Re: invalid syslinux config in case of usb-hdd for sid/lenny
On Tue, Jun 26, 2007 at 10:08:22PM +0200, Daniel Baumann wrote: fixed differently, because we still need that stupid workaround on etch, Danial, I am sorry, but since You don't reply, I am not sure if You see, that You replace one bug with another. There is logical error in Your if/then statement, where first part executed in case of more then one kernels, and second one go in case of one only kernel. But this true for etch only. For sid/lenny the second part executed in all cases. This mean that in case of more then one kernels You doing things designed for case of one only kernel. This is wrong and produce syntax error in mv command making image unusable. Please, look at the bottom of lh_binary_syslinux carefully and fix this. By the way, as I pointed earlier, this is useful bug, because it overcome the 8.3 format problem: in case of one only kernel You make short filenames for all distributions. -- * Vladimir Stavrinov ** *** [EMAIL PROTECTED] ** * ___ Debian-live-devel mailing list Debian-live-devel@lists.alioth.debian.org http://lists.alioth.debian.org/mailman/listinfo/debian-live-devel
Re: invalid syslinux config in case of usb-hdd for sid/lenny
don't be that inpatient ;) otavio changed the code of lh_binary_syslinux in svn, right now it has a new bug with memtest entry, and the numbering is now wrong too. means, i will have to look at it anyway. in the meanwhile, patches are welcome. -- Address:Daniel Baumann, Burgunderstrasse 3, CH-4562 Biberist Email: [EMAIL PROTECTED] Internet: http://people.panthera-systems.net/~daniel-baumann/ ___ Debian-live-devel mailing list Debian-live-devel@lists.alioth.debian.org http://lists.alioth.debian.org/mailman/listinfo/debian-live-devel
Re: invalid syslinux config in case of usb-hdd for sid/lenny
On Fri, Jun 29, 2007 at 10:34:29AM +0200, Daniel Baumann wrote: don't be that inpatient ;) I am not. I was not sure if You know about this bug. It would be enough to send short reply, recognized this bug like Yes, I see and will fix it. in the meanwhile, patches are welcome. You've got one in the very first message of this thread, but You reject it. It is simple and leave one only alternative: one kernel or more for all distributions. But You want long filenames and this become source of this bug. -- * Vladimir Stavrinov ** *** [EMAIL PROTECTED] ** * ___ Debian-live-devel mailing list Debian-live-devel@lists.alioth.debian.org http://lists.alioth.debian.org/mailman/listinfo/debian-live-devel
Image QA
Hey guys, I've been looking into QA for the images I generate with Debian-Live. Webconverger.com generates several new ISOs on an update and I am worried sick something might have gone wrong. ;) What usually goes wrong for me is that sometimes the package(s) live-helper needs is misplaced (my fault) or perhaps failed to be fetched (network fault). I need to also probably come up with a couple of fault finding greps of the debian-live log output. Brendan Sleight (see mail below) from the morphix project does video stuff. I think it could be useful to quickly watch a movie in fast forward if it looks right. (black box testing?) Also Martin Ferrari tells me he is working on a lintian fully automated style testing of images. (white box testing?) Other elements of testing I would like to pursue (if I had more time) is the PC World test. I would like to plug a USB stick of my image and test it on all the machines on demo in the store. I probably just need to create a persistent USB image for the live.log and where perhaps a `hwinfo` dump could live. This was talked about during Debconf7 as perhaps being also part of Debian popcon stuff. Anyone have some QA tips or thoughts, please share. Best wishes, - Forwarded message from Brendan Sleight [EMAIL PROTECTED] - From: Brendan Sleight [EMAIL PROTECTED] To: [EMAIL PROTECTED] Subject: Re: iceweasel-webconverger Date: Thu, 28 Jun 2007 23:58:18 +0100 X-Original-To: [EMAIL PROTECTED] Delivered-To: [EMAIL PROTECTED] X-Forwarded-To: [EMAIL PROTECTED] X-Forwarded-For: [EMAIL PROTECTED] [EMAIL PROTECTED] Delivered-To: [EMAIL PROTECTED] X-Spam-Checker-Version: SpamAssassin 3.0.3 (2005-04-27) on spunkymail-mx7.dreamhost.com On 28/06/07, Kai Hendry [EMAIL PROTECTED] wrote: think up a good way of testing all these ISOs I am creating at I wrote a very ugly script to AutoTest the Morphix ISOs. http://morphix.svn.sourceforge.net/viewvc/morphix/trunk/mmaker/utils/auto-testing/auto_testing.sh?view=markup I have tweaked the script to be less Morphix specific:- http://morphix.svn.sourceforge.net/viewvc/morphix/trunk/mmaker/utils/auto-testing/auto_testing.webconverger.sh?view=markup In a nutshell, the script runs qemu, viewed via vnc session, then captures screenshots. Then converts the screenshots into a video. I am using vnc, so one day I can drop the require to be running a local x-server. (Google for vncrec) http://en.wikipedia.org/wiki/QEMU Required packages - imagemagick ffmpeg ffmpeg2theora mencoder qemu The Autotesting is a big tough on tmp disk space and CPU usage. I have a spare old Laptop that runs the Autotesting overnight. Each full test takes ~1 hour, depending on how many frames are grabbed. http://www.barwap.com/morphix/webconverger/webc-2.17.cn.iso.ogg http://www.barwap.com/morphix/webconverger/webc-2.21.en.iso.ogg I have space on the machine to test an webconverger.iso per day. In a week of cron job I could get through most of all language versions. I would be glad to help webconverger, I have been using your package Iceweasel-webconverger_6-16_all.deb. It works really well. http://morphix.svn.sourceforge.net/viewvc/morphix/trunk/morphixlivekiosk/iceweasel-webconverger-6/ Hops this helps, Brendan - End forwarded message - signature.asc Description: Digital signature ___ Debian-live-devel mailing list Debian-live-devel@lists.alioth.debian.org http://lists.alioth.debian.org/mailman/listinfo/debian-live-devel
Re: Some corrections/suggestions to live-helper
Antonio Amorim wrote:
First of all, my thanks. Live-helper is really becoming an excellent tool!
welcome.
1 - On /etc/cron.daily/live-helper
the ARCHITECTURE options -a amd64 or -a i386 are missing from
the make-live statment.
Something like:
if [ $ARCHITECTURE == amd64 ]
then
OPTIONS=${OPTIONS} -a amd64
fi
if [ $ARCHITECTURE == i386 ]
then
OPTIONS=${OPTIONS} -a i386
fi
not needed. the architecture is, if unset, guessed by the host system.
as long as you don't want to build for another architecture (like
building i386 on amd64), this is not needed.
2 - For amd64, in lh_binary_debian-installer, downloading the elilo
package raises an exception that is caught by trap and messes
everything. Please remove it.
done.
3 - If one has a package list that contains many packages aptitude can
go into an infinite loop while guessing the dependences. I have seen
this behavior systematically in etch.
Everything is fine if aptitude installs one package at the time.
My patch goes in lh_chroot_packageslists:
i'm not sure if we want this by default. aptitude does much better
conflicts handling when all packages are installed in one shot and the
loop/segfault of aptitude afaik only happens with really, really long lists.
I hope this can help a bit your excellent work.
jup, thanks for reporting.
--
Address:Daniel Baumann, Burgunderstrasse 3, CH-4562 Biberist
Email: [EMAIL PROTECTED]
Internet: http://people.panthera-systems.net/~daniel-baumann/
___
Debian-live-devel mailing list
Debian-live-devel@lists.alioth.debian.org
http://lists.alioth.debian.org/mailman/listinfo/debian-live-devel
Re: Debian live booting trouble
Hi thomas, I think the problem is live-initramfs_1.91.1-1_i386.deb. Remove that and try again. monty:/tmp/t% tree config config |-- binary |-- binary_grub |-- binary_local-hooks |-- binary_local-includes |-- binary_local-udebs |-- binary_rootfs |-- binary_syslinux |-- bootstrap |-- chroot |-- chroot_local-hooks |-- chroot_local-includes |-- chroot_local-packages | `-- live-initramfs_1.91.1-1_i386.deb |-- chroot_local-packageslists |-- chroot_sources |-- common |-- includes |-- source `-- templates 13 directories, 6 files Best wishes, ___ Debian-live-devel mailing list Debian-live-devel@lists.alioth.debian.org http://lists.alioth.debian.org/mailman/listinfo/debian-live-devel
Re: Image QA
Kai Hendry wrote: I've been looking into QA for the images I generate with Debian-Live. basically, there are four types of errors: * build-process is buggy (bugs in live-helper) * build-process was interrupted (flacky network connection etc.) * run-time error (bugs in live-initramfs) * formal errors (like incorrect .disk/* files on the image) the first and the second one can be easily catched by checking for a complete build and approximately size; together with the possibility of reviewing the build log, there is not much to be done here. the third one could be done with qemu as they do with morphix (thanks for asking brendan about the script he uses), which is really, really great. i'm looking forward to setup a machine to build daily svn sid images soon (about three weeks from now), hopefully. the fourth one could be tested with martin ferraris tool. i had a quick chat about his lintian-for-cd google summer of code project, and he wasn't against including livecd specific tests once he has some working code published. looking forward to do that too. Anyone have some QA tips or thoughts, please share. Best wishes, well, not actually.. you said everything already :) -- Address:Daniel Baumann, Burgunderstrasse 3, CH-4562 Biberist Email: [EMAIL PROTECTED] Internet: http://people.panthera-systems.net/~daniel-baumann/ ___ Debian-live-devel mailing list Debian-live-devel@lists.alioth.debian.org http://lists.alioth.debian.org/mailman/listinfo/debian-live-devel
