spins work:
- Each day, a compose is being attempted.
- A spin that succeeded is listed with a green background, red then obviously is a spin that failed.
- The name of the spin is a link to the spins "root" directory;
- .torrent file for downloading the spin results - iso/ directory with the .iso so that you can see the size of the product - log/ directory with some debugging output as well as reporting on package sizes, packages included, etc. - os/ directory with the binary and source RPMs used.
Please take some time to either;
- Download the .torrent and go at testing the compose (the .torrent is going to be rather slow as it runs on my home internet connection), or
- compose yourself against the master branch.
Please do so before Monday, and let us know if you're OK, or whether there's things that still need fixing, so that we can add your spin to the agenda and ACK.
Thank you in advance,
Kind regards,
Jeroen van Meeuwen -kanarip
On Fri, Apr 24, 2009 at 12:48:58 +0200, Jeroen van Meeuwen kanarip@kanarip.com wrote:
Please take some time to either;
- Download the .torrent and go at testing the compose (the .torrent is
going to be rather slow as it runs on my home internet connection), or
Is just testing i386 OK? I don't have an x86_64 system at home.
Is there going to be one more push of the spin-kickstarts package before the release or is that going to wait for updates? (I don't think there has been a release since I dropped nexuiz after its new version had a very large size increase.)
Am Fri, 24 Apr 2009 12:48:58 +0200 schrieb Jeroen van Meeuwen kanarip@kanarip.com:
All spins have succeeded in composing against rawhide today (20090424):
I assume the oversized KDE live images are still related to the revisor issue not to understand wildcard removals (like "-scim*)? I see at least some ibus packages which shouldn't be there:
http://www.kanarip.com/revisor/20090424/rawhide-i386-live-kde-respin/log/rpm...
Sebastian
On Fri, 24 Apr 2009 19:44:58 +0200, Sebastian Vahl deadbabylon@googlemail.com wrote:
Am Fri, 24 Apr 2009 12:48:58 +0200 schrieb Jeroen van Meeuwen kanarip@kanarip.com:
All spins have succeeded in composing against rawhide today (20090424):
I assume the oversized KDE live images are still related to the revisor issue not to understand wildcard removals (like "-scim*)? I see at least some ibus packages which shouldn't be there:
http://www.kanarip.com/revisor/20090424/rawhide-i386-live-kde-respin/log/rpm...
I'll investigate this...
Kind regards,
Jeroen van Meeuwen -kanarip
Am Sat, 25 Apr 2009 02:01:16 +0200 schrieb Jeroen van Meeuwen kanarip@kanarip.com:
On Fri, 24 Apr 2009 19:44:58 +0200, Sebastian Vahl deadbabylon@googlemail.com wrote:
Am Fri, 24 Apr 2009 12:48:58 +0200 schrieb Jeroen van Meeuwen kanarip@kanarip.com:
All spins have succeeded in composing against rawhide today (20090424):
I assume the oversized KDE live images are still related to the revisor issue not to understand wildcard removals (like "-scim*)? I see at least some ibus packages which shouldn't be there:
http://www.kanarip.com/revisor/20090424/rawhide-i386-live-kde-respin/log/rpm...
If you need a package list for comparison, this one is for my push from yesterday: http://www.deadbabylon.de/fedora/livecd/packagelists/f11/long-F11-KDE-086-i6...
After a short look it seems that's the problem I raised in IRC some time ago, that wildcards are not working properly (like -m17n*, -kaffeine*,-ibus*).
Sebastian
On 04/25/2009 10:09 AM, Sebastian Vahl wrote:
If you need a package list for comparison, this one is for my push from yesterday: http://www.deadbabylon.de/fedora/livecd/packagelists/f11/long-F11-KDE-086-i6...
After a short look it seems that's the problem I raised in IRC some time ago, that wildcards are not working properly (like -m17n*, -kaffeine*,-ibus*).
I think I may have solved this now, so it should show in tomorrow's compose.
The size of the .iso is 715M in my last compose attempt, does that seem about right?
Kind regards,
Jeroen van Meeuwen -kanarip
Am Sun, 26 Apr 2009 00:24:42 +0200 schrieb Jeroen van Meeuwen kanarip@kanarip.com:
On 04/25/2009 10:09 AM, Sebastian Vahl wrote:
If you need a package list for comparison, this one is for my push from yesterday: http://www.deadbabylon.de/fedora/livecd/packagelists/f11/long-F11-KDE-086-i6...
After a short look it seems that's the problem I raised in IRC some time ago, that wildcards are not working properly (like -m17n*, -kaffeine*,-ibus*).
I think I may have solved this now, so it should show in tomorrow's compose.
The size of the .iso is 715M in my last compose attempt, does that seem about right?
It's still too big. i686 should be ~693M and x86_64 ~699M. At least with my latest kickstart which I've pushed on friday. But also with a former one the difference shouldn't be more than 5-7M and 715M would also be too big.
I'll have a look at tomorrows compose.
Sebastian
On 04/26/2009 12:48 AM, Sebastian Vahl wrote:
Am Sun, 26 Apr 2009 00:24:42 +0200 schrieb Jeroen van Meeuwenkanarip@kanarip.com:
The size of the .iso is 715M in my last compose attempt, does that seem about right?
It's still too big. i686 should be ~693M and x86_64 ~699M. At least with my latest kickstart which I've pushed on friday. But also with a former one the difference shouldn't be more than 5-7M and 715M would also be too big.
Using livecd-tools, I can confirm the resulting image is 699M.
Kind regards,
Jeroen van Meeuwen -kanarip
I'll have a look at tomorrows compose.
The compose from today looks much better now. But when comparing your and my x86_64 version, your one does have 11 packages more. Some of them are ibus packages which shouldn't be installed:
ibus ibus-gtk ibus-libs im-chooser imsettings imsettings-libs iso-codes libgxim pyxdg wget xml-common
Most of them are dependencies of ibus. But I don't know why wget is there.
(I've took this file for comparision: rpms-rawhide-x86_64-live-kde-respin.log)
Sebastian
Em Dom, 2009-04-26 às 22:31 +0200, Sebastian Vahl escreveu:
I'll have a look at tomorrows compose.
The compose from today looks much better now. But when comparing your and my x86_64 version, your one does have 11 packages more. Some of them are ibus packages which shouldn't be installed:
ibus ibus-gtk ibus-libs im-chooser imsettings imsettings-libs iso-codes libgxim pyxdg wget xml-common
Most of them are dependencies of ibus. But I don't know why wget is there.
(I've took this file for comparision: rpms-rawhide-x86_64-live-kde-respin.log)
The BrOffice.org Spin is also oversized when composed using revisor. But I did a test with livecd-creator and I got a 668MB ISO for i586 and 680MB for x86_64 what sounds pretty reasonable to me. The spin is in a good shape for F11. There is a general bug in F11 affecting the abnt-br keyboard layout (and probably others), but this should be fixed in a different level and doesn't prevent the spin from compose or boot.
Regards, Igor Pires Soares
On Fri, Apr 24, 2009 at 12:48:58 +0200, Jeroen van Meeuwen kanarip@kanarip.com wrote:
All spins have succeeded in composing against rawhide today (20090424):
Please take some time to either;
- Download the .torrent and go at testing the compose (the .torrent is
going to be rather slow as it runs on my home internet connection), or
- compose yourself against the master branch.
Is this just the spin-kickstarts master branch with the rawhide livecd-tools or do we need to get a git version of that too?
I tried the download option, but the torrent was invalidated before the download finished and at the rate it was getting pushed trying again won't complete before the end of the weekend.
So I'll test a local version. Mostly I have been having video card related issues and space concerns in my previous testing. Space looks OK right now with Nexuiz removed. The video issues I had aren't directly related to the spin.
On Fri, Apr 24, 2009 at 12:48:58 +0200, Jeroen van Meeuwen kanarip@kanarip.com wrote:
- compose yourself against the master branch.
I built both i686 and x86_64 versions of the game spin with livecd-creator and they are both OK size wise. I will do some testing of the i686 version tonight. I probably won't get a chance to test the x86_64 version before the meeting.
I expect I'll see some radeon driver problems show up, but they aren't game spin specific.
On Fri, Apr 24, 2009 at 12:48:58 +0200, Jeroen van Meeuwen kanarip@kanarip.com wrote:
- compose yourself against the master branch.
I tested the i686 games spin.
Sound worked.
3D games mostly crashed. (I have an rv280 based card and the games tremulous and alien arena don't work normally on that machine.)
Booting was very slow. (Several times normal.)
I noticed a udev warning that did not seem specific to the games spin, but related to live images and filed a bug.
On Sun, Apr 26, 2009 at 22:41:33 -0500, Bruno Wolff III bruno@wolff.to wrote:
On Fri, Apr 24, 2009 at 12:48:58 +0200, Jeroen van Meeuwen kanarip@kanarip.com wrote:
- compose yourself against the master branch.
I tested the i686 games spin.
Sound worked.
3D games mostly crashed. (I have an rv280 based card and the games tremulous and alien arena don't work normally on that machine.)
Booting was very slow. (Several times normal.)
I tested an x86_64 spin on a system with an rv530 based card and tremulous worked. Booting was still slow. There were some graphics artifacts but the -6 version of the ati driver has some regressions caused by attempted r690 fixes. Airlie built -7 last night and I have that installed on the normal os for the machine and things look better.
I won't be able to attend the meeting today because of a conflicting meeting at work. I plan to leave an irc session open so I can read through the discussion later, but I won't actually be present.
-
Bruno Wolff III -
Igor Pires Soares -
Jeroen van Meeuwen -
Sebastian Vahl