Re: Lowering down my involvment in D-I
On Sun, Dec 06, 2009 at 02:14:04PM +0100, Christian Perrier wrote: I has been obvious for the past two months and even more for about the last twelve months: I can't really dedicate the time I would need to dedicate, to D-I work. I would like to thank you for all the time and energy you've put into d-i. I sincerely hope that, at some time, new energies will show up in D-I development. We desperately need it. It seems like that's always the case. :) Peace, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! step...@marenka.net -- To UNSUBSCRIBE, email to debian-boot-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: [RFC] Dropping architectures alpha and m68k from the Installation Guide
On Tue, Oct 13, 2009 at 08:58:43PM +0200, Frans Pop wrote: m68k was dropped as a release architecture for Etch, and alpha is being dropped for Squeeze. For m68k the intention was to bring it back as a release architecture, but now, over 2 years later, I think we can say that's unlikely to happen. Indeed. Opinions, comments and suggestions welcome. No objections. Peace, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! step...@marenka.net -- To UNSUBSCRIBE, email to debian-boot-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: [RFC] Packages that looks ready for uploading
On Thu, Jun 18, 2009 at 06:08:51PM +0200, Wouter Verhelst wrote: On Thu, Jun 18, 2009 at 02:54:48PM +0100, Colin Watson wrote: Yes please. How does one go about uploading m68k-only packages? Don't you have to do a source+m68k upload, which presumably the main archive would reject because it doesn't know about m68k? Yes, indeed. We can upload to debian-ports, but not to the main archive (which is why I suggested not to worry about it ;) Argh! I forgot about that. Also uploading source to d-ports is only allowed for unreleased not for sid. Peace, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! step...@marenka.net signature.asc Description: Digital signature
Re: [RFC] Packages that looks ready for uploading
On Mon, Jun 15, 2009 at 10:05:04AM -0300, Otavio Salvador wrote: Wouter Verhelst wou...@debian.org writes: On Sun, Jun 14, 2009 at 07:32:56PM +0200, Christian Perrier wrote: Quoting Otavio Salvador (ota...@debian.org): [...] The following was not uploaded because they need to be done by porters: [...] vmelilo-installer vmelilo-installer is specific to a rather rare m68k subarch. It should not be kicked from the repository (we still do care about m68k), but it should not be of much concern to d-i development anymore these days. Yes I agree that we shouldn't worry too much about it but if we can keep it in sync with rest it is better and makes porter's life easier too. I can upload it if that would make life easier. :) Peace, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! step...@marenka.net signature.asc Description: Digital signature
Re: The future of the D-I team
On Mon, Mar 02, 2009 at 06:43:03AM +0100, Christian Perrier wrote: Quoting Stephen R Marenka (step...@marenka.net): FWIW: m68k still builds too [1]. [1] http://people.debian.org/~smarenka/d-i/m68k/images/daily/ Hmmm, I can't check right now (offline) but I don't remember seeing mention of m68k builds on the general build status page. It's not. I don't know if it ever was or if it got dropped at some point. Peace, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! step...@marenka.net signature.asc Description: Digital signature
Re: The future of the D-I team
On Sun, Mar 01, 2009 at 12:53:03PM +0100, Frans Pop wrote: is worrysome. You only have to look at the daily build page [2] to see that. We currently only have 4 arches building reliably (i386, armel, sparc and s390). The rest have been absent or failing for some weeks or months. The only arches for which there is any real excuse for this are FWIW: m68k still builds too [1]. [1] http://people.debian.org/~smarenka/d-i/m68k/images/daily/ Peace, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! step...@marenka.net signature.asc Description: Digital signature
network console
Is there already a way to fire up just the ssh server portion of network console or something equivalent? I don't really want to run the whole installer under network-console, but I would like to be able to ssh in when the installer hangs. Thanks, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: Installing with D-I from d-ports.org
On Thu, Sep 25, 2008 at 09:52:40PM +0200, Frans Pop wrote: During installation you basically want to avoid displaying the mirror list. But you also want to allow users to use a local mirror. I've always wondered why there wasn't an option to set a local mirror interactively. I've had the occasion on one-off installs to need to use a local mirror, but had to restart the installer with preseeding. Is it just that nobody ever wrote the code? Easiest solution for that is probably to use initrd preseeding [1]. This should do the above and avoid the need to modify any components. The preseed file should contain the following three lines (ignoring security mirrors for now as you only have unstable): d-i mirror/country string manual d-i mirror/http/hostname string debian-ports.org d-i mirror/http/hostname seen false The first line ensures that the country list is not displayed. The last two lines ensure that the hostname _is_ asked, but has the correct default. The preseed file can be included in the images by using a mechanism that was in place for the Etch installer, but which has been dropped in trunk. Check out installer/build from the etch branch, and see the unstable.cfg file and grep for PRESEED_SUITE. IMO it should be no problem to again add something like that for d-ports. The variable (DEBIAN_PORTS maybe?) could simply be set in config/m68k.cfg. Sounds like a reasonable solution. Interesting that I hadn't really thought about *that* bit of the switch to debian-ports. I use approx, so I changed it once and forgot about it. Thanks, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: Installing with D-I from d-ports.org
On Fri, Sep 26, 2008 at 11:30:31PM +0200, Frans Pop wrote: On Friday 26 September 2008, Stephen R Marenka wrote: On Thu, Sep 25, 2008 at 09:52:40PM +0200, Frans Pop wrote: During installation you basically want to avoid displaying the mirror list. But you also want to allow users to use a local mirror. I've always wondered why there wasn't an option to set a local mirror interactively. I've had the occasion on one-off installs to need to use a local mirror, but had to restart the installer with preseeding. Is it just that nobody ever wrote the code? Huh? I always install using my local mirror. We have that nice option Enter information manually option at the top of the country list exactly for that purpose :-) Sigh, learn something new every day. :) Thanks, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: Uploads of daily D-I builds (was: on .ssh/authorized_keys files)
On Sat, May 31, 2008 at 01:12:09PM -0700, Steve Langasek wrote: On Sat, May 31, 2008 at 07:35:58PM +0200, Frans Pop wrote: On Friday 30 May 2008, Stephen R Marenka wrote: On Thu, May 29, 2008 at 10:08:02PM +0200, Frans Pop wrote: So basically this is what needs to be done to get uploads for daily D-I builds working again for remaining architectures. Does anybody who has a build running want to coordinate that? Maybe setup a (more) common system for it? I'm willing to coordinate if that will help. I just emailed weasel to find out what the procedure should be. Great. Note that Joey and Steve (vorlon) have now already gotten this done, so you could also ask one of them. Main thing is to try to get it done ASAP for all remaining arches. Well, for me the only procedure was ping weasel on IRC... :) I guess that makes me a bad candidate for success. :( -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: Uploads of daily D-I builds (was: on .ssh/authorized_keys files)
On Thu, May 29, 2008 at 10:08:02PM +0200, Frans Pop wrote: So basically this is what needs to be done to get uploads for daily D-I builds working again for remaining architectures. Does anybody who has a build running want to coordinate that? Maybe setup a (more) common system for it? I'm willing to coordinate if that will help. I just emailed weasel to find out what the procedure should be. Thanks, Stephen -- Forwarded Message -- Subject: on .ssh/authorized_keys files Date: Thursday 29 May 2008 From: Peter Palfrader [EMAIL PROTECTED] To: [EMAIL PROTECTED] The use of ~user/.ssh/authorized_keys files has been disabled since DSA1571 was announced. While our initial plan was to allow them again eventually some bad experience with DDs' key handling has led us to reconsider that intent. So ~user/.ssh/authorized_keys will remain disabled. If you want to login to debian.org hosts using keys you should send them to the LDAP as outlined at URL:https://db.debian.org/doc-mail.html, which allows us to do at least some quality control. Should you need keys only on specific hosts for automated tasks like updating stuff or syncing files between project machines or similar we can enable a user editable authorized_keys file for specific users on specific hosts. Usually we would expect those keys to be limited to use only from certain hosts (using from=xyz) and limited to allow execution of only certain commands (using command=foobar). Contact DSA if you have such a case. Your sysadmins --- -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: Bug#474698: partman-base: Inconsistent output if sector size != 512 bytes
On Tue, Apr 22, 2008 at 10:56:06AM -0300, Otavio Salvador wrote: Ferenc Wagner [EMAIL PROTECTED] writes: Otavio Salvador [EMAIL PROTECTED] writes: Frans Pop [EMAIL PROTECTED] writes: On Tuesday 08 April 2008, Ferenc Wagner wrote: I could go on to replace PED_SECTOR_SIZE_DEFAULT with dev-sector_size throughout partman-base/parted_server.c, but I'm getting suspicious. Isn't there a good reason for this state of affairs? Looks like this could possibly be a bug introduced by revision r37468 [1], which changed PED_SECTOR_SIZE to PED_SECTOR_SIZE_DEFAULT (partman-base 84). Looks like that change was made for compatibility with libparted 1.7. Whether changing it to dev-sector_size is correct depends of course on what the actual definition of PED_SECTOR_SIZE was in prior libparted versions. It could also be that the problem has always been there and that the use of PED_SECTOR_SIZE was always not correct (though maybe only in some places). This certainly needs to be checked very carefully. A good start could be to check whether the Sarge installer also has the bug as that has older versions of libparted and partman-base. Otavio: you did that change; any comments? I'll look into it. Nothing from the top of my head. Hi Otavio, did you find the time to look into this? I didn't. I was busy finishing a work task and plan to look into it this week. I hope to get parted OK for d-i this week (this includes the s390 regression). And maybe the atari patch? Pretty please? ;) Thanks, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: Hold the upload until Beta1 test images
On Sat, Mar 01, 2008 at 12:28:52PM -0300, Otavio Salvador wrote: I noticed that you tagged and uploaded m68k kernel udebs however, as you know, we're in mid of Beta1 releasing. Sure. As you know m68k doesn't have a beta to test. 2.6.24 fixes an atari/aranym bug present in 2.6.22. Moreover, I wasn't aware that multiple versions of kernel udebs would cause anybody any grief. We usually upload them before we switch over d-i, don't we? It's a bad time for it so I asked for the package rejecting to ftp-masters. Another thing that they noticed is that your upload was wrongly done since it lacked the source and would be nice if you could be more careful in next upload. Well I ran kernel-wedge build-all which is what I've done in the past. It spit out two changes files [0], which I don't remember. So I signed them and uploaded. Apparently wrong again, not my week. :( [0] linux-kernel-di-m68k-2.6_0.89_m68k.changes and linux-kernel-di-m68k-2.6_0.89_source.changes I'll upload kernel-wedge when it's ok for 2.6.24 uploads to start. Okay, hopefully I'll have time whenever that is, otherwise feel free to upload the m68k binaries. I'm already happy with the way the packaging works for now (at least until I've tested it by using it). I'll go crawl back under my rock for now. :) Peace, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: Subject: Debian Installer Lenny Beta1: Status Update 3
On Thu, Feb 28, 2008 at 03:28:48AM +0100, Frans Pop wrote: On Sunday 24 February 2008, Otavio Salvador wrote: I've sent debian-installer package and I will now try to get it built for all arches as soon as possible so we can change dailies to use those images. Just to keep people informed. I've today uploaded a new version of d-i (on request of Otavio as he was struggling with a dead mainboard) to fix the build error for arm. Before doing that, I reverted a commit by Stephen Marenka for m68k for two reasons. One is that his workaround would not result in working images and two is that I did not think the hack was suitable for official uploads. I later discussed it with Otavio and he told me that he would have done the same. See the commit message for details. | Revert 'm68k doesn't have a lenny, so try to build from sid' | This hack will not result in working images because they will still have | the DEBIAN_RELEASE set to lenny, which means that installation of additional | udebs will fail. As long as m68k is not in lenny, it is better to not | have it built at all than having semi-official builds that deviate from all | other arches. Sorry. I'd rather hoped that having sid images would be better than no images at all. Maybe we'll have a lenny to build from one day. :) Peace, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: Debian Installer Lenny Beta2: Kernel update work
On Fri, Feb 15, 2008 at 12:42:58PM -0200, Otavio Salvador wrote: Please porters, let's start the work for upgrading Debian Installer to 2.6.24. Those changes (for kernel-wedge and kernel modules packages) should remain on SVN repository and _CANNOT_ _BE_ _UPLOADED_ until we does have Beta1 images ready. As soon as it gives no trouble for us, I'll ping you all back to uploads start to be done. Is m68k just weird or does kernel-wedge need a patch? Index: modules/crypto-modules === --- modules/crypto-modules (revision 51488) +++ modules/crypto-modules (working copy) @@ -1,7 +1,10 @@ -aes +aes ? +aes_generic ? blowfish twofish serpent -sha256 +sha256 ? +sha256_generic ? +sha512 ? cbc ? ebc ? Also while building against 2.6.24-4 I ran into the following. Does anyone have any pointers? I need isofs. I hate not to have ppp. | kernel-wedge find-dups 2.6.24-1-amiga | debian/ppp-modules-2.6.24-1-amiga-di lib/modules/2.6.24-1-amiga/kernel/lib/zlib_inflate/zlib_inflate.ko | debian/ide-modules-2.6.24-1-amiga-di lib/modules/2.6.24-1-amiga/kernel/lib/zlib_inflate/zlib_inflate.ko | some modules are in more than one package | command exited with status 1 Thanks, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: Debian Installer Lenny Beta1: Status Update 1
On Fri, Feb 15, 2008 at 11:33:05AM -0200, Otavio Salvador wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Otavio Salvador [EMAIL PROTECTED] writes: Please porters, do the uploads for respective architectures that has only l10n updates. Following are the packages, architecture specific ones, that needs uploading: arch/m68k/vmelilo-installer I'm taking care of this for wouter. Can someone point me in the right direction please? I run output-l10n-changes -d . but no translations seem to have changed. Is there something I'm missing? Thanks, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: Debian Installer Lenny Beta2: Kernel update work
On Fri, Feb 15, 2008 at 08:02:53PM -0200, Otavio Salvador wrote: Stephen R Marenka [EMAIL PROTECTED] writes: On Fri, Feb 15, 2008 at 12:42:58PM -0200, Otavio Salvador wrote: Please porters, let's start the work for upgrading Debian Installer to 2.6.24. Those changes (for kernel-wedge and kernel modules packages) should remain on SVN repository and _CANNOT_ _BE_ _UPLOADED_ until we does have Beta1 images ready. As soon as it gives no trouble for us, I'll ping you all back to uploads start to be done. Is m68k just weird or does kernel-wedge need a patch? Index: modules/crypto-modules === --- modules/crypto-modules (revision 51488) +++ modules/crypto-modules (working copy) @@ -1,7 +1,10 @@ -aes +aes ? +aes_generic ? blowfish twofish serpent -sha256 +sha256 ? +sha256_generic ? +sha512 ? cbc ? ebc ? Yes. It needs changes due the modules renaming. Is sha512 really required? Not by me. Also while building against 2.6.24-4 I ran into the following. Does anyone have any pointers? I need isofs. I hate not to have ppp. | kernel-wedge find-dups 2.6.24-1-amiga | debian/ppp-modules-2.6.24-1-amiga-di lib/modules/2.6.24-1-amiga/kernel/lib/zlib_inflate/zlib_inflate.ko | debian/ide-modules-2.6.24-1-amiga-di lib/modules/2.6.24-1-amiga/kernel/lib/zlib_inflate/zlib_inflate.ko | some modules are in more than one package | command exited with status 1 humm, I can't check it now but will try to look at it later tomorrow Thanks, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: Debian Installer Lenny Beta1: Status Update 1
On Fri, Feb 15, 2008 at 09:47:04PM +0100, Frans Pop wrote: On Friday 15 February 2008, Stephen R Marenka wrote: On Fri, Feb 15, 2008 at 11:33:05AM -0200, Otavio Salvador wrote: Following are the packages, architecture specific ones, that needs uploading: arch/m68k/vmelilo-installer I'm taking care of this for wouter. Can someone point me in the right direction please? I run output-l10n-changes -d . but no translations seem to have changed. Is there something I'm missing? You have to do this _before_ changing 'UNRELEASED' to 'unstable' in the changelog or, if there are no pending other changes, before adding a new changelog entry. Thanks! -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: Preparation to mass translation uploads, porters help needed
On Wed, Feb 13, 2008 at 10:54:39AM -0200, Otavio Salvador wrote: Tomorrow we need to mass upload packages that has translation updates. All modules that has no other changes, but l10n ones, ought to be updated. If there're modules with other changes, those need to be checked with me. We'll decide together if we do or not each upload. Would partman-base and partman-partitioning fall into this somewhere? Thanks, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: Preparation to mass translation uploads, porters help needed
On Wed, Feb 13, 2008 at 02:32:11PM -0200, Otavio Salvador wrote: Stephen R Marenka [EMAIL PROTECTED] writes: On Wed, Feb 13, 2008 at 10:54:39AM -0200, Otavio Salvador wrote: Tomorrow we need to mass upload packages that has translation updates. All modules that has no other changes, but l10n ones, ought to be updated. If there're modules with other changes, those need to be checked with me. We'll decide together if we do or not each upload. Would partman-base and partman-partitioning fall into this somewhere? I don't plan to upload parted before beta1. Won't your changes break the installation if it doesn't gives support for m68k? They don't break anything and only affect m68k/atari, which isn't well supported atm anyway. Thanks, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: Call for Debian Installer testing, before D-I Beta1 release
On Mon, Feb 11, 2008 at 06:04:24PM -0800, Brad Boyer wrote: On Mon, Feb 11, 2008 at 01:29:33PM -0600, Stephen R Marenka wrote: On Mon, Feb 11, 2008 at 04:27:56AM -0800, Brad Boyer wrote: I did download and try to boot the m68k build of this installer. The Which initrd? cdrom and hd-media want a cd image. nativehd will download everything from the net. I'm actually not sure at this point. They all have the same file name on the server, and I didn't change the name. Did you use the kernel that was with the initrd's? Interesting, I believe the scsi cdrom used to be builtin. If that's not the case, then I need to add those modules onto the initrd. It didn't show up in the driver list in /sys, so I presume it wasn't loaded. The SCSI controller detected the CD-ROM drive and it showed up in /proc/scsi/scsi, but no /dev/sr0 or any equivalent. I did find the IDE CD-ROM driver in the appropriate directory in /sys, but this system only has SCSI. odd thing is that there doesn't appear to be a network only install. nativehd OK. I'll make sure to download that one and try it. Would it be inappropriate to suggest that this is a somewhat unintuitive choice? I'd be happy to give it a different name, please feel free to recommend one. It's very similar to netboot, but as we can't actually netboot anything I didn't want to be misleading in that direction either. There's a netinst iso that I don't want to be confused with since this initrd wouldn't use that (or any) iso. Still, it does look like we're not in as bad of shape as we thought. I'm just curious how much RAM I would need to avoid the low-memory mode, and if there's anything I'm missing because of that. It's not like we could do a graphical install on m68k. We can barely get X running after the install. It looks like lowmem level1 kicks in at 64Mb and level2 at 32Mb. I haven't pushed any of those numbers in quite a long time, so we can likely get by with less than that. That makes sense, I guess. I don't think I have any m68k systems with more than 64M of RAM total. The Q840AV could have 128M, but I haven't filled it up completely. Some of my older systems only have 16M. Is there an easy way to force it into a particular mode for testing? AFAIK, I need to build an initrd with a modified lowmem udeb. That's not really a big deal. If someone's willing to test it, I'll build it. Thanks, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: Call for Debian Installer testing, before D-I Beta1 release
On Mon, Feb 11, 2008 at 04:27:56AM -0800, Brad Boyer wrote: On Tue, Feb 05, 2008 at 03:04:31PM +0100, Frans Pop wrote: On Tuesday 05 February 2008, Eugen Paiuc wrote: Good news , m68k is reentered in d-i ! Well, the building of the daily D-I images has never really been disabled as far as I know. But http://cdimage.debian.org/cdimage/daily-builds/daily/arch-latest/m68k/iso -cd/ are forbidden for users (temporary I hope). Afraid not. As long as m68k is not in testing, we cannot build CD images (all daily/weekly CD images use debs from testing). I have now removed the links to the m68k CD images from the D-I project webpage (only thing we can do really). If there are any installation methods that don't require CD images, you could try that. You could also see if someone in the Debian m68k community would be willing to build CD images using packages from unstable. I did download and try to boot the m68k build of this installer. The Which initrd? cdrom and hd-media want a cd image. nativehd will download everything from the net. system in question is a Macintosh Quadra 840AV (64M RAM, 1G SCSI drive). The mac flavor kernel boots with no issues and sees the major hardware (internal video, keyboard/mouse, internal SCSI, internal ethernet) and the installer does start in low-memory mode. I haven't downloaded anything but the basic installer, so I haven't been able to actually complete the install. The CD-ROM method looks like it could work if the SCSI CD-ROM module was available. It looks like only IDE CD-ROM drives are supported. The Interesting, I believe the scsi cdrom used to be builtin. If that's not the case, then I need to add those modules onto the initrd. odd thing is that there doesn't appear to be a network only install. nativehd The installer sees and can configure the network, but I couldn't figure out a way to get it to pull files from it. I haven't tried a hard drive only install with the files downloaded to a local partition at this point. Still, it does look like we're not in as bad of shape as we thought. I'm just curious how much RAM I would need to avoid the low-memory mode, and if there's anything I'm missing because of that. It's not like we could do a graphical install on m68k. We can barely get X running after the install. It looks like lowmem level1 kicks in at 64Mb and level2 at 32Mb. I haven't pushed any of those numbers in quite a long time, so we can likely get by with less than that. Thanks, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: Call for Debian Installer testing, before D-I Beta1 release
On Tue, Feb 05, 2008 at 03:04:31PM +0100, Frans Pop wrote: On Tuesday 05 February 2008, Eugen Paiuc wrote: Good news , m68k is reentered in d-i ! Well, the building of the daily D-I images has never really been disabled as far as I know. But http://cdimage.debian.org/cdimage/daily-builds/daily/arch-latest/m68k/iso -cd/ are forbidden for users (temporary I hope). Afraid not. As long as m68k is not in testing, we cannot build CD images (all daily/weekly CD images use debs from testing). I have now removed the links to the m68k CD images from the D-I project webpage (only thing we can do really). If there are any installation methods that don't require CD images, you nativehd should work (I'd use suite=etch-m68k). could try that. You could also see if someone in the Debian m68k community would be willing to build CD images using packages from unstable. What's the procedure for that? At the moment, sid is pretty broken for installation (we're far behind), but no reason not to start the process. Thanks, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: partman support for m68k atari
On Sun, Jan 13, 2008 at 12:42:59AM +0100, Frans Pop wrote: (CCing Stephen as I'm not sure if he's subscribed) I'm subscribed. Recently you committed this change in r50759: +partman-base (116) UNRELEASED; urgency=low + + [ Stephen R. Marenka ] + * m68k: don't make partman uninstallable for atari However, There is still this code in partman-partitioning/lib/disk-label.sh: m68k) case $sub in amiga) echo amiga;; atari|q40) # unsupported by parted echo UNSUPPORTED;; I just updated that right before I read this email. :) | $ svn ci -m add m68k/atari support | Sendingpartman-partitioning/debian/changelog | Sendingpartman-partitioning/lib/disk-label.sh | Transmitting file data .. | Committed revision 50863. Has libparted really gained support for atari? It would surprise me as parted hasn't been updated since the release of Etch. There's a patch in 239816 which seems to work. I've been updating the rest of d-i with the hope that one day parted will include that patch or one like it. Meanwhile, I can run a full installation with my own version of parted. If it has, shouldn't the code above be updated too? Which disk label does atari use? partman-partitioning/lib$ svn diff -rPREV disk-label.sh Index: disk-label.sh === --- disk-label.sh (revision 50862) +++ disk-label.sh (working copy) @@ -69,8 +69,7 @@ amiga) echo amiga;; atari|q40) - # unsupported by parted - echo UNSUPPORTED;; + echo atari;; mac) echo mac;; *vme*) Thanks, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Bug#458965: [m68k] allow installation of etch-m68k
Package: debootstrap Version: 1.0.7 Please apply this one line patch to allow etch-m68k to be installed. Thanks, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] --- Makefile.orig 2007-11-13 19:51:03.0 -0600 +++ Makefile 2008-01-03 15:40:55.0 -0600 @@ -19,6 +19,7 @@ # no special script for etch anymore ln -s sid $(DSDIR)/scripts/etch + ln -s sid $(DSDIR)/scripts/etch-m68k ln -s sid $(DSDIR)/scripts/lenny ln -s gutsy $(DSDIR)/scripts/hardy signature.asc Description: Digital signature
Re: m68k mklibs problem?
The _rtld_global problem is confirmed fixed for m68k. Thanks! Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: m68k mklibs problem?
On Thu, Jun 07, 2007 at 04:07:56AM +0200, Bastian Blank wrote: On Wed, Jun 06, 2007 at 02:49:37PM -0500, Stephen R Marenka wrote: Please show objdump -p /lib/libc.so.6. [...] NEEDED ld.so.1 [...] And mklibs-readelf -n /lib/libc.so.6. $ mklibs-readelf -n /lib/libc.so.6 ld.so.1 -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: m68k mklibs problem?
On Tue, Jun 05, 2007 at 02:59:37PM -0500, Stephen R Marenka wrote: On Tue, Jun 05, 2007 at 09:29:49PM +0200, Bastian Blank wrote: On Tue, Jun 05, 2007 at 07:24:11AM -0500, Stephen R Marenka wrote: Is there any workaround I can do to get this working again? No idea. A not uptodate build environment? Are you saying the symbol version is missing from glibc? I don't understand. The chroot was up-to-date for m68k as of 1 June when those tests were run. m68k does not have the latest glibc built. As a follow up, the missing symbols seem to be on the initrd, which seems like it would have been stripped by mklibs? I just built an initrd using mklibs 0.1.22 and get the same results. D *UND* _rtld_global D *UND* _rtld_global_ro Thanks, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: m68k mklibs problem?
On Wed, Jun 06, 2007 at 05:50:15PM +0200, Bastian Blank wrote: On Tue, Jun 05, 2007 at 09:29:49PM +0200, Bastian Blank wrote: On Tue, Jun 05, 2007 at 07:24:11AM -0500, Stephen R Marenka wrote: Is there any workaround I can do to get this working again? No idea. A not uptodate build environment? For some reason, libc.so.6 is not listed as linked against the ld.so. So it is not added to the library list for the libc link and therefor loose the version. Please show objdump -p /lib/libc.so.6. /lib/libc.so.6: file format elf32-m68k Program Header: PHDR off0x0034 vaddr 0x0034 paddr 0x0034 align 2**2 filesz 0x0100 memsz 0x0100 flags r-x INTERP off0x000ed574 vaddr 0x000ed574 paddr 0x000ed574 align 2**0 filesz 0x000d memsz 0x000d flags r-- LOAD off0x vaddr 0x paddr 0x align 2**13 filesz 0x000f3f80 memsz 0x000f3f80 flags r-x LOAD off0x000f3f80 vaddr 0x000f5f80 paddr 0x000f5f80 align 2**13 filesz 0x46c8 memsz 0x6e28 flags rw- DYNAMIC off0x000f584c vaddr 0x000f784c paddr 0x000f784c align 2**2 filesz 0x00e0 memsz 0x00e0 flags rw- NOTE off0x0134 vaddr 0x0134 paddr 0x0134 align 2**2 filesz 0x0020 memsz 0x0020 flags r-- EH_FRAME off0x000ed584 vaddr 0x000ed584 paddr 0x000ed584 align 2**2 filesz 0x0d64 memsz 0x0d64 flags r-- STACK off0x vaddr 0x paddr 0x align 2**2 filesz 0x memsz 0x flags rw- Dynamic Section: NEEDED ld.so.1 SONAME libc.so.6 INIT0x1ce84 FINI_ARRAY 0xf5f80 FINI_ARRAYSZ 0x4 HASH0x154 STRTAB 0xbf54 SYMTAB 0x3424 STRSZ 0x5589 SYMENT 0x10 PLTGOT 0xf792c PLTRELSZ0xfc PLTREL 0x7 JMPREL 0x1cb50 RELA0x1292c RELASZ 0xa224 RELAENT 0xc VERDEF 0x12644 VERDEFNUM 0x14 VERNEED 0x128fc VERNEEDNUM 0x1 VERSYM 0x114de RELACOUNT 0xd24 Version definitions: 1 0x01 0x0865f4e6 libc.so.6 2 0x00 0x0d696910 GLIBC_2.0 3 0x00 0x0d696911 GLIBC_2.1 GLIBC_2.0 4 0x00 0x09691f71 GLIBC_2.1.1 GLIBC_2.1 5 0x00 0x09691f72 GLIBC_2.1.2 GLIBC_2.1.1 6 0x00 0x09691f73 GLIBC_2.1.3 GLIBC_2.1.2 7 0x00 0x0d696912 GLIBC_2.2 GLIBC_2.1.3 8 0x00 0x09691a71 GLIBC_2.2.1 GLIBC_2.2 9 0x00 0x09691a72 GLIBC_2.2.2 GLIBC_2.2.1 10 0x00 0x09691a73 GLIBC_2.2.3 GLIBC_2.2.2 11 0x00 0x09691a74 GLIBC_2.2.4 GLIBC_2.2.3 12 0x00 0x09691a76 GLIBC_2.2.6 GLIBC_2.2.4 13 0x00 0x0d696913 GLIBC_2.3 GLIBC_2.2.6 14 0x00 0x09691972 GLIBC_2.3.2 GLIBC_2.3 15 0x00 0x09691973 GLIBC_2.3.3 GLIBC_2.3.2 16 0x00 0x09691974 GLIBC_2.3.4 GLIBC_2.3.3 17 0x00 0x0d696914 GLIBC_2.4 GLIBC_2.3.4 18 0x00 0x0d696915 GLIBC_2.5 GLIBC_2.4 19 0x00 0x0963cf85 GLIBC_PRIVATE GLIBC_2.5 20 0x00 0x0b792650 GCC_3.0 Version References: required from ld.so.1: 0x0d696911 0x00 22 GLIBC_2.1 0x0963cf85 0x00 21 GLIBC_PRIVATE private flags = 0: -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: m68k mklibs problem?
On Sat, Jun 02, 2007 at 11:59:43PM +0200, Bastian Blank wrote: On Sat, Jun 02, 2007 at 04:24:14PM -0500, Stephen R Marenka wrote: Please find attached. Yeah. | D *UND* _rtld_global vs. | DO *UND* 05ae GLIBC_PRIVATE _rtld_global The symbol version is lost. Bastian Is there any workaround I can do to get this working again? Thanks, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: m68k mklibs problem?
On Tue, Jun 05, 2007 at 09:29:49PM +0200, Bastian Blank wrote: On Tue, Jun 05, 2007 at 07:24:11AM -0500, Stephen R Marenka wrote: Is there any workaround I can do to get this working again? No idea. A not uptodate build environment? Are you saying the symbol version is missing from glibc? I don't understand. The chroot was up-to-date for m68k as of 1 June when those tests were run. m68k does not have the latest glibc built. Thanks, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: r47134 - trunk/installer/build/config/m68k
On Mon, Jun 04, 2007 at 03:38:14PM +0200, Frans Pop wrote: On Sunday 03 June 2007 16:10, Stephen Marenka wrote: Author: smarenka Date: Sun Jun 3 14:10:24 2007 New Revision: 47134 Log: [m68k] drop keymaps that aren't at, start simplifying [...] +++ trunk/installer/build/config/m68k/cdrom.cfg Sun Jun 3 14:10:24 -EXTRANAME = cdrom/ +EXTRANAME = cdrom- [...] +++ trunk/installer/build/config/m68k/hd-media.cfg Sun Jun 3 14:10:24 -EXTRANAME = hd-media/ +EXTRANAME = hd-media- Author: smarenka Date: Mon Jun 4 12:50:12 2007 New Revision: 47139 [...] +++ trunk/installer/build/config/m68k/nativehd.cfg Mon Jun 4 -EXTRANAME = nativehd/ +EXTRANAME = nativehd- I'm not sure that these changes are a good idea. Currently, debian-cd looks in the cdrom subdir for the initrd and is thus currently broken. With the new name, debian-cd would probably have to rename the initrd too because the cdrom- prefix is rather meaningless on a CD. Did you have a particular reason for these changes? IMO it would be better, also for consistency with other arches, to keep the subdirs. No particular reason other than it seemed silly to have a directory containing one file. I'll put it back. I'm fine with consistency -- I was sort of striving for a bit of that with my changes anyway. I sort of forgot about debian-cd, although I don't believe it's currently building anything for m68k. I hope to look at that soonish. Thanks, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: m68k mklibs problem?
On Sat, Jun 02, 2007 at 09:40:23AM +0200, Bastian Blank wrote: On Fri, Jun 01, 2007 at 10:07:48AM -0500, Stephen R Marenka wrote: I fired up the latest daily for m68k and ran into the following problem. | /bin/sh: symbol lookup error: /lib/libc.so.6: undefined symbol: _rtld_global Is this a mklibs problem? Any ideas? This symbol is defined in the dynlinker /lib/ld.so.1. Can you please compare the files in the image and the system with objdump -T? No difference, see attached. Thanks, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] /lib/ld.so.1: file format elf32-m68k DYNAMIC SYMBOL TABLE: 1d0c ld .text .text 00011314 ld __libc_freeres_fn __libc_freeres_fn 00011370 ld .rodata .rodata 00014318 ld .eh_frame_hdr .eh_frame_hdr 00014374 ld .eh_frame .eh_frame 00017060 ld .data .data 000176b4 ld __libc_subfreeres __libc_subfreeres 000176b8 ld .bss .bss 00017060 gDO .data 05ae GLIBC_PRIVATE _rtld_global d384 gDF .text 00f2 GLIBC_PRIVATE _dl_make_stack_executable 000177a8 gDO .bss 0004 GLIBC_2.1 __libc_stack_end 000177d4 gDO .bss 0004 GLIBC_2.4 __stack_chk_guard f87c w DF .text 00c6 GLIBC_2.0 __libc_memalign f942 w DF .text 0018 GLIBC_2.0 malloc gDO *ABS* GLIBC_2.1 GLIBC_2.1 000177a0 gDO .bss 0004 GLIBC_PRIVATE __libc_enable_secure gDO *ABS* GLIBC_2.4 GLIBC_2.4 f9dc w DF .text 001c GLIBC_2.0 calloc bc3a gDF .text 0008 GLIBC_PRIVATE _dl_debug_state 000176b8 gDO .bss 0004 GLIBC_PRIVATE _dl_argv gDO *ABS* GLIBC_2.0 GLIBC_2.0 gDO *ABS* GLIBC_PRIVATE GLIBC_PRIVATE 00017614 gDO .data 00a0 GLIBC_PRIVATE _rtld_global_ro f95a w DF .text 0082 GLIBC_2.0 realloc 7092 gDF .text 0194 GLIBC_PRIVATE _dl_rtld_di_serinfo c788 gDF .text 01fa GLIBC_2.1 _dl_mcount 000176bc gDO .bss 0004 GLIBC_PRIVATE __pointer_chk_guard 000177ec gDO .bss 0014 GLIBC_2.0 _r_debug gDO *ABS* GLIBC_2.3 GLIBC_2.3 f83c w DF .text 0040 GLIBC_2.0 free /mnt/lib/ld.so.1: file format elf32-m68k DYNAMIC SYMBOL TABLE: 1d0c ld .text .text 00011314 ld __libc_freeres_fn __libc_freeres_fn 00011370 ld .rodata .rodata 00014318 ld .eh_frame_hdr .eh_frame_hdr 00014374 ld .eh_frame .eh_frame 00017060 ld .data .data 000176b4 ld __libc_subfreeres __libc_subfreeres 000176b8 ld .bss .bss 00017060 gDO .data 05ae GLIBC_PRIVATE _rtld_global d384 gDF .text 00f2 GLIBC_PRIVATE _dl_make_stack_executable 000177a8 gDO .bss 0004 GLIBC_2.1 __libc_stack_end 000177d4 gDO .bss 0004 GLIBC_2.4 __stack_chk_guard f87c w DF .text 00c6 GLIBC_2.0 __libc_memalign f942 w DF .text 0018 GLIBC_2.0 malloc gDO *ABS* GLIBC_2.1 GLIBC_2.1 000177a0 gDO .bss 0004 GLIBC_PRIVATE __libc_enable_secure gDO *ABS* GLIBC_2.4 GLIBC_2.4 f9dc w DF .text 001c GLIBC_2.0 calloc bc3a gDF .text 0008 GLIBC_PRIVATE _dl_debug_state 000176b8 gDO .bss 0004 GLIBC_PRIVATE _dl_argv gDO *ABS* GLIBC_2.0 GLIBC_2.0 gDO *ABS* GLIBC_PRIVATE GLIBC_PRIVATE 00017614 gDO .data 00a0 GLIBC_PRIVATE _rtld_global_ro f95a w DF .text 0082 GLIBC_2.0 realloc 7092 gDF .text 0194 GLIBC_PRIVATE _dl_rtld_di_serinfo c788 gDF .text 01fa GLIBC_2.1 _dl_mcount 000176bc gDO .bss 0004 GLIBC_PRIVATE __pointer_chk_guard 000177ec gDO .bss 0014 GLIBC_2.0 _r_debug gDO *ABS* GLIBC_2.3 GLIBC_2.3 f83c w DF .text 0040 GLIBC_2.0 free signature.asc Description: Digital signature
Re: m68k mklibs problem?
On Sat, Jun 02, 2007 at 11:01:00PM +0200, Bastian Blank wrote: On Sat, Jun 02, 2007 at 10:10:14AM -0500, Stephen R Marenka wrote: No difference, see attached. Okay. Please show the same from libc.so.6. Please find attached. Thanks, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] initrd-libc.dump.bz2 Description: Binary data di-libc.dump.bz2 Description: Binary data signature.asc Description: Digital signature
m68k mklibs problem?
I fired up the latest daily for m68k and ran into the following problem. | /bin/sh: symbol lookup error: /lib/libc.so.6: undefined symbol: _rtld_global Is this a mklibs problem? Any ideas? Thanks, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: m68k mklibs problem?
On Fri, Jun 01, 2007 at 05:20:50PM +0200, Frans Pop wrote: On Friday 01 June 2007 17:07, Stephen R Marenka wrote: I fired up the latest daily for m68k and ran into the following problem. | /bin/sh: symbol lookup error: /lib/libc.so.6: undefined symbol: | _rtld_global Is this a mklibs problem? Any ideas? It could be. There have been changes in libc/gcc which broke library reduction on most other arches. Bastian has switched mklibs to use a different (custom) readelf. Do you already have the latest mklibs installed on your buildd? Does the problem get solved if you revert to the previous version of mklibs (0.1.20)? Please make sure to test that if it does work with the old mklibs, that the image actually works, as that currently is _not_ the case for alpha. If the new mklibs does not work for m68k, then please provide the output of a build (for a single target) with two extra '-v' added in the mklibs call in the Makefile (look for MKLIBS). Cheers, FJP P.S. Could you also have a look at the space problem for the m68k floppy images? I'll run some tests and then look at the floppy problem. Thanks! Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: m68k mklibs problem?
On Fri, Jun 01, 2007 at 05:20:50PM +0200, Frans Pop wrote:
On Friday 01 June 2007 17:07, Stephen R Marenka wrote:
I fired up the latest daily for m68k and ran into the following
problem.
| /bin/sh: symbol lookup error: /lib/libc.so.6: undefined symbol:
| _rtld_global
Is this a mklibs problem? Any ideas?
Do you already have the latest mklibs installed on your buildd?
Does the problem get solved if you revert to the previous version of
mklibs (0.1.20)?
Yes, the failed version was with 0.1.21.
Please make sure to test that if it does work with the old mklibs, that
the image actually works, as that currently is _not_ the case for alpha.
Downgrading to 0.1.20 causes the following ftbfs.
| calling mklibs-readelf --print-symbols-undefined
./tmp/cdrom/tree/bin/cdrom-checker
| Traceback (most recent call last):
| File /usr/bin/mklibs, line 363, in ?
| needed_symbols.update(undefined_symbols(obj))
| File /usr/bin/mklibs, line 123, in undefined_symbols
| name, weak_string = line.split(' ', 2)
| ValueError: need more than 1 value to unpack
| make[3]: *** [stamps/tree-cdrom-stamp] Error 1
| make[2]: *** [_build] Error 2
| make[1]: *** [build_cdrom] Error 2
| make: *** [rebuild_cdrom] Error 2
However, it used to work. I downloaded an initrd from 7 May (0.1.21 was
installed to the archive on May 9) and verified that it at least lets me
chroot without the undefined symbol error (it also boots farther with
aranym). I can do a full boot test if you think it will help.
If the new mklibs does not work for m68k, then please provide the output
of a build (for a single target) with two extra '-v' added in the mklibs
call in the Makefile (look for MKLIBS).
http://people.debian.org/~smarenka/cdrom-1.21.log.bz2
P.S. Could you also have a look at the space problem for the m68k floppy
images?
There seems to be a /usr/share/acorn directory, but I haven't figured
out where it came from yet. It didn't look like the acorn keymaps were
installed. I'm sure I can drop the mac keymaps. I'll have to check on the
amiga and atari ones.
Thanks,
Stephen
--
Stephen R. Marenka If life's not fun, you're not doing it right!
[EMAIL PROTECTED]
signature.asc
Description: Digital signature
Re: m68k mklibs problem?
On Fri, Jun 01, 2007 at 09:49:11PM -0500, Stephen R Marenka wrote: On Fri, Jun 01, 2007 at 05:20:50PM +0200, Frans Pop wrote: P.S. Could you also have a look at the space problem for the m68k floppy images? There seems to be a /usr/share/acorn directory, but I haven't figured out where it came from yet. It didn't look like the acorn keymaps were installed. I'm sure I can drop the mac keymaps. I'll have to check on the amiga and atari ones. Errr. Hmmm. It appears that we're getting all kinds of keymaps, including acorn, on what should be an atari only floppy. Hopefully that will fix it. Peace, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: linux-kernel-di-m68k-2.6 help
On Wed, Nov 29, 2006 at 03:27:22PM +0100, Frans Pop wrote: On Wednesday 29 November 2006 04:02, Stephen R Marenka wrote: I was wondering if someone can help me out. I started to update linux-kernel-di-m68k-2.6 to 2.6.18 and I ran into the following error. I suppose something changed since I last did this, any ideas? Looks like this is a minor error in the control.sub file: remove the blank line above the line for XS-Vcs-Svn and the build should go OK. That's it! Thanks, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
linux-kernel-di-m68k-2.6 help
I was wondering if someone can help me out. I started to update
linux-kernel-di-m68k-2.6 to 2.6.18 and I ran into the following error. I
suppose something changed since I last did this, any ideas?
Thanks,
Stephen
$ kernel-wedge build-all
| dpkg-buildpackage: source package is linux-kernel-di-m68k-2.6
| dpkg-buildpackage: source version is 0.79
| dpkg-buildpackage: source changed by Stephen R. Marenka [EMAIL PROTECTED]
| dpkg-buildpackage: source version without epoch 0.79
| fakeroot debian/rules clean
| kernel-wedge gen-control debian/control
| dh_testdir
| dh_clean `find modules -type l`
| dpkg-source -ICVS -I.svn -I{arch} -I.arch-ids -b linux-kernel-di-m68k-2.6
| dpkg-source: error: per-package paragraph 1 in control info file is missing
Package line
--
Stephen R. Marenka If life's not fun, you're not doing it right!
[EMAIL PROTECTED]
signature.asc
Description: Digital signature
Re: [D-I] mass kernel udeb update and preparations for RC1
On Fri, Sep 22, 2006 at 02:31:43PM +0200, Frans Pop wrote: (Reply-to set to debian-boot; please only add relevant port if needed.) /me wonders why there have been almost no reactions to this mail The first part is mostly information (though a cool or thanks would be appreciated), but the second part has some issues that need attention. Have D-I porters actually read the mail? Is it useful that I send such mails at all? I was happy to hear about it. About the only reponse I have is thanks! :) Stephen (for m68k) On Sunday 17 September 2006 14:28, Frans Pop wrote: Dear (d-i) porters, First mass upload of kernel udebs = Today I have uploaded kernel udeb updates to 2.6.17-9 _for all arches_. This is the first time using the 'massbuild' [1] script I wrote recently. Effectively this means that d-i porters won't really have to worry anymore about updating kernel udebs after uploads by the kernel team. Only if the kernel major/minor changes will I request porters to do the upload themselves. For stable releases (including ABI changes) I intend to do these mass builds and do the uploads myself. Hopefully this will help the speed with which kernel udebs are updated and allow you all to spend more time testing d-i ;-) Of course porters are still responsible for maintaining which modules will be included for each arch/flavor. If you have changes between kernel major/minor releases you can either commit them and upload, or commit them as UNRELEASED and they will be automatically included in the next mass build. The massbuild script can be used for single-arch builds too. Its main advantage is that kernel images don't need to be installed and the certainty that the correct kernel version will be used. Feel free to contact me to help you get started. Some comments on today's upload: - I have used the last released version of kernel-wedge and will normally do that in the future too - I have not really checked or tested the udebs [2], so there could be some surprises; please be alert for them - m68k: I had to update the dependencies from kernel-image to linux-image The road to RC1 === We are slowly moving towards RC1. I plan to post an initial planning later this week. As we get closer to Etch, testing the installer for all arches gets to be more important. Any time you can spend on that is very much appreciated. There are some issues that need attention: * type of initrd used Some arches have already switched to using initramfs for d-i initrds, other arches are still using cramfs or ext2. Please check if a change could/should be made for your architecture. * 2.4 support now officially dropped Starting with RC1 d-i will no longer support 2.4 based installations. All arches have been switched now and some cleanup has been started; more cleanup is expected and this may cause unexpected breakage. * support for non-devfs device names Colin Watson has committed a series of changes to make d-i support non-devfs device names. We will be slowly moving away from using devfs names, but the most intrusive work will be postponed until after Etch. Please check for unexpected breakage though. * partman-auto using LVM and crypto partman-auto-lvm now has been available for some time, but is still not available for all arches. LVM support is a prerequisite for partman-auto-crypto support which will be uploaded soon. Note: swap on LVM should be possible now and is even required for partman-auto-crypto. If you would like to add support for it, please see [3]. Feel free to contact me or David Härdeman (Alphix) for help. * mips: keyboard issues We've had a report about a dead keyboard on installation (#382983). This needs to be investigated. * powerpc: oldworld boot problems with recent kernels If there are other architecture specific issues that we should be aware of, please let me know. Cheers, FJP [1] http://svn.debian.org/wsvn/d-i/people/fjp/massbuild?op=filerev=0sc=0 [2] The script does have a number of sanity checks though. [3] http://lists.debian.org/debian-boot/2006/01/msg01054.html -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: r38030 - trunk/installer/build/config/m68k
On Mon, Jun 12, 2006 at 05:16:31PM -0400, Joey Hess wrote: I hope there's a linux-kernel-di-m68k-2.6 upload pending using the -2 kernels, as it's still at -1 in svn... Already in the archive. Rats, I failed to commit my changes. Cluebat applied, changes committed. (Must've had my brain fried at the beach.) -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: A web-based frontend for the cdebconf
On Sun, Jun 11, 2006 at 05:14:38PM +0200, Attilio Fiandrotti wrote: I would like to have some feedback from you: do you think this frontend module could be somehow useful and is worth to be furtherly developed? is there anyone interested in working on it? I think it sounds pretty cool. m68k would certainly benefit from it, whereas I doubt we'll ever get gtk working on all m68k subarchs. -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: removing 2.4 from etch/sid
On Thu, May 18, 2006 at 12:24:57AM +0200, Frans Pop wrote: I could also understand if m68k would keep a 2.4 kernel for a while longer. As it is not a release arch, it might be acceptable for release managers to keep a m68k 2.4 kernel in testing/unstable. 2.2 should probably be dropped completely. Whether we will be able to keep 2.4 support in d-i if all other architectures have dropped it is uncertain though [1]. The m68k porters just had a discussion about this and we're fine with whatever ya'll decide to do. We can always keep some unofficial kernels and installers around if necessary. In other words, don't worry about holding on to any kernels just for us. I hope to keep some 2.4 and 2.2 support in d-i until we get everything moved, but I'll try to keep it out of the way or in a branch. -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Bug#360074: installation-report: ipw2200 wireless driver badconfigured
On Thu, Mar 30, 2006 at 10:21:19PM +0200, Vicente Benjumea wrote: After reboot, the driver seems working nicely, however, I had to manually run # iwconfig eth2 essid MY_ESSID mode Managed key MY_HEX_KEY to make it run, and then #dhclient How can I configure to make all these things automatically at boot time (or automatically when the wireless switch is enabled)? man 7 wireless man 5 interfaces Is how you configure it for when the interface goes up (or down). So something like the following in /etc/network/interfaces. iface eth2 inet dhcp wireless-essid MY_ESSID wireless-mode Managed wireless-key MY_HEX_KEY If you're using resolvconf, you'll also need entries like dns-nameservers and dns-search. -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
d-i beta2 other images
On the d-i homepage, the link to other images under beta 2 points to ftp.debian.org, which is either currently broken or at least doesn't have m68k (or much else) any more. -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: D-I Etch Beta2 - Status update (2)
On Wed, Feb 22, 2006 at 11:46:21PM +0100, Frans Pop wrote: Today the 20060218 builds have been installed for all arches, so tomorrow Technically still not true, as the m68k build continues to sit in byhand. -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: D-I Etch Beta2 - Status update (2)
On Mon, Feb 27, 2006 at 03:29:38PM +0100, Frans Pop wrote: On Monday 27 February 2006 14:39, Stephen R Marenka wrote: On Wed, Feb 22, 2006 at 11:46:21PM +0100, Frans Pop wrote: Today the 20060218 builds have been installed for all arches, so tomorrow Technically still not true, as the m68k build continues to sit in byhand. I'm aware of that, but as the only difference between 20060217 and 20060218 is a build dep fix that was needed for alpha and 20060217 is linked from current for m68k, I did not think that was worth bothering about. That would have been different if 20060218 had been the final build for Beta2 of course. Okay, just wanted to be sure we're on the same page. I haven't been doing to well with that lately, I guess. -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
m68k/atari: new atari-bootstrap and daily d-i builds
The new atari-bootstrap is in unstable. I'd appreciate it if some atari folks would try it out and make sure it works. If there aren't any problems, then I'd like to apply to have it included in sarge if possible, since it allows more machines to boot. The daily d-i builds [1] are also being built with the new atari- bootstrap for your testing pleasure. [1] http://people.debian.org/~smarenka/d-i/images-m68k/daily/ Thanks, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: release update and branching
On Tue, Nov 30, 2004 at 09:19:15PM -0500, Joey Hess wrote: - Some problems with the CDs, I belive the amiga kernel is missing, there may be some other mising things. May or may not be fixed in debian-cd CVS. The amiga kernel fix was orignally commited to an automatically generated file and lost. Needs followup. Rats. (Why are automatically generated files in cvs?) Anyway, thanks to manty for fixing my fix. m68k/amiga kernels are on the daily sarge cds, so I think this is fixed. -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Bug#283680: wishlist: allow file-preseed and network-preseed on same ramdisk
Package: preseed Version: 1.01 Severity: wishlist I'd like to be able to install both file-preseed and network-preseed udebs on the same ramdisk. I did this and got the following warning while building the ramdisk, with the net result that file-preseed vanished from the ramdisk. | Selecting previously deselected package network-preseed. | (Reading database ... 370 files and directories currently installed.) | Unpacking network-preseed (from udebs/network-preseed.udeb) ... | dpkg - warning, overriding problem because --force enabled: | trying to overwrite `/bin/preseed_command', which is also in package file-preseed | dpkg - warning, overriding problem because --force enabled: | trying to overwrite `/usr/lib/prebaseconfig.d/05preseed', which is also in package file-preseed | dpkg - warning, overriding problem because --force enabled: | trying to overwrite `/lib/preseed/preseed.sh', which is also in package file-preseed | (Noting disappearance of file-preseed, which has been completely replaced.) Thanks, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: announcing debian-installer release candidate 2
m68k rc2 errata: * m68k/amiga only: debian-cd wasn't updated to latest cvs for rc2, which means the netinst iso doesn't include the kernel for amiga. This means the q40 kernel gets installed, no amiga modules get installed, and the reboot may fail or come up without a network or such. -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Bug#281813: [m68k] [20041711] [cdrom] Failure with Debian installer on a Macintosh IIci
On Wed, Nov 17, 2004 at 11:38:28PM +0100, Bastian Appenzeller wrote: Method: Booted the Sarge Netboot-CD using kernel vmlinuz-2.25-mac and initrd.22 from the /cdrom directory. Kernel and ram disk were not from the netboot-cd but from S. Marenkas page. No network connection was used. Problems: The installation did not finish correctly, there was an error with mounting or unmountig partitions. However, I switched to console used the poweroff Can you describe this further? Do you remember what the message was? I did not come to the part where to set the root password and create a user. When bootin in the new installed system the root password was nothing (just pressing the enter key). But when in the system, everything seems to be OK. So you were never prompted to setup the timezone, root password, or anything after it came back up? H. I missed a man page for partman. For what were you looking? When reading README.html of the Netbood CD the links to tools is wrong (it goes to \tools and not to \install\tools). Okay, I'll have a look at that. The Mac graphic card runs in Linux without a problem. It is connected to pc monitor. Not having a mac monitor I did not remove it. Where did you get this information? After having some strange installation problems, I have the feeling that - the Linux root partition must be previous to the swap on the hard disk - the order of the boot parameter in Penguin does matter. Maybe I did not find this in the manual. No reason that partitions should be in a particular order. What kind of strange problems did you have? With no in-depth Linux and Mac experience I have to say: The Debian installer is great! It was much harder to install MacOS on the machine than Debian Linux afterwards. Cool! Thanks for your feedback. -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Bug#246877: about your Debian installation report
On Thu, Nov 18, 2004 at 06:51:25PM +0100, Christian T. Steigies wrote: The StartInstall scripts on the cd do not find the vmlinuz file nor the initrd files, simple because the path is not correct in the scripts. I can discuss that with Stephen. I thought Geert had me straight on this. Please correct the following line. amiboot-5.6 -d -k vmlinuz -r //cdrom/initrd22.gz root=/dev/ram ramdisk_size=15000 debian-installer/framebuffer=false With that fixed, the installation runs fairly smoothly, I did not get an IP with DHCP, but hostname and domain where set. This is probably due to my local network with two DHCP servers running. My freshly installed server does not have a complete config file yet, whereas my old desktop has. Do you mean a dhcp config? have been installed, and not the modules from kernel-image-amiga. And even worse, on the CD I have a lot of kernel-image debs, but there is no kernel-image for the amiga, so I can not install this to get a working network. This is a major problem, if the net were working, I could send a proper installation report... If the kernel image was there, you'd probably have modules too. I'll go look at debian-cd. The only other slight problem I noted where two segfaults shown on the third console, I think they happend when the partitioner was started, but I did not have any problems creating linux partitions on the new disk. I'm seeing those too, but I haven't figured out what that's about yet. Thanks, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Bug#268792: Unicode (UTF-8) bug still present and failure to download libc6-udeb
On Thu, Oct 28, 2004 at 11:31:29PM +0200, Erik C.J. Laan wrote: Stephen R Marenka wrote: I've tried to document all the magic in http://d-i.alioth.debian.org/manual/en.m68k/index.html. But, I tried finding the string initrd22 in the document from the URL above. In chapter 3.3.1.1 in the document I found reference to a .txt and a .pdf version. The .txt version gave me a 404, so I downloaded the .pdf version. My acrobat reader could not find the string in there. Next I checked all HTML documents, also to no avail. So it looks like the magic is not yet in the manual ;-). May I suggest chapter 4.4 Preparing Rats, you're right! Okay, that's on the TODO. P.S. The installation crashed with an oops, and the second attempt has stalled. Both at trying to write the partition table. Low-mem troubles? Likely. I'll retry at a later time, and update the bugreport accordingly. Thanks, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Bug#268792: Unicode (UTF-8) bug still present and failure to download libc6-udeb
On Thu, Oct 28, 2004 at 10:50:44AM +0200, Erik C.J. Laan wrote: As per Joey's request I've downloaded the daily build of the debian installer initial ramdisk image, and the kernel, and retried the installation. Exactly which image from which day? Thanks, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Bug#268792: Unicode (UTF-8) bug still present and failure to download libc6-udeb
On Thu, Oct 28, 2004 at 08:25:03PM +0200, Erik C.J. Laan wrote: Stephen R Marenka wrote: On Thu, Oct 28, 2004 at 10:50:44AM +0200, Erik C.J. Laan wrote: As per Joey's request I've downloaded the daily build of the debian installer initial ramdisk image, and the kernel, and retried the installation. Exactly which image from which day? Hi Stephen, I used your image downloaded about 10 hours ago from: http://people.debian.org/~smarenka/d-i/images-m68k/2004-10-27/nativehd/initrd.gz. Please try the initrd22.gz image in the same directory. It is intended for 2.2 kernels, whereas the initrd.gz kernel is intended for 2.4 kernels. I'll check that image on my box too. Thanks, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Bug#268792: Unicode (UTF-8) bug still present and failure to download libc6-udeb
On Thu, Oct 28, 2004 at 09:27:57PM +0200, Erik C.J. Laan wrote: Erik C.J. Laan wrote: Stephen R Marenka wrote: On Thu, Oct 28, 2004 at 08:25:03PM +0200, Erik C.J. Laan wrote: Stephen R Marenka wrote: On Thu, Oct 28, 2004 at 10:50:44AM +0200, Erik C.J. Laan wrote: As per Joey's request I've downloaded the daily build of the debian installer initial ramdisk image, and the kernel, and retried the installation. Exactly which image from which day? Hi Stephen, I used your image downloaded about 10 hours ago from: http://people.debian.org/~smarenka/d-i/images-m68k/2004-10-27/nativehd/initrd.gz. Please try the initrd22.gz image in the same directory. It is intended for 2.2 kernels, whereas the initrd.gz kernel is intended for 2.4 kernels. I'll check that image on my box too. OK, will do. I was already wondering that the -22 was about. Hi Stephen, Magic hint that. The Unicode trouble is gone, and it has just gotten past retreiving libc6-udeb en unpacking libc6-udeb successfully. I'll update the installation report after the installation has finished completely (probably tomorrow morning Central European Time). Excellent! I've tried to document all the magic in http://d-i.alioth.debian.org/manual/en.m68k/index.html. Please feel free to give it a read and give me some feedback. Thanks for testing, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: udebs that need upload (non-i386)
On Wed, Oct 20, 2004 at 02:09:15PM -0400, Joey Hess wrote: The following udebs that cannot be build on i386 need to be uploaded soon if they're to make the release: partitioner (Code change is required for mips, tested by ths. m68k mips or s390) Apparently someone already gotten this one. -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: d-i floppy builds requring root
On Wed, Oct 13, 2004 at 01:42:37PM -0400, Joey Hess wrote: d-i has just failed to build on the m68k autobuilder because the vme_boot target tries to use sudo. We cannot rely on there being root access to build the installer; I don't know how the last build worked, just lucky I guess. I wasn't aware that there was an m68k autobuilder that didn't have sudo. Interesting and rats! We have about 1 week to resolve this and to get the final d-i builds. Maybe you can come up with a better patch than this one? I think 250578 can be fixed pretty easily, but no one's done it and at this point we'd be changing base again. Index: floppy.cfg === --- floppy.cfg(revision 23089) +++ floppy.cfg(working copy) @@ -258,5 +258,6 @@ $(atari_root) -TARGET = $(TEMP_INITRD) $(BVMEROOT) $(BVMEBOOT) $(MVME147BOOT) $(MVME147ROOT) $(MVME16XBOOT) $(MVME16XROOT) $(BVMECDROM) $(ATARIBOOT) $(ATARIROOT) $(ATARITOOLS) - +TARGET = $(TEMP_INITRD) $(BVMECDROM) $(ATARIBOOT) $(ATARIROOT) $(ATARITOOLS) +# vme boot floppies need root to build, so cannot be built by default +# $(BVMEROOT) $(BVMEBOOT) $(MVME147BOOT) $(MVME147ROOT) $(MVME16XBOOT) $(MVME16XROOT) If you want to punt vme, so be it. I'd rather see the release than have me hold it up for anything. Hopefully, we can catch it in a point release. -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: adding an outdated warning to the installation manual for some arches
On Wed, Oct 06, 2004 at 07:43:26PM -0400, Joey Hess wrote: What other architectures besides i386 and powerpc are updated well enough to avoid such a warning? I think m68k is in pretty good shape. -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: Final status of Debian Installer string freeze
On Wed, Oct 06, 2004 at 04:22:09PM -0400, Joey Hess wrote: partitioner (m68k mips or s390) (has code changes, should be tested before upload) vmelilo-installer (m68k) done. -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
rootskel and #269077
I've just committed a workaround for #269077 to rootskel. It only affects 2.2 kernels. It should be properly fixed, but now probably isn't the time. Would someone please upload? Thanks, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: Introduction and questions about a failure while calling dpkg-buildpackage linux-kernel-di-powerpc-2.6-0.71
On Tue, Sep 21, 2004 at 12:01:59AM +0200, Niklaus Giger wrote: What did I miss? Do have to copy manuall /boot/vmlinuz-2.6.8-pegasos to /lib/modules/2.6.8-pegasos ? Is there somewhere a newer version of linux-kernel-di-powerpc-2.6? Any hints would be greatly appreciated. linux-kernel-di packages an existing kernel in a udeb. I think you're expected to have the kernel package specified in kernel-versions installed on your system. Short of that, you probably can place the kernel and modules in the same type place with the same names. -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: status of d-i 2.4.27/2.6.8 kernel transition
On Fri, Sep 10, 2004 at 09:22:57AM -0400, Joey Hess wrote: m68k 2.2.25/2.4.27 2.2.25/2.4.26 [4] [4] m68k build not using new udebs, or is it? The common-kernels file has 2.4.27, but the other .cfgs mention 2.4.26. Mess.. Sorry for the mess. I haven't removed all the old stuff from the m68k reorg yet. 2.2.25/2.4.27 is correct. common-kernels is the only place need to change kernel versions now. -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Please Help with m68k Debian-Installer Testing
First, I'd like everyone to followup to [EMAIL PROTECTED] That's what I'll be doing. Debian-installer is working on releasing rc2 and m68k has recently undergone a major reorganization (dropping build times by 2/3's). We now have support for amiga, atari, bvme6000, mac, mvme147, mvme16x, and q40/q60. Further, we're going to be releasing with 2.4.27 instead of the better- tested, but security-vulnerable 2.4.26. Daily ramdisk builds are at http://people.debian.org/~smarenka/d-i/images-m68k/daily/. Daily iso's are at http://cdimage.debian.org/pub/cdimage-testing/sid_d-i/m68k/current/. The d-i home page is at http://www.debian.org/devel/debian-installer/. The Installation Guide has also been updated, particularly chapter 5, please read over it and tell me what you'd like to see fixed (both in chapter 5 and elsewhere). http://d-i.alioth.debian.org/manual/en.m68k/index.html. Minimum RAM seems to be about 24MB, although 2.4.x kernel users may get away with less (let me know either way). Now would be way better than later to get things fixed. CAVEATS: 1) 2.2.25 does support ext3, although I've had some reports that ext3 doesn't work right on vme. 2) atari-bootstrap needs an update and we're working on it. If the current doesn't work, please try http://people.debian.org/~smarenka/d-i/bootloaders/atari/ataboot.ttp. It allows booting linux on AfterBurner040, CT2, and CT60 accelerator cards with small ST-RAM, allows loading the ramdisk to TT/FastRAM, fixes Load-to-TTRAM GEMDOS header flag and more. (Thanks Petr!) 3) q40 hasn't been tested. I haven't even told Richard Zidlicky that I'm building images yet. No bootloader or documentation. I'm slack. Hopefully we can get that cleaned up soon. All comments welcome. Thanks, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Bug#269077: serial check
On Wed, Sep 01, 2004 at 10:16:36PM +0200, Wouter Verhelst wrote: Do you really need to delete the extra vc devices? Yes; because busybox init currently checks whether the devices are there (whether the files exist), and if they are, tries to open them. Repeatedly. Issuing an error every time it fails (which, of course, happens every time). Removing the files once those error messages start to appear only changes the error that appears; and with those messages going over the screen, the usefulness of d-i is quite, uh, low. This obviously really is a bug in busybox init, hence my previous suggestion to try and fix that one. I wonder if it's not because inittab has respawn for vc/3 vc/4? Do you see an easier solution than to pass a kernel parameter? Is there one you're already passing that we could overload for 2.2 kernels (console= maybe)? -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Bug#269077: serial check
On Wed, Sep 01, 2004 at 01:36:24AM +0200, Wouter Verhelst wrote: [...serial check thinks real consoles are serial as well...] Hm. I just referenced the code which was already in rootskel, supposing it was working OK, but that doesn't seem to be the case. Checking on my VME box, however, the /proc/self/fd/* files also link to /dev/console, so there is no way this problem could be fixed by slightly modifying the check. The difference is that the original code was called by debian-installer and the new code is called in sbin/init, which has a different context. Can the following sbin/init code be moved to lib/debian-installer.d/S30term? if [ $TERM_TYPE = serial ]; then for i in 2 3 4 5; do rm -f /dev/vc/$i done fi This seems to solve the detection problem for me. Do you really need to delete the extra vc devices? -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Bug#268792: [m68k] [20040828, rc1(?)] [nativehd] Failure with Debian installer RC1 on Macintosh Quadra650
On Tue, Aug 31, 2004 at 08:07:48PM +0200, Erik C.J. Laan wrote: Mind you my setup was not with a serial console, just with debugging output send over the serial port (debug=ser console=tty0 (not ttyS0)). From what I heard serial console's on Mac's work very bad. Oh right, then this combined with it being a recent daily image means you are probably seeing bts#269077 (incorrectly detecting non-serial-console as serial console). Should be fixed soonish. The debug serial and rc1 subject line had me looking elsewhere. Glad you finally got it installed. Also worth noting is that with the d-i beta3 I tested I could not start the partitioning tool, because it needed more than 4 MB (ramdisk was 20 MB then). Now with the ramdisk_size dropped down to 13000 I have 11 MB RAM left for the installer and the partitioning tool worked, and on top of that the installation was 1,5 to 2 times faster. Progress! -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Bug#268792: [m68k] [20040828, rc1(?)] [nativehd] Failure with Debian installer RC1 on Macintosh Quadra650
On Sun, Aug 29, 2004 at 07:51:01AM +0200, Erik C.J. Laan wrote: no http proxy configured otherwise. Booted with ramdisk_size=13000 debug=ser console=tty0 added to my 'normal' Linux/mac68k bootoptions (mac53c9x=1,0). cool, a serial install on a mac. The low memory mode textbox looked fine. After that a message appeared the Unicode was set on console and from that moment on a very annoying bug appeared: Each line-drawing character in every one of the 4 borders of each text-box was replaced by the 3 character sequence of an a with a caret ^, and 2 black box-like characters. This made correctly selecting and entering information tricky. Examples: May have to setup a serial console to check this out. ext3 to ext2 at first. I need ext2 because the last stable kernel on Macintoshes is 2.2.25-mac, and ext3 support has not been compiled into that AFAIK. ext3 is indeed compiled into the 2.2.25-mac kernel (as of kernel-image-2.2.25-mac 2.2.25-4). I use it regularly. Glad you finally got it installed. -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
m68k images, opinions please
First thing, I don't want to break m68k building, so any changes might well not make sarge. That said Currently m68k generates 14 initrd's arranged by subarch, except for cdrom and cdrom22 (see [0]). I pretty much like the arrangement, everybody shares cdrom images and you look in your subarch's directory for everything else (and thus ignore everyone else's junk). But 14 initrd's take a while to build on m68k hardware and I plan to add more. It seems that most of the build time comes from building the trees. If we instead look at how many separate trees are actually needed, there would be five. All the 2.2 kernel stuff looks just like 2.4 stuff in the tree except for /lib/modules (which turns out to be minimal). The images only differ in actual image generation (same for floppy). I could now add q40 without building a new initrd, for instance. cdrom - all subarchs cdrom22 - all subarchs, 2.2 kernel hd-media- amiga, atari (not currently built) hd-media22- mac nativehd- amiga, atari nativehd22- mac (atari?) floppy atari bvme6000 mvme147 mvme16x netboot - all vme The only cost in size is /lib/modules and keymaps. The build system would also have to change to accommodate building multiple initrds out of one tree (how to do change the makefile to do that doesn't seem obvious to me). So do I go with the an image-based layout? I like the way it looks, but users will have to be in different directories to download the pieces. What other problems am I getting into? cdrom/ bvme6000-boot.img.gz initrd.gz initrd22.gz floppy/ atari-boot.img.gz atari-root.img.gz bvme6000-boot.img.gz bvme6000-root.img.gz mvm147-boot.img.gz mvm147-root.img.gz mvm16x-boot.img.gz mvm16x-root.img.gz hd-media/ initrd.gz initrd22.gz nativehd/ initrd.gz initrd22.gz atari/ bootargs bootstra.prg netboot/ initrd.gz tftplilo.bvme tftplilo.conf tftplilo.mvme tftplilo.txt kernels/ vmlinuz-2.4.26-amiga vmlinuz-2.2.25-atari vmlinuz-2.4.26-atari vmlinuz-2.2.25-bvme6000 vmlinuz-2.4.26-bvme6000 vmlinuz-2.2.25-mac vmlinuz-2.2.25-mvme147 vmlinuz-2.4.26-mvme147 vmlinuz-2.2.25-mvme16x vmlinuz-2.4.26-mvme16x tools/ rawwrite.ttp [0] Current m68k d-i directory listing. cdrom-initrd.gz cdrom22-initrd.gz amiga/ nativehd-initrd.gz vmlinuz-2.4.26-amiga atari/ bootargs bootstra.prg floppy-boot.img.gz floppy-root.img.gz nativehd-initrd.gz rawwrite.ttp vmlinuz-2.2.25-atari vmlinuz-2.4.26-atari bvme6000/ cdrom-boot.img.gz floppy-boot.img.gz floppy-root.img.gz netboot-initrd.gz tftplilo.bvme tftplilo.conf tftplilo.txt vmlinuz-2.2.25-bvme6000 vmlinuz-2.4.26-bvme6000 mac/ hd-media-initrd.gz nativehd-initrd.gz vmlinuz-2.2.25-mac mvme147/ floppy-boot.img.gz floppy-root.img.gz netboot-initrd.gz tftplilo.conf tftplilo.mvme tftplilo.txt vmlinuz-2.2.25-mvme147 vmlinuz-2.4.26-mvme147 mvme16x/ floppy-boot.img.gz floppy-root.img.gz netboot-initrd.gz tftplilo.conf tftplilo.mvme tftplilo.txt vmlinuz-2.2.25-mvme16x vmlinuz-2.4.26-mvme16x Thanks, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Bug#265955: [sjf@utvinternet.ie: Bug#265955: Amiga 4000 sarge rc1 install problems]
Ok, my problem not really with the install, but with the sarge iso. It seems to be missing some files. Previously when I tried to install with woody CD I used amiboot-5.6 to start the installer. On this CD the whole /install/amiga directory is missing. In /install/ there is root.bin, root22.bin and manifest. amiboot and linux.bin aren't there. Probably would be handy if I had these files :) Hopefully I have fixed this bug in cvs. I'll let you know when it's fixed for sure. Thanks, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: Debian-Installer release candidate 1
On Sat, Aug 07, 2004 at 11:17:22PM -0300, Joey Hess wrote: - Added support for more subarches for arm (riscpc, riscstation), m68k (bvme6000, mvme147, mvme16x), mipsel (cobalt) and powerpc (oldworld). m68k isn't currently planning on releasing with vme right now. They simply haven't been tested and apparently don't work. wouter is working on fixing them. Maybe in a point release. I would also like to rev. the kernel for atari as it fixes problems for some hardware, but I don't want to do anything to mess with the release. The new kernel is already in linux-kernel-di-m68k_0.61 and in the daily builds. The following from the checklist are okay. mac cdrom22 mac hd-media mac nativehd Thanks, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: Debian-Installer release candidate 1
On Tue, Aug 10, 2004 at 11:54:39PM +0200, Wouter Verhelst wrote: On Tue, Aug 10, 2004 at 06:49:03AM -0500, Stephen R Marenka wrote: On Sat, Aug 07, 2004 at 11:17:22PM -0300, Joey Hess wrote: - Added support for more subarches for arm (riscpc, riscstation), m68k (bvme6000, mvme147, mvme16x), mipsel (cobalt) and powerpc (oldworld). m68k isn't currently planning on releasing with vme right now. They simply haven't been tested and apparently don't work. wouter is working on fixing them. Maybe in a point release. Yes; but I had told Joey that my tests were successful (and they were). Everything seems to work perfectly, except for vmelilo-installer. That explains that. If you had told me that, I might have released them. I must add, though that I tested the netboot version, not the floppy one; I don't have a floppy for this box. I don't thing atari's floppy boot works either. Of course, I do agree that not releasing in this state is probably not a good idea anyway. Let me know when you change your mind. I've been thinking about reorganizing the build process so there aren't quite some many separate initrds to build. Of course, I hate to make that kind of change right before a release. :-\ -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Bug#262108: Debian Installation Report
Does this machine usually run linux? If so, what version of the kernel? Thanks, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Installation Manual
I was cleaning up the m68k portion of the installation manual and deleted some old entities. Unfortunately, that causes all the translations to fail to build. Should I add the entities back, even if in dummy form, or should those translations just not build for the time being? Thanks, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: Installation Manual
On Wed, Jul 28, 2004 at 01:38:57PM -0400, Joey Hess wrote:
Stephen R Marenka wrote:
I was cleaning up the m68k portion of the installation manual and
deleted some old entities. Unfortunately, that causes all the
translations to fail to build.
Should I add the entities back, even if in dummy form, or should those
translations just not build for the time being?
I ran into the same thing while overhauling the manual yesterday.
Currently
build/{rescue-images,root-tarballs,tftp-images,driver-images,driver-tarballs}.ent
are all unused except by translations and maybe some !i386 arches. I
left them in because the installation manual is built as part of the
official d-i image build process and the only other alternative was to
kill the translation build until they update.
Okay, I've removed all the old m68k entities except the three required
to build translations (in install-docs.ent with comment). It looks like
everything still builds, although some of the translations have linkend
errors (I think some of those where there before).
Thanks,
Stephen
--
Stephen R. Marenka If life's not fun, you're not doing it right!
[EMAIL PROTECTED]
signature.asc
Description: Digital signature
Re: /etc/fstab problem on S/390: which package? Userdevfs?
On Fri, Jul 23, 2004 at 12:03:00PM -0500, Adam Thornton wrote: I *think* this should go into subarch-dev in userdevfs, but I definitely would like some confirmation of that before committing any changes. userdevfs was written for systems that couldn't use devfs, 2.2 kernels for instance. I far as I can tell, m68k is the only arch using it. -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Bug#260396: PowerMac G3 (beige) sarge d-i (18 July 2004) Installation Report using 2.6 kernel
On Thu, Jul 22, 2004 at 01:31:13AM +0100, Clive Menzies wrote: On (20/07/04 10:31), Stephen R Marenka wrote: nc -l -p 2080 -e /bin/sh on tty2 or whatever suits. Thanks Stephen I used the above on tty2, which from the man page I understand means that this machine will listen for shell commands on port 2080. However, when try to contact this host from my workstation using: [EMAIL PROTECTED]:~$ nc -t Homer 2080 Homer.PRIORYROAD [192.168.1.21] 2080 (?) : No route to host Either the machine isn't really on the network or you have a routing problem. Does ifconfig show you have a configured network on Homer? What about route -n? It's good to see a default gateway. -t isn't necessary, I'm not sure what effect it would have. Can you ping 192.168.1.21? Can Homer wget some random web page? -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: Minutes of Debian Installer IRC meeting of 20040717
On Mon, Jul 19, 2004 at 05:47:34PM +1000, Anibal Monsalve Salazar wrote: m68k joeyh has got confirmation that the genext2fs bug is still there. Also, Florian Lohoff posted to -boot and reported the ext filesystem error problem on boot. I cannot confirm that this bug remains. If ya'll don't have further evidence, then this bad boy seems closed. Woo hoo! -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Bug#260396: PowerMac G3 (beige) sarge d-i (18 July 2004) Installation Report using 2.6 kernel
On Tue, Jul 20, 2004 at 03:56:12PM +0100, Colin Watson wrote: On Tue, Jul 20, 2004 at 03:44:18PM +0100, Clive Menzies wrote: again tomorrow. Meanwhile, assuming I can get a network connection what's the best way to issue the above commands from my workstation (which will allow me to cut and paste the output you've requested). Can I telnet into the machine in its half installed state - I note that I can do some stuff on tty2? I'm afraid you'll probably have to type it in on tty2. netcat is built into busybox now, so you can always go with something like nc -l -p 2080 -e /bin/sh on tty2 or whatever suits. hth, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: m68k install daily build broken ?
On Sat, Jul 17, 2004 at 06:31:11PM +0200, Florian Lohoff wrote: On Sat, Jul 17, 2004 at 05:33:57PM +0200, Florian Lohoff wrote: Hi, i tried to install on a Qudra 840av and basically it doesnt work how i try it. There is a complete lack of documentation on how to unstuff the .sit files - The documentations merely mentions the existance of penguin. I haven't even gotten to look at the documentation yet, so as far as I know, there has been no updates to the m68k documentation for d-i. Sorry. Though in the install/mac folder there are 2 penguin versions (18 and 19) and at least 4 initrds. Additionally there are root and root22 in the parent directory but no documentation on when to use what. I gather that you're using one of the cdrom images? It would be nice to no exactly what image you're using. Now i have ran penguin with linux-mac-2.2.5 with the initrd-linuxmedia.gz. Where did you get initrd-linuxmedia.gz from? If you're using a cdrom and a mac, then root22 should be the right image (22 for 2.2 kernel). Bugs are in chronologic order - EXT3-fs error - Wrong free blocks / bitmap .. after mount of ramdisk - umount /initrd/dev no such file or directory - umount /initrd invalid argument - languagechooser shows chars not displayable (utf8) - hang in archdetect - Last line in logfile on tty4 DEBUG: configure archdetect, status: 2 Okay, that's potentially worrisome. Can you tell me exactly what image you're using so I can try to reproduce it? Thanks, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: Minutes of Debian Installer IRC meeting of 20040717
On Mon, Jul 19, 2004 at 05:47:34PM +1000, Anibal Monsalve Salazar wrote: Simon Richter will look after m68k/Amiga a bit. Yeah!!! -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Bug#258098: partman disk format prompting
Package: partman Version: 40 Severity: wishlist I'd like to see better prompting when choosing to format a partition. After the usage method is set to format, prompting should be automatic for file system and mount point (ala partconf). Another option would be to default the file system and perhaps the mount point to / for the first partition formatted. Thanks, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: m68k vs. d-i tc1 -- bummer, help needed
If this was a normal package, I'd binNMU it to get the fixed images available for testing. Since I can't do that, I've put them at http://people.debian.org/~smarenka/d-i/test/tc1/. mv cdrom22-initrd-tc1-take2.gz cdrom22-initrd.gz mv nativehd-initrd-tc1-take2.gz mac/nativehd-initrd.gz Thanks, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: m68k vs. d-i tc1 -- bummer, help needed
On Wed, Jun 02, 2004 at 10:11:38AM -0300, Joey Hess wrote: Stephen R Marenka wrote: If this was a normal package, I'd binNMU it to get the fixed images available for testing. Since I can't do that, I've put them at http://people.debian.org/~smarenka/d-i/test/tc1/. mv cdrom22-initrd-tc1-take2.gz cdrom22-initrd.gz mv nativehd-initrd-tc1-take2.gz mac/nativehd-initrd.gz Well it is a normal package, as far as being bunNMUable. But do we have a fix for this that doesn't involve hand-hacking and that can work on the autobuilders? Well on my autobuilders buildd has sudo rights. I don't know if this is still policy everywhere or just a holdover. I can switch to a dd, mke2fs scheme, at least for the initrd's destined for 2.2 kernels. I think we're getting out-of-scope for a binNMU though. What would you prefer? Thanks, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: m68k vs. d-i tc1 -- bummer, help needed
On Wed, Jun 02, 2004 at 11:04:05AM -0300, Joey Hess wrote: Stephen R Marenka wrote: Well on my autobuilders buildd has sudo rights. I don't know if this is still policy everywhere or just a holdover. I can switch to a dd, mke2fs scheme, at least for the initrd's destined for 2.2 kernels. I think we're getting out-of-scope for a binNMU though. What would you prefer? I doubt that requiring root on the autobuilders is workable. I'm mostly worried about continued maintenance. It's fine if you do a bin-NMU for this, but what if I have to release a new version and you're not available? Is it possible that the recent fixes in genext2fs led to this problem on m68k? Or did we just get very unlucky? I think it's possible that the recent fixes made the problem more explicit. I had them before, but more random. I think we're a bit unlucky. I didn't use to have them at all. Long term I'd like to get genext2fs fixed. Perhaps I need to divert my d-i development time to that. I'd rather spend it on lowmem, polish, packaging, and documentation. But hey, whatever it takes. Short term, perhaps I should do a conditional build where a second set of images is built if sudo is available? It need only apply to the three 2.2-kernel images. It won't increase the build time much if I base it on the existing trees. As for build access when I'm not around (and I'll be hitting vacation time later this month), I'd be happy to give accounts on my d-i build box to interested dd's. It has both buildd chroots and my daily d-i chroot. Thanks, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
m68k vs. d-i tc1 -- bummer, help needed
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Periodically while testing daily d-i images, I've run into the following problem on m68k/mac while running a 2.2 kernel. Many of the initrd's give the following example error when mounting on a 2.2 kernel system (either loopback or as initrd). | EXT2-fs error (device loop(7,0)): ext2_check_blocks_bitmap: Wrong free | blocks count for group 0, stored = 4515, counted = 4511 | EXT2-fs error (device loop(7,0)): ext2_check_blocks_bitmap: Superblock | in group 1 is marked free | EXT2-fs error (device loop(7,0)): ext2_check_blocks_bitmap: Descriptor | block #0 in group 1 is marked free | EXT2-fs error (device loop(7,0)): ext2_check_blocks_bitmap: Block bitmap | for group 1 is marked free | | EXT2-fs error (device loop(7,0)): ext2_check_blocks_bitmap: Inode bitmap | for group 1 is marked free | EXT2-fs error (device loop(7,0)): ext2_check_blocks_bitmap: Wrong free | blocks count for group 1, stored = 5054, counted = 5050 | EXT2-fs error (device loop(7,0)): ext2_check_blocks_bitmap: Wrong free | blocks count in super block, stored = 9569, counted = 9561 Often nothing comes as a result of the errors and I used to ignore them. But recently, the permissions have started to get messed up too. Featuring the dreaded busybox message | Bummer, could not run '/sbin/debian-installer': Permission denied scrolling across the screen. It doesn't happen with all images always, but if an image is broken, it stays broken. This same image looks fine on a machine with a 2.4 kernel and even looks fine when mounted loopback on a 2.2 machine, although you get the above ext2-fs errors in dmesg. I think this may be genext2fs bug #232023. Preliminary testing seems to support this. If I create a filesystem using dd and ext2fs and copy a broken initrd to it via loopback: the new image seems to be fine. Which brings me to d-i, the mac/nativehd-initrd.gz and cdrom22-initrd.gz images are so afflicted. I plan on copying them to a fresh filesystem and testing them, but that doesn't help what's in the archive. Unfortunately, the only build workaround that I'm aware of requires root. dd mke2fs sudo mount -o loop sudo cp sudo umount gzip Ideas welcome. Thanks, Stephen - -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] -BEGIN PGP SIGNATURE- Version: GnuPG v1.2.4 (GNU/Linux) iD8DBQFAvJ8MVKM4J7YoSbMRAlBlAJ46yij5QNlQf5BzhNR8iKt9mBOfMgCgjJpn AiUUgKmNBuWI+i6PWW652Ms= =x1nb -END PGP SIGNATURE- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: m68k vs. d-i tc1 -- bummer, help needed
On Tue, Jun 01, 2004 at 07:21:16PM +0100, Colin Watson wrote: Do you have enough space and kernel support to use cramfs as a workaround? 2.2 doesn't seem to have cramfs. It also doesn't have tmpfs, so we need to be able to write to the initrd (unless someone has a better idea). -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: m68k cdrom build failure
On Fri, May 28, 2004 at 09:25:26PM +0100, Colin Watson wrote: On Wed, May 26, 2004 at 09:48:57AM -0500, Stephen R Marenka wrote: It seems to be eject-udeb. I'd file a bug, but I'm not sure what to say. You should now just need to arrange for 2.0.13deb-5.1 to be built on m68k. Looks like it's already been installed. Cool. Thanks, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: Suggestions for udeb development?
On Thu, May 27, 2004 at 10:43:23PM -0700, Matt Zimmerman wrote: I am interested in getting the evms udeb into a working state, and would appreciate any pointers regarding an efficient way to organize my development environment so that I can iteratively make changes to the udeb and test. For instance, is there a way that I can get d-i to download a udeb over the network, or from a mounted filesystem, in preference (or in addition) to those on the installation media? Pointers to any existing resources on this subject appreciated. My solutions was to build my own little mirror on the network and burn it to cd for testing. The scripts, such as they are, are at http://marenka.net/~stephen/di/bin/. Basically, debmirror will build a d-i mirror. You need a few .disk files and md5sums if you want a cd. You need to update the release file after you add in your stuff. hth, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Call for m68k/vme debian-installer testers
Unofficial debian-installer (d-i) test images and iso's are available for m68k/vme at http://people.debian.org/~smarenka/d-i/test/vme/. You need a minimum of something like 24 MB of ram to test these images. We plan on pushing that number down, but it's not there yet. The home page for d-i has pointers to the Installation HOWTO and Manual at http://www.debian.org/devel/debian-installer/. The images include cdrom (bootable only under bvme6000), tftp, and floppy (which should be bootable). All the images are setup to work with a 2.4 kernel. I'd like to know if the 2.4 kernel works, since the 2.2 kernel may get dropped from sarge. If all else fails, I can build images which work with a 2.2 kernel. To get regulary daily building support for vme, someone will need to figure out how to install vmelilo without root, see http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=250578. As far as I know, these images have never been tested and may well not work. I've completed as much of the m68k/vme port of debian-installer as I can for now. The rest is up to you. -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: 2.4 kernels
On Fri, May 28, 2004 at 06:21:32PM +0100, Colin Watson wrote: debian-installer completely relies on devfs at the moment, unconditionally. For instance, many components walk /dev/discs. Unconditionally, unless you're using userdevfs. So far userdevfs has only been used with 2.2 kernels, but it could be made to work otherwise. -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
m68k cdrom build failure
Can anyone help me out here? The cdrom images on m68k are failing to build with the following. | LOCPATH=./tmp/cdrom/tree/usr/lib/locale LC_ALL=C.UTF-8 reduce-font /usr/src/unifont.bdf ./tmp/cdrom/all.utf ./tmp/cdrom/unifont.bdf | setlocale: C.UTF-8 | FYI: MB_CUR_MAX/MB_LEN_MAX: 6/16 | error -1 at position 523495 (bytes: 16 Used chars: 1928 (523495 processed) The full log is at http://people.debian.org/~smarenka/d-i/images-m68k/daily/build_cdrom.log. Thanks, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
m68k ports update
The m68k/atari port is now building with the dailies, so both the image and iso links may be updated appropriately. The m68k/vme ports now have a test build with iso's available at http://http://people.debian.org/~smarenka/d-i/test/vme/daily/. Please update the port links. Thanks, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: m68k ports update
On Tue, May 25, 2004 at 04:24:06PM -0300, Joey Hess wrote: Stephen R Marenka wrote: The m68k/atari port is now building with the dailies, so both the image and iso links may be updated appropriately. The m68k/vme ports now have a test build with iso's available at http://http://people.debian.org/~smarenka/d-i/test/vme/daily/. Please update the port links. Here are the changes I've committed, please check. Looks good, thanks! Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: tenative timeline for rc1 release (and DebConf)
On Tue, May 18, 2004 at 02:37:52AM -0300, Joey Hess wrote: The next release of d-i should not be a beta. It should be something that we feel is suitable to be used in a real Debian release. This means that we can't afford to put out a release with significant known bugs of the kind documented in the errata, if fixes are known for the bugs. Then would someone please upload base-installer_0.81? .80 has already made it into testing and .81 fixes the m68k cdrom errata. Note that the 21-24 may time period is more important than it appears. When we're checking out the test candidate, it's essential that we cover all arches and as many installation methods as we can. An important I'll only be near my test box on the 24th, so that limits what m68k/mac testing is going to get done by me. Thanks, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: m68k/atari ports update
Unfortunately, I didn't pay enough attention to the fact that partconf uses parted also. Since no support for atari exists in parted, d-i simply can't support atari until that is fixed -- partman vs. partconf isn't even the issue. :-( http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=239816 Maybe a link to the bug should go on the ports page? Thanks, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: m68k/atari ports update
On Tue, May 18, 2004 at 03:04:53PM +0100, Martin Michlmayr wrote: * Stephen R Marenka [EMAIL PROTECTED] [2004-05-18 08:44]: Unfortunately, I didn't pay enough attention to the fact that partconf uses parted also. Since no support for atari exists in parted, d-i simply can't support atari until that is fixed -- partman vs. partconf isn't even the issue. :-( Are you sure partconf using parted matters actually matters? As far as I know, partconf uses partman in a way that the disk label doesn't matter. I could be wrong, though. It turns out that yes, I am indeed, an idiot. :-) I was just informed that partconf seems to work fine and that the problem is elsewhere. Oh well, live and learn. Thanks, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
Re: m68k/atari ports update
On Thu, May 13, 2004 at 10:52:32AM -0300, Joey Hess wrote: Stephen R Marenka wrote: Please update the ports page to show that m68k/atari nativehd, floppy, and cdrom images and iso's are at http://people.debian.org/~smarenka/d-i/test/atari/current/. I suppose they should be listed as building or test or some such. As soon as I get some confirmation that these work, I'll add them to the daily builds. Done. Please let me know when the CD builds are available from the standard place so I can update it. I think Manty said that the m68k CD builds failed BTW. I'll let you know. I don't plan on adding it to the daily or cd build until I'm sure atari works to some extent. Hopefully that's fixed now (at least they build for me). Will there be atari netboot stuff, or should I remove that from the table? Remove it for now, if someone wants it bad enough, then I'll get it working. Thanks, Stephen -- Stephen R. Marenka If life's not fun, you're not doing it right! [EMAIL PROTECTED] signature.asc Description: Digital signature
