Re: OEM-Install for Debian?
On Wed, Mar 12, 2008 at 11:58:00PM +0200, Eddy Petrișor wrote: Wolfgang Lonien wrote: Hi Matt dear DDs, Hello, I could need something like Matt Zimmermans OEM-Installer (which Ubuntu has since 2k5 or so) in Debian. For an explanation, please see https://help.ubuntu.com/community/Ubuntu_OEM_Installer_Overview This *looks* really like is Debian Installer with some additional module and some supplementary package. (Note that I said looks, but I haven't dug into the issue to see if it really is D-I). If, indeed, is a modified Debian Installer, I guess we could get the D-I module and add an option to boot with such an option. Ditto for the package(s) in the system. Matt, most likely, can confirm or infirm the assumptions above :-) . I can't take personal credit for this bit of technology, but yes, can say that yes, it is built on debian-installer. Colin Watson is the primary author, and would be a good point of contact for someone interested in helping to get this upstream. -- - mdz
Re: Fwd: [i18n] Input Method and Fonts improvements for Gutsy
On Sat, Aug 11, 2007 at 09:10:57AM -0500, Ming Hua wrote: On Sat, Aug 11, 2007 at 11:26:08AM +0100, Matt Zimmerman wrote: Ubuntu does in fact use tasks, though they aren't presented to the user by default in the installer, as in Debian. I'm curious why you feel that the distinction between main and universe is an issue here. Oh, I didn't know that Ubuntu also uses tasks. But I think most Ubuntu installation use language-support packages instead of tasks to install input methods anyway. That much is correct, yes. My understanding (which may be wrong) is that all language-support packages in Ubuntu are main packages, and therefore all their dependencies are also main packages. I think this also makes Ubuntu developers want to minimize the number/size of supported (does supported seed equal to main?) packages, leading to the decision of using SCIM for input method support for all languages. However, as I wrote, it may not be optimal from a user's point of view. language-support packages are part of the default installation (depending on the language selected), and therefore they may depend only on fully supported packages (i.e. main). It is naturally preferable to support fewer packages, if they can do the same job as a larger number of packages, but the ultimate priority is of course the user experience. If the scim methods do not work for some languages, and there exists an alternative which is maintained and supportable, then we would consider using it instead. In Debian, however, tasks can depend on (include? -- since missing dependency in tasks in not fatal) input method packages as long as they are in the archive, without going through the main inclusion process as they need to do in Ubuntu. So many tasks include more than one input method packages, and quite some of them don't use scim at all. As im-switch is essentially an alternative system, users can then use im-switch to choose the input method he/she prefers. I doubt Ubuntu will be willing to support multiple input method packages. We would prefer to standardize on one approach per language, yes, and devote our efforts to ensuring a good quality implementation of that one approach, rather than offering many alternatives. SCIM doesn't have the best support for all languages, its biggest advantage is multi-language support. SCIM also has its own shortcomings, the most infamous one being causing crash in GTK+ applications linked to libstdc++5 [1,2,3], which means firefox from mozilla.org, Adobe acroread, ATi proprietary video driver, etc. This sounds more like a bug than a shortcoming; can it not be fixed? It sounds like a dynamic linking issue, and there are folks around who know that part of the stack much better than I do. For users only using one language (or English and another language only), they have many reasons to prefer another input method. Debian's task system can support this easily. Ubuntu's language-support package system doesn't seem to be supporting this now, and I don't feel it will support this unless there are significant changes in the way input method packages are developed/tested and in the main inclusion process. Your concerns seem to have nothing to do with the language-support packages or with tasks, but rather which packages we choose to use by default to provide input method support. Our policy should be to use the best method for each language, while respecting the need for these packages to be supportable. We do not wish to ship unmaintained software, but this does not mean that we cannot consider reasonable alternatives where they exist. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Fwd: [i18n] Input Method and Fonts improvements for Gutsy
On Fri, Aug 10, 2007 at 10:33:09PM -0500, Ming Hua wrote: (Dropping pkg-fonts-devel list, as they are not likely interested in input method discussions. Adding ubuntu-devel list.) On Tue, Aug 07, 2007 at 09:05:20AM +0200, Christian Perrier wrote: Quoting Daniel Glassey ([EMAIL PROTECTED]): Hi, I think it would be good to discuss this with Debian folks at well to share their expertise and I think these issues should be addressed for lenny as well. And, given that this highly involves packages beings installed by default, this should be discussed with the D-I team as such default installations should be handled by tasksel in Debian. (please note that Ubuntu does not use tasksel and, therefore, solutions suitable for Ubuntu will, there, not be suitable for Debian and vice-versa) Since Debian doesn't have the constraint of the main/universe separation, Debian can use a very different approach than Ubuntu's. And AFAIK, on CJK front, etch already has a rather good input method support in default desktop task installation. Ubuntu does in fact use tasks, though they aren't presented to the user by default in the installer, as in Debian. I'm curious why you feel that the distinction between main and universe is an issue here. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#389881: RC-ness of this bug
On Wed, Mar 07, 2007 at 10:50:46AM -0300, Otavio Salvador wrote: peter green [EMAIL PROTECTED] writes: I don't know how invasive those changes might be. AFAIK Ubuntu already does it (Colin?) and wouldn't be too hard to pick the changes from them but we would also need RM and Frans approval :( ubuntu already does what? there are four possible soloutions proposed aren't there (labels in fstab and the 3 different /dev/by-* trees) labels on fstab IIRC. Ubuntu uses mount by UUID, not labels. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#279994: Installation report for Ubuntu
On Sat, Nov 06, 2004 at 04:03:09PM +0100, Danilo Piazzalunga wrote: Package: installation-reports INSTALL REPORT Debian-installer-version: Ubuntu 4.10/Warty Warthog Final (2004-10-20) How did you come to send an Ubuntu installation report to Debian? Is there some documentation or software that we need to update in order to avoid this? The correct place to send Ubuntu installation reports is [EMAIL PROTECTED] -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#261778: Detects two video cards when the system has only one
Package: discover1 Version: 1.6.1 Severity: normal debian:/home/mdz# discover video Toshiba America Info Systems 601 S3 Inc. ViRGE/MX The latter is my video card. The former is actually a Host bridge device: :00:00.0 Host bridge: Toshiba America Info Systems 601 (rev a2) Subsystem: Toshiba America Info Systems: Unknown device 0001 Flags: bus master, medium devsel, latency 0 -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#258307: Create ubd devices for UML
Package: base-installer Severity: wishlist Tags: patch With this patch, base-installer will create the necessary ubd devices so that UML can boot from the installed system. They are not created by debootstrap or makedev by default, so they will not exist unless explicitly created. Behaviour with the patch is only different if the installation is happening under UML. http://www.no-name-yet.com/patches/base-installer.ubd.diff -- System Information: Debian Release: unstable APT prefers unstable APT policy: (500, 'unstable'), (1, 'experimental') Architecture: i386 (i686) Kernel: Linux 2.6.6-deb1-skas3 Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#258136: Allow reboot action to halt instead
Package: di-utils-reboot Severity: wishlist Tags: patch I find it convenient sometimes (especially when using UML) to have d-i halt the system as its final step, rather than rebooting it. Rebooting in UML re-execs the kernel with exactly the same command line, which causes the installer to be started, rather than booting from the installed system. I could also see this being useful for embedded systems, where it is sometimes common to perform the first stage of installation and then stop, shipping an image which will be configured by the end user after flashing it into a device such as a PDA. Anyway, here's a patch: http://www.no-name-yet.com/patches/debian-installer-utils.reboot-halt.diff -- System Information: Debian Release: unstable APT prefers unstable APT policy: (500, 'unstable'), (1, 'experimental') Architecture: i386 (i686) Kernel: Linux 2.6.6-deb1-skas3 Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#258136: Allow reboot action to halt instead
On Wed, Jul 07, 2004 at 05:32:44PM -0400, Joey Hess wrote: Matt Zimmerman wrote: Anyway, here's a patch: http://www.no-name-yet.com/patches/debian-installer-utils.reboot-halt.diff So the idea is to set di-utils-reboot/halt at the boot line or similar, and never really ask the question? Yes; perhaps there is a better way to do that, but this seemed to make sense to me. This way, it can trivially be set as a boot parameter, and also via debconf. It's a lot like kbd-chooser's debian-installer/serial-console, so I assumed this was a reasonable implementation. Is it OK? -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#258176: Support for UML ubd devices
Package: libdebian-installer Severity: wishlist Tags: patch With this patch, di_system_devfs_map_from can generate correct mappings for UML ubd devices. http://uml-goodness.no-name-yet.com/patches/libdebian-installer.ubd.diff -- System Information: Debian Release: unstable APT prefers unstable APT policy: (500, 'unstable'), (1, 'experimental') Architecture: i386 (i686) Kernel: Linux 2.6.6-deb1-skas3 Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#257383: User-mode Linux support
Package: kbd-chooser Severity: wishlist Tags: patch This patch enables kbd-chooser to detect that it is running under UML, and allows configuration to proceed without errors. The UML console doesn't require keyboard configuration. This makes testing debian-installer under UML more convenient. http://uml-love.no-name-yet.com/kbd-chooser.uml.diff -- System Information: Debian Release: unstable APT prefers unstable APT policy: (500, 'unstable'), (1, 'experimental') Architecture: i386 (i686) Kernel: Linux 2.6.6-deb1-skas3 Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#230630: SCSI HD, installation with bootfloppy and usb flash stick, i386
On Sat, Jun 05, 2004 at 11:41:52AM +0200, Bastian Blank wrote: On Fri, Jun 04, 2004 at 05:35:54PM -0700, Matt Zimmerman wrote: I am interested in getting it working, but d-i development seems rather awkward to me. So far, I know of no alternative to burning a CD-RW over and over. hmm - use apt-ftparchive to build your own archive and use netboot - use qemu/bochs/vmware with cd images aren't that enough? I asked on debian-boot for suggestions for udeb development using a network repository, and got only responses suggesting CD-RW. Is there a document somewhere which describes the setup? -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#230630: SCSI HD, installation with bootfloppy and usb flash stick, i386
On Tue, May 25, 2004 at 04:54:09PM +0100, Martin Michlmayr wrote: The evms module in the installer does not work. Yeah, evms support in debian-installer is not very well tested/developed at all. LVM will work, though. I am interested in getting it working, but d-i development seems rather awkward to me. So far, I know of no alternative to burning a CD-RW over and over. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#252164: Package: installation-reports
On Wed, Jun 02, 2004 at 09:11:11PM -0700, [EMAIL PROTECTED] wrote: I just today discovered the IBM donation of EVMS to the LINUX world. I'll bet I also need to select evms-udeb? I have not been because I didn't know what EVMS is. If not required, what does evms-udeb add? evms-udeb is incomplete. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Suggestions for udeb development?
On Fri, May 28, 2004 at 01:16:10PM +0100, Colin Watson wrote: 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? Personally I use the 'monolithic' build target type, which lets you build all necessary udebs into the CD-ROM image. If you have CD-RWs, this works out well. Does that help? Actually, that is sort of the situation I want to avoid: re-burning a CD-RW each time I want to test a change. I would rather be able to modify a network repository of udebs, if that is possible, and have the existing media grab my new udeb. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Suggestions for udeb development?
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. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#248399: dhcp-found dns servers not placed in resolv.conf after default install
On Tue, May 11, 2004 at 09:48:35PM +1000, Anthony Towns wrote: On Tue, May 11, 2004 at 11:46:48AM +0100, Colin Watson wrote: I don't know why the installer is installing dhcp-client rather than dhcp3-client. I'll reassign this to debian-installer. That's easy: dhcp3-client didn't used to exist. (If dhcp3-client is to replace dhcp-client, why isn't it just being called dhcp-client, with dhcp2-client kept around if it's really needed?) Because the debian-installer folks asked the dhcp*-client folks to leave things as they are for sarge. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#248399: dhcp-found dns servers not placed in resolv.conf after default install
On Tue, May 11, 2004 at 05:23:06PM +0200, Thomas Hood wrote: On Tue, 2004-05-11 at 17:10, Matt Zimmerman wrote: Because the debian-installer folks asked the dhcp*-client folks to leave things as they are for sarge. So can should debootstrap switch to installing dhcp3-client? I expect not, but ask debian-boot. In order for dhcp-client to be updated to v3, the various upgrade issues (dhcp-* v2 - dhcp-* v3, dhcp3-* - dhcp-* v3) need to be worked out, and I think it's a bit late in the game for that anyway. We asked about it some time ago and were told to leave it alone for d-i's sake. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#248399: dhcp-found dns servers not placed in resolv.conf after default install
On Tue, May 11, 2004 at 09:34:45AM -0700, Scott Webster wrote: Perhaps I'm missing something here, but doesn't this mean that the default install with d-i is now broken by using both dhcp-client and resolvconf? Maybe. Doesn't this affect everyone using dhcp? No? It certainly works fine for me (I don't use resolvconf). -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Full Debian install impressions and facts (another one)
On Fri, Apr 16, 2004 at 11:25:42AM +0200, Christian Perrier wrote: On this install, the noflushd problem didn't happen. So I did the whole install at high priority. Thus, the debconf screens count is exact. Package name Useful Useless Total fr-Translated Should use d-i [...] libssl0.9.70 1 1 1 This question should not be asked on a new install of libssl0.9.7, only on upgrades. Looking at the postinst, it seems correct: if [ $1 = configure ] then if [ ! -z $2 ]; then if dpkg --compare-versions $2 lt 0.9.7d-1; then -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Full Debian install impressions and facts (another one)
On Mon, Apr 19, 2004 at 12:37:42PM -0400, Joey Hess wrote: Matt Zimmerman wrote: This question should not be asked on a new install of libssl0.9.7, only on upgrades. Looking at the postinst, it seems correct: It's possible that an upgrade was involved in the install, if the CD had a slightly older version. if [ $1 = configure ] then if [ ! -z $2 ]; then if dpkg --compare-versions $2 lt 0.9.7d-1; then Surely that $2 should be quoted for safty's sake. I agree. As-is, though, the code seems to work as expected, and so I don't know why that debconf note is displayed. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#240331: apt: CDROMs must be manually ejected and mounted on Apple/Mac/PowerPC
On Sat, Mar 27, 2004 at 05:58:05AM -0700, toff wrote: On Fri, Mar 26, 2004 at 08:35:27PM -0800, Matt Zimmerman wrote: On Fri, Mar 26, 2004 at 09:01:30PM -0700, toff wrote: On Fri, Mar 26, 2004 at 01:31:50PM -0800, Matt Zimmerman wrote: You had to mount the CD in order to eject it? Yes, if I don't mount it first I get a message like 'unable to eject, last error: inappropriate ioctl for device' That sounds completely backwards. You should not be able to eject a mounted CD; you should have to unmount it first. All mounting and unmounting is handled by base-config, not apt, in this case. If you want the CD to be automatically ejected, that would need to be done in base-config as well. I see you reassigned it. Yet the same problem comes up using aptitude, apt-get and apt-cdrom. If what you say is true, it sounds like a kernel problem. -- - mdz Is there something I can do to pin it down further? Would strace help? I don't know; ask debian-powerpc. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#240331: apt: CDROMs must be manually ejected and mounted on Apple/Mac/PowerPC
On Fri, Mar 26, 2004 at 09:01:30PM -0700, toff wrote: On Fri, Mar 26, 2004 at 01:31:50PM -0800, Matt Zimmerman wrote: You had to mount the CD in order to eject it? Yes, if I don't mount it first I get a message like 'unable to eject, last error: inappropriate ioctl for device' That sounds completely backwards. You should not be able to eject a mounted CD; you should have to unmount it first. All mounting and unmounting is handled by base-config, not apt, in this case. If you want the CD to be automatically ejected, that would need to be done in base-config as well. I see you reassigned it. Yet the same problem comes up using aptitude, apt-get and apt-cdrom. If what you say is true, it sounds like a kernel problem. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#238653: debian-installer network config evms
On Thu, Mar 18, 2004 at 09:05:43AM +0100, Stefan Tibus wrote: - The evms-udeb links /lib/libevms-2.3.so.0 to libevms.so instead of libevms-2.3.so.0.0. Furthermore libncurses(5?) is missing and the EVMS module for debian-installer does not work at all. Fixing the two library issues makes evmsn run (though it does not really work, due to lacking kernel support), the d-i module remains broken. I haven't had time to finish this udeb; if there is some way to hide it by default without removing it entirely, I would like to do that. Is the d-i kernel really lacking device-mapper? That has been a standard part of Debian kernel-source since 2.4.23, and is needed for lvm2 as well (which I believe has a working udeb). -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
s390 installation support in Debian 'sarge'
s390 fans: As you can see from the attached debian-installer status report, s390 is one of a minority of Debian architectures which are currently not supported by debian-installer. Part of the reason for this is that access to create and reboot Linux/390 guests (as would be needed for installation testing) is hard to come by for most Debian developers. Without access to a development platform, it is unlikely that we will be able to support installation of Debian on s390 in the Debian 'sarge' release. I would like to ask that if any of you are in a position to provide Debian developers with access to such a system, and are willing to do so, please contact either the installer development mailing list ([EMAIL PROTECTED]) or myself if you would prefer private mail ([EMAIL PROTECTED]) to work out the details. Feel free to forward this message to anyone else who might be in a position to help. -- - mdz ---BeginMessage--- The debian-boot team is working hard to produce a decent next generation installer for sarge, and we're making good progress. But in many areas we don't have enough people to do all the work. We're reaching the point where if some things are not done soon, they will simply not be supported in time for the release of sarge. This email lists some well-defined jobs, many of them easy for developers who are unfamiliar with the internals of the installer to take on, that can significantly improve debian-installer. 1. installation report processing With the release of betas 1 and 2 of d-i, we've gotten many installation reports from users. These tend to collect many problems, glitches, bugs, and observations into one bug report, and so they need to be processed, cloned off into separate bugs which are retitled and assigned to the right packages in debian-installer. We have over 150 of these that need processing. We are hoping to process at least 75 of these installation reports this week, with the help of the larger group of Debian developers, and so we have written a tutorial that should get developers quickly up to speed on processing installation reports. You can find it here: http://cvs.alioth.debian.org/cgi-bin/cvsweb.cgi/debian-installer/doc/installation-reports.txt?rev=1.2content-type=text/x-cvsweb-markupcvsroot=d-i 2. The installation manual is half complete, and we need writers familiar with XML to work on the unfinished sections, as well as update and proofread what's already there. The manual has its own TODO list, here: http://cvs.alioth.debian.org/cgi-bin/cvsweb.cgi/debian-installer/doc/manual/TODO?rev=1.1content-type=text/x-cvsweb-markupcvsroot=d-i 3. PCMCIA Do you have more than two PCMCIA cards? Have you ever edited config.opts by hand? Then you're more qualified than anyone on the d-i team to get our PCMCIA support working. Currently it fails about 75% of the time. Stealing the parts of pcmcia-cs's postinst that figure out what module to use for the bridge would be a good first step. 4. low memory support d-i barely supports installations on systems with 32 mb of memory. It's unlilkely to ever support lesser systems unless someone steps up to work on it. We have some ideas, that should work, but no time. 5. arm, s/390, sparc These are our lagging architectures, and if d-i is to support installing them for sarge, we need at least one person working on each. Currently we have none. hppa, m68k, and mipsel are a bit further along, but also need more developers. 6. lintian/linda checks for udebs debian-installer uses udebs, which are small, non-policy compliant packages in deb format. Actually we have our own minimal policy[1] for udebs, which prohibits them having preinst scripts, conffiles, pre-depends, conflicts, documentation, etc. It's kind of the anti-policy, and we occasionally screw up and put something into a udeb we should not. So we hope to get support in lintian or linda to check udebs for these problems. 7. graphical boot screen We would like to drag Debian kicking and screaming into the .. er, late 90's by giving its installer a fancy graphical boot screen. We have two candidates, but would be glad to see something even better. Note that it's limited to 16 colors and 640x480 resolution, and see the syslinux documentation for details. 8. ppp support If you care about ppp support for the first stage install, you need to put some time into getting it working in d-i. Otherwise, it just won't happen. 9. everything else Here is the rest of our short list for the next release. While we're working on everything in here, it's likely that at least some of it will not happen in time, unless we get more developers. - fix all beta2 errata (a must) - security fixed, 2.4.24 kernel (with SATA support) (done for stock i386 (but not SATA, probably?)) - discover 2 - partman (fixes many issues
Re: debian-installer help wanted
On Sat, Jan 31, 2004 at 01:43:29PM -0500, Joey Hess wrote: - security fixed, 2.4.24 kernel (with SATA support) (done for stock i386 (but not SATA, probably?)) I actually installed woody (bf24) onto a system with only SATA disks a few days ago. It worked, but things were a little strange (trickle of unexpected interrupt messages on the console), and was sometimes unstable. Upgrading to kernel-image-2.4.24-i386 from unstable cleared everything up, and the system has been rock solid ever since, so I think that if you use a kernel at least that recent, SATA should be fine. - better wireless support wireless-tools-udeb is available, but is not loaded by default on netinst CD no easy configurator yet (Joshua Kwan) Isn't the usual problem with wireless that the drivers for most cards must be built from source because they aren't in the standard kernel? Are we going to try to provide binary modules? We still need more people testing d-i on more strange and wonderful hardware. See http://www.debian.org/devel/debian-installer for details. Is there any hardware that particularly needs testing? -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: debian-installer help wanted
On Sat, Jan 31, 2004 at 06:22:49PM -0500, Joey Hess wrote: Matt Zimmerman wrote: Isn't the usual problem with wireless that the drivers for most cards must be built from source because they aren't in the standard kernel? Are we going to try to provide binary modules? Between orinoco and orinoco_pci and a few others, the kernel actually has reasonably good coverage. linux-wlan-ng only supports a few oddball adaptors, and some USB ones that arn't supported by the stock kernel. hostap doesn't support much not in the stock kernel either. Of course this ignores all the newer stuff that has only binary module support. OK, I'll take your word for it. I only have experience with two wireless adapters, and both of them required third-party drivers (linux-wlan-ng for a prism-based PCMCIA card whose model I don't recall, and madwifi for a Netgear PCI card). -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Debconf Templates Style Guide
On Fri, Oct 31, 2003 at 07:43:09PM +0100, Henning Makholm wrote: Scripsit Matt Zimmerman [EMAIL PROTECTED] Of course it would, and I would never recommend doing so. So instead you're recommending an approach that forces the user to choose *before* reconfiguring between a) not being told what the maintainer thought was a sensible default b) not being told how he has currently configured the package I don't see why that would help anyone. You aren't making sense. You said that the problem was that the user wanted the safe defaults and couldn't tell what they were. I provided an idea for a solution, which was to give the user an option to forget their current configuration and confirm the safe defaults. Now you're complaining that the user has too much choice. Which is not terribly helpful to a user who wants to make an informed choice *between* the safe default and his own prior customizations. Writing a paragraph of text attempting to tell the user what the safe default is, without making reference to any UI-specific widgets, is a waste of time and space. Changing the debconf interface for the purpose of creating a horrific UI which attempts to present the user with three sets of options for each possible question would be equally silly. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Debconf Templates Style Guide
On Tue, Oct 28, 2003 at 11:11:13PM +0100, Henning Makholm wrote: There used to be, somewhere, a guideline that told maintainers to let themselves be inspired by the descriptions in the kernel source's make fooconfig, especially with regard to telling the user what the conservative default choice is. Many of the kernel option descriptions do indeed say If unsure, answer No or the like. Or do I misremember? If we used a sentence like this, it would read If unsure, accept the default, which would be redundant. The default choice should always be what the user wants if they are unsure. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Debconf Templates Style Guide
On Wed, Oct 29, 2003 at 03:52:27AM +0100, Henning Makholm wrote: That would be a horrible sentence to put into a debconf description. Of course it would, and I would never recommend doing so. The default choice should always be what the user wants if they are unsure. I'm afraid you need to redesign the entire debconf system then. No, you don't. You just need a command line option to dpkg-reconfigure which says to forget the current/previous configuration. This is a bit shy of redesigning the entire system. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Bug#211397: Comments on this bug
On Sat, Oct 18, 2003 at 10:49:50AM +0900, Junichi Uekawa wrote: I'm using rootstrap as an almost in-place replacement for debootsrtap in pbuilder. It's working fine as it is; rootstrap is scoped to create Debian root filesystem as debootstrap does. I would consider it to be an added bonus if debian-installer could be ran to configure the system. (or is running base-config after rootstrap enough?) I think base-config is what would be needed. I do not know whether base-config is currently in a state where it can be used noninteractively, or to what extent the database would need to be seeded ahead of time. I'm copying the base-config maintainer, but if you are interested in seeing this implemented, you might also want to look at base-config and see what would be required. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#212072: Build-depends cannot be met in unstable
Package: efi-reader Version: 0.2 (not installed) Severity: grave Build-Depends: debhelper (= 4.1.16), libcdebconf-dev, libdebian-installer3-dev, po-debconf (= 0.5.0), iso-codes (= 0.012) libcdebconf-dev does not exist. -- System Information: Debian Release: unstable Architecture: i386 Kernel: Linux mizar 2.4.21-evms2.1.0-skas-3 #1 Thu Jul 17 09:01:34 EDT 2003 i686 Locale: LANG=en_US, LC_CTYPE=en_US -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
[pere@hungry.com: Bug#210275: evms-udeb: Incorrect menu item value?]
How do I fix this? - Forwarded message from Petter Reinholdtsen [EMAIL PROTECTED] - Date: Wed, 10 Sep 2003 14:59:38 +0200 From: Petter Reinholdtsen [EMAIL PROTECTED] Resent-From: Petter Reinholdtsen [EMAIL PROTECTED] To: [EMAIL PROTECTED] Subject: Bug#210275: evms-udeb: Incorrect menu item value? Package: evms-udeb Version: 2.1.1-1 When installing the evms udeb into debian-installer, it is positioned at the very top of the main menu, before Choose language. This is not good, as it end up as the default menu entry. It should be further down in the menu, probably around the location of the partitioning tools. - End forwarded message - -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [pere@hungry.com: Bug#210275: evms-udeb: Incorrect menu item value?]
On Thu, Sep 11, 2003 at 12:59:34AM +0200, Petter Reinholdtsen wrote: [Matt Zimmerman] How do I fix this? You change this line in the debian/control file: XB-Installer-Menu-Item: # Change it to some sensible value based on the content in CVS file debian-installer/doc/menu-item-numbers.txt, and make a new upload. lvmcfg uses value 44. I suspect evms-udeb should use the range 40-50. That file seems to list a specific value for evms (47), so I've changed it to that in 2.1.1-2. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Bug#137560: apt: build-deps not resolved
On Sun, Aug 03, 2003 at 09:55:00AM +0100, Alastair McKinstry wrote: On Sun, 2003-08-03 at 08:47, Matt Zimmerman wrote: The following packages have unmet dependencies: slang1-pic: Depends: slang1-dev (= 1.4.5-2.1) but it is not going to be installed E: Broken packages zsh: exit 100 sudo apt-get install slang1-pic libnewt-dev A separate issue is that you are building boot-floppies in a sid environment (presuming this is deliberate, for the moment; if you are building for 3.0r2, then do so in woody) I was only trying to reproduce the bug, not trying to actually build boot-floppies. Thanks, though. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#140579: Report: tftpboot install successfull
On Sun, Dec 29, 2002 at 08:48:48PM +0100, nb wrote: Thanks to Michael R. Schwarzbach and he's post I finally can install debian via tftp. I think the need is going growing to do that and it's really time to update documentation. I've spent 24 hours to do that and sentences like NOT YET WRITTEN are not acceptable today. I hope this will go better. Since you have spent much time on this, and been successful at it, it seems that perhaps you are better qualified to write that documentation than most. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
d-i testing
My test machine was a Dell Dimension 4100 with an IDE HDD and 3c59x NIC. I used this floppy image: f710d54705563e73a4e74b4bc86d6e9d net-1440.img From http://people.debian.org/~tfheen/d-i/ What worked: - The floppy booted and the frontend started - The NIC was discovered - I was able to statically configure the network - I was able to select a mirror, and the udebs were downloaded This is as far as I was able to get within the installer. At this point, I had to use a shell to try to make further progress. What almost worked: - After running depmod and modprobe, I was able to see the HDD. I'm not sure at which point (or if) this was supposed to happen automatically. What did not work: - DHCP. The error message seems to indicate that the kernel does not contain the necessary support. - fdisk. I get a symbol relocation error on bindtextdomain, and fdisk would not start. cfdisk would not start because libslang was not present. - This was my first opportunity to try the EVMS udebs that I hastily assembled. evms would not run due to lack of libreadline, and evmsn would not run due to lack of libncurses. I didn't have kernel support anyway, of course. How are dependencies satisfied in this environment? -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: d-i testing
On Wed, Oct 16, 2002 at 06:16:31PM -0400, Matt Zimmerman wrote: My test machine was a Dell Dimension 4100 with an IDE HDD and 3c59x NIC. I used this floppy image: f710d54705563e73a4e74b4bc86d6e9d net-1440.img From http://people.debian.org/~tfheen/d-i/ OK, I'm a moron. I missed the option to enter the mirror location manually, so I wasn't getting the packages from people.d.o/~tfheen/. Those work much better. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: d-i testing
On Wed, Oct 16, 2002 at 06:28:55PM -0400, Matt Zimmerman wrote: OK, I'm a moron. I missed the option to enter the mirror location manually, so I wasn't getting the packages from people.d.o/~tfheen/. Those work much better. More specifically, everything that worked on the previous attempt still worked, and fdisk also ran, after I selected the option to detect disks (way down at #7?). I didn't get any farther than that due to the lack of a spare partition on the machine, which I thought it had. I'll try again on another system in the near future. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Partition tools (Re: debian-installer status -- 2002-07-29)
On Mon, Aug 19, 2002 at 05:42:30PM -0700, Jim Lynch wrote: On Mon, 29 Jul 2002 18:46:41 -0400 Matt Zimmerman [EMAIL PROTECTED] wrote: I am in complete agreement that we would not want to include EVMS, XFS or similar in our default kernel unless (or until) they are part of the official kernel. Please find the reason for EVMS not being incorporated. Also, is LVM not going to be part of the kernel in the future? I'm not totally sure about this part, but I thought I had read that Linus wants LVM out; note that there have been fairly nasty core-level bugs in LVM in the recent past (the last one I knew of involved main stack overflow causing big filesystem problems: I recall patching to kill that particular bug on my personal machine). Pay special attention to the word 'default' in my sentence above; it is the core of my opinion on this subject. I believe that the default kernel should be sufficiently generic to enable a wide variety of users to install the system, but not so overfeatured as to be difficult to support, or to introduce too many unknowns into the installation process. There is no reason why these systems should not be options if someone is willing to do the work, but I would caution against their inclusion in the default kernel. From what I know about LVM and EVMS, the latter's development is being carried out by people who understand the kernel; the former, not so much: some of the LVM developers had a falling out with others; the group who understood kernel issues were harping on technical issues. They got kicked off the mailing list for doing so. Hence, they are left with people who don't understand the kernel quite as much. I don't keep up with LVM, or with kernel politics, and I am not qualified to judge anyone's kernel prowess, but as far as I know, LVM is no longer being actively developed, and current development is focused on LVM2. EVMS is, of course, a separate project. I propose that we measure the stability of each, and make a decision based on stability, rather than whether or not a certain module passed political muster to be included in the kernel. I would add to this that we should be very conservative in our decisions. I further propose that we take looks at these technologies more often, looking at the development status and trying things out. This is what kernel-patch packages are for. There don't seem to be any for LVM2 at the moment, but I would like to see one. There is, of course, one for EVMS. However, I am willing to do some work to support EVMS in d-i, and to provide EVMS-enabled installation media for those who are interested in trying it. That's definitely more like it... there are many who would want to try such things. It would also be good to work toward standardizing naming of volumes (err, allowing such naming of volumes to happen in the installer), and allowing the creation of EVMS, LVM, LVM2 (worth a look!) volumes. Current installer does not allow these things; I think work in these areas should begin as quickly as possible. Should the kernel VFS be extended to allow things along these lines? I had planned to try to build some boot-floppies which could be used with woody to install directly onto EVMS volumes. However, I think the time (if I find that I have it) might be better spent getting something more permanent into debian-installer. Given that I know almost nothing about d-i, this would probably be a significant investment of resources to get up to speed and implement something. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Partition tools (Re: debian-installer status -- 2002-07-29)
On Mon, Jul 29, 2002 at 08:26:51AM +0200, Eduard Bloch wrote: #include hallo.h Karl M. Hegbloom wrote on Sun Jul 28, 2002 um 11:01:42PM: Tollef Fog Heen [EMAIL PROTECTED] writes: [...], we need some more partition tools; there has been some discussion on debian-boot regarding it but no final decision has been reached, though several solutions have been presented. I wonder if the partition tools can do EVMS, or if a plugin should be made for that? Anyone else want to use EVMS at some point? I object. EVMS is nice and it is the future. But currently, I miss some features and the guys are preparing a new version with major changes. Lvm10 is known to be stable enough, EVMS is not (yet). What is the basis for your objection? Are you going to try to stop someone else from doing this development? -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Partition tools (Re: debian-installer status -- 2002-07-29)
On Tue, Jul 30, 2002 at 12:09:08AM +0200, Eduard Bloch wrote: #include hallo.h Matt Zimmerman wrote on Mon Jul 29, 2002 um 09:00:42AM: What is the basis for your objection? Are you going to try to stop someone else from doing this development? Pretty simple. The same reason as for not having XFS in the default kernel: - development stage - not in the kernel - changes lots of stuff in the kernel - not stable or has other nasty bugs - future stable version breaks things I know that you like EVMS, and I do too, but it's still beta-ware. You may argue that it will be stable to the time when Sarge is ready - but we do also release with 2.2.x as default 1.5years after 2.4 release. I am in complete agreement that we would not want to include EVMS, XFS or similar in our default kernel unless (or until) they are part of the official kernel. However, I am willing to do some work to support EVMS in d-i, and to provide EVMS-enabled installation media for those who are interested in trying it. If I can use this experience to improve volume management support in d-i, then I would hope that my contributions would be welcome. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: woody install: NFS hangs
On Thu, Jul 18, 2002 at 07:41:14PM +0200, Andrea Mennucc wrote: problems was, vga does not work. I had to boot with the option video=vga16:off Actually, I would propose to NOT USE vga for the rescue disk: this is the 3rd woody install I tried, and all had problems with vga Did you try the default (vanilla) kernel, instead of bf2.4? -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: boot cd
On Mon, Jul 01, 2002 at 08:12:00AM -0600, [EMAIL PROTECTED] wrote: What other files are required on the boot cd besides the *.iso file? The .iso is an image file which contains an entire filesystem for the CD; it should not be placed on a CD like an ordinary file, but instead used as the data track. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Woody installation on IBM 44P-170 - really close
On Wed, May 22, 2002 at 11:27:15AM -0500, Rolf Brudeseth wrote: Does anybody know where I may obtain the source code for boot-floppies 3.0.23? PowerPC if it makes a difference. The source is the same for all architectures, and can be found in /pool/main/b/boot-floppies/ on any Debian mirror. Note that you do not need this in order to make a new rescue.bin. I will have to create my own rescue.bin image. What kind of image is this? Do I need to create a true rescue image with bootloader etc, or is this really just the zImage kernel? If it is a true rescue image, I would assume that I use yaboot? I know how to make one; however, does Debian have documentation how they made theirs? This chapter in the installation manual: http://www.debian.org/releases/woody/i386/ch-boot-floppy-techinfo.en.html explains how the rescue disk works, specifically how to replace the kernel. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Load kernel - unable to mount Rescue Floppy
On Tue, May 21, 2002 at 05:33:51PM +0100, David Murphy wrote: I'd really appreciate any guidance, as I've been hitting my head against this for a couple of days now! This mailing list is for discussions regarding development of the installation system, not for technical support. Try [EMAIL PROTECTED] -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#144825: boot-floppies: invalid contents of /debian/dists/woody/main/disks-s390/3.0.22-2002-04-03/tape/parmfile.debian
On Tue, Apr 30, 2002 at 06:01:57AM -0700, Matt Kraai wrote: On Mon, Apr 29, 2002 at 09:09:41PM -0400, Matt Zimmerman wrote: On Mon, Apr 29, 2002 at 07:56:09AM -0700, Matt Kraai wrote: I've fixed this. Does this also need to be done for parmfile.example1: dasd=192-191 ro hostname=debian.domain.com ip=192.168.0.42 peer=192.168.0.1 dns=192.168.0.2 netinterface=iucv1 netmodule=netiucv iucv=tcpip:gateway and parmfile.example2: dasd=192-191 ro hostname=debian.domain.com ip=192.168.0.42 netmask=255.255.255.0 gateway=192.168.0.1 dns=192.168.0.2 netmodule=lcs chandev=noauto;lcs0,0x1000,0x1001,0,1,0 ? Yes, it can't hurt. These numbers are usually changed by the administrator to match the system being installed, but a more correct example is helpful. OK, fixed. I don't suppose you know how to handle the TDF file? TDFs are used for supplying data to an emulated tape device; the file that Jochen Hein supplied looks OK, but I haven't tested it. The pathnames typically have to be edited by the admin anyway, so it should be considered an example. It would probably do fine under the tape/ directory, by analog to vmrdr/debian.exec (which is the same kind of metadata). -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#144825: boot-floppies: invalid contents of /debian/dists/woody/main/disks-s390/3.0.22-2002-04-03/tape/parmfile.debian
On Tue, Apr 30, 2002 at 07:20:19AM -0700, Matt Kraai wrote: OK, I'll do so. What should it be named? debian.tdf would be good. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#144825: boot-floppies: invalid contents of /debian/dists/woody/main/disks-s390/3.0.22-2002-04-03/tape/parmfile.debian
On Tue, Apr 30, 2002 at 09:49:56AM -0700, Matt Kraai wrote: On Tue, Apr 30, 2002 at 05:17:07PM +0200, Stefan Gybas wrote: On real hardware yes. You can use the CD-ROM of a Multiprice 3000 as emulated tape. But it might also be useful for Hercules although you could also specifiy the indivial files that should be on the emulated tape in hercules.cnf. A TDF file might have the problem that you need a tapes directory in / when used with Hercules. Should we include an example hercules.cnf, or document it? It sounds like all of the original raised issues are moot. I don't think that it's necessary, but here is a minimal version of the one that I use for installation testing. -- - mdz # # Sample configuration file for Hercules ESA/390 emulator # CPUSERIAL 002623# CPU serial number CPUMODEL 3090 # CPU model number MAINSIZE 64# Main storage size in megabytes XPNDSIZE 0 # Expanded storage size in megabytes CNSLPORT 3270 # TCP port number to which consoles connect NUMCPU1 # Number of CPUs OSTAILOR LINUX # OS tailoring PANRATE FAST # Panel refresh rate ARCHMODE ESA/390 # Architecture mode S/370, ESA/390 or ESAME # .---Device number # | .-Device type # | | .-File name and parameters # | | | # V V V #--- # Card reader with data from the vmrdr subdirectory 000C3505bf/kernel.debian bf/parmfile.debian bf/initrd.debian autopad eof # A DASD device to install on (must be created with dasdinit) # e.g., dasdinit -bz2 debian.3390 3390 DEBIAN 500 # for a 500-cylinder, bzip2-compressed image 03003390debian.3390 # Adjust the following two lines for networking using TUN on Linux 2.2+ # deviceMTU Hercules IP Host IP Netmask 0A003088 CTCI /dev/misc/net/tun 2000 192.168.10.2 192.168.10.1 255.255.255.252 0A013088 CTCI /dev/misc/net/tun 2000 192.168.10.2 192.168.10.1 255.255.255.252
Bug#144825: boot-floppies: invalid contents of /debian/dists/woody/main/disks-s390/3.0.22-2002-04-03/tape/parmfile.debian
On Mon, Apr 29, 2002 at 07:56:09AM -0700, Matt Kraai wrote: I've fixed this. Does this also need to be done for parmfile.example1: dasd=192-191 ro hostname=debian.domain.com ip=192.168.0.42 peer=192.168.0.1 dns=192.168.0.2 netinterface=iucv1 netmodule=netiucv iucv=tcpip:gateway and parmfile.example2: dasd=192-191 ro hostname=debian.domain.com ip=192.168.0.42 netmask=255.255.255.0 gateway=192.168.0.1 dns=192.168.0.2 netmodule=lcs chandev=noauto;lcs0,0x1000,0x1001,0,1,0 ? Yes, it can't hurt. These numbers are usually changed by the administrator to match the system being installed, but a more correct example is helpful. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#144680: boot-floppies: s390 chandev configuration fails due to lack of /usr/bin/editor
On Sun, Apr 28, 2002 at 09:19:33AM -0400, Adam Di Carlo wrote: Matt Zimmerman [EMAIL PROTECTED] writes: /target/usr/sbin/modconf: /usr/bin/editor: not found [...] Can't we also fix this by setting the EDITOR environment variable correctly? Yes, that is correct and would be the most straightforward. I didn't look at modconf to see what other options there could be. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#144680: boot-floppies: s390 chandev configuration fails due to lack of /usr/bin/editor
Package: boot-floppies Version: 3.0.22-2002-0403 Severity: normal In order to set up chandev parameters, modconf runs /usr/bin/editor. This does not yet exist in the base system when modconf is run from debootstrap, causing this option to immediately return to the menu. If you are quick (or running hercules on a slow system) you can see: /target/usr/sbin/modconf: /usr/bin/editor: not found This means that in many configurations, the network will not come up after the initial reboot because chandev parameters need to be specified. This could be corrected either by temporarily creating an editor link to nano-tiny, or (?) by re-running modconf from base-config. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#144566: boot-floppies: s390 login message gives incorrect path to debootstrap.log
Package: boot-floppies Version: N/A; reported 2002-04-25 Severity: normal Tags: patch --- s390-specials/issue.old Thu Apr 25 20:02:38 2002 +++ s390-specials/issue Thu Apr 25 20:03:02 2002 @@ -1,8 +1,9 @@ -Please open a telnet session to launch the installation program. -You may open a second telnet session and issue tail -f /var/log/messages -to monitor the system log or tail -f /tmp/debootstrap.log to monitor the -installation progress. /tmp/debootstrap.log will only be created in the -final steps of the installation when the base packages are installed. +Please open a telnet session to launch the installation program. You may +open a second telnet session and issue tail -f /var/log/messages +to monitor the system log or tail -f /target/tmp/debootstrap.log to +monitor the installation progress. /target/tmp/debootstrap.log will only +be created in the final steps of the installation when the base packages +are installed. If your connection gets closed and you wish to restart the installation program please enter dbootstrap after reconnecting. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: screenshots from BFs?
On Fri, Apr 19, 2002 at 07:42:33PM +0200, Michael Bramer wrote: How can I make screenshots from the BF-system? There was a thread about this no more than a couple of days ago on -devel; a search should find it. Personally, I would use user-mode-linux, but plex86 and bochs are also reported to work. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Bochs can be helpful for bootdisk testing
On Fri, Apr 12, 2002 at 12:27:46PM +0200, Robert Millan wrote: Bochs (the x86 emulator) can be quite helpful for bootdisk testing, saving from lots of reboots. It's relatively fast and supports el Torito for booting from CD (or ISO image, no need to burn) Also there have been some major reworks to make configuration easier, so while you are at testing bootdisks you can test bochs and ensure that it'll be working well for Woody. :) It also doesn't test booting on real hardware, which is the most important test for this particular change I think. For boot-floppies, I should think that user-mode-linux would be as useful as bochs, and much, much faster. I talked a bit about testing boot-floppies with UML here: http://lists.debian.org/debian-boot/2001/debian-boot-200112/msg00583.html And I know that others have had success with it as well. As far as testing bochs, that's beneficial as well. I've only managed to get it to boot MS-DOS on my PowerPC machine, but that is an accomplishment in itself. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Bochs can be helpful for bootdisk testing
On Fri, Apr 12, 2002 at 11:49:46AM -0400, Dale Scheetz wrote: My needs are a bit eccentric ;-) I'm more interested in the process than in the final result, as the product I'm looking for is the screen shots of each step along the path to a completed installation. If this is all you need, try UML instead. There will be no framebuffer stuff, but I should think that a basic text console would be enough for your purposes. Contact me if you need any assistance in this area, and see my other message in this thread. I'm mildly interested in getting the Ethernet connection working, so I can test the network install, but without help I'm probably not up for it, even though it appears from bochs list traffic that this is possible. This should be significantly simpler under UML than bochs. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Bochs can be helpful for bootdisk testing
On Sat, Apr 13, 2002 at 11:16:32PM +0200, Christian Leber wrote: On Sat, Apr 13, 2002 at 05:05:57PM -0400, Matt Zimmerman wrote: Absolutely; UML has been able to be used for installation testing for quite some time now. Even more if initrd would be compiled in. Consider it done. (why didn't you say so earlier?) -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Please test this woody cd image
I tested the isolinux boot image on 3 random workstations (Dell and HP) as well as an IBM ThinkPad T21. All of them worked fine with both idepci and bf24 kernels. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Please test this woody cd image
On Thu, Apr 11, 2002 at 10:15:08AM +0200, Eduard Bloch wrote: #include hallo.h Matt Zimmerman wrote on Wed Apr 10, 2002 um 09:31:12PM: I have installed many SCSI systems (including the one that I'm using right now) with potato CD #1, which I assume has a similar configuration. I've Your assumption is WRONG. The old vanilla Flavor used there had lots of drivers. Idepci does not. In that case, you are making a completely fallacious argument by using idepci as justification for isolinux. That is a completely different decision. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Please test this woody cd image
On Thu, Apr 11, 2002 at 10:12:28AM +0200, Eduard Bloch wrote: Did YOU help introducing ANY new features in boot-floppies? I18n? Kernel 2.4, other filesystems, RAID support, lots of bugfixes in the existing code, etc. etc.? Either you can continue your mission of distruction, or you stop crying about lost time. You have never helped with my projects, but I still listen when you report bugs or express concerns. This is not too much to ask. If we should work out and discuss every feature (needing some weeks), then test it some months, we won't release Woody before 2003. Good idea. We should definitely stop discussing features before implementing them, and cut back on testing. I'm finished with this argument, and I'm going to go test the isolinux images. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Please test this woody cd image
On Wed, Apr 10, 2002 at 11:06:44AM +0200, Raphael Hertzog wrote: Hello everybody, release is coming soon ... and we need a bit of feedback about a new feature we plan to use on CD1 of Debian woody for i386. On CD *1*? Surely you don't plan to introduce new features in such a core component mere weeks before the scheduled release? There is no way that it could receive sufficient testing in that time. If we must have a new experimental boot feature, it should go on one of the other CDs in the set. (that said, I see no problem with making the first CD contain fewer packages so that it can fit on a 5cm CD; that sounds useful indeed) -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Please test this woody cd image
On Wed, Apr 10, 2002 at 08:33:00PM +0200, Eduard Bloch wrote: Matt Zimmerman wrote on Wed Apr 10, 2002 um 10:45:35AM: On CD *1*? Surely you don't plan to introduce new features in such a core component mere weeks before the scheduled release? There is no way that it Why not? Weeks != days. 10 days != long enough to test a completely different _primary_ way of booting the installation system. Why not use _exactly_ this on the first CD, and pure idepci on 5th CD (for the few cases where isolinux may break though I have _never_ heard about problems with isolinux). Sure, once it has been proven to work _for Debian_ on a wide variety of systems. It would have been a great idea a couple of months ago. Why did you wait until there was so much pressure to finish the release? And sorry, IMHO is idepci the worst kernel-image to be used for CD#1 as the only available flavor. It seems to work for a large number of users, and that is its only job, is it not? -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Please test this woody cd image
On Wed, Apr 10, 2002 at 11:23:26PM +0200, Eduard Bloch wrote: #include hallo.h Matt Zimmerman wrote on Wed Apr 10, 2002 um 04:45:08PM: 10 days != long enough to test a completely different _primary_ way of booting the installation system. a) isolinux is not completely new. It is syslinux, extended with ability of reading iso9660. Show me one failure (caused not by general problems, i.e. with some laptops not beeing able to boot _any_ mkisofs-made cdrom) and I will shut up. The burden of proof is on you, who want to make a change with such broad effect so late in the release cycle. I have never even tried isolinux, while I have used syslinux many times. b) You can insert another CD and boot, if you got problems. I am sure most people will do anyways. Exactly the same argument can be made for putting isolinux on the second CD. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#141102: Not release notes
On Wed, Apr 10, 2002 at 07:24:39PM -0700, Chris Tillman wrote: Does this address the issue? Index: inst-methods.sgml === RCS file: /cvs/debian-boot/boot-floppies/documentation/en/inst-methods.sgml,v retrieving revision 1.117 diff -u -r1.117 inst-methods.sgml --- inst-methods.sgml 2002/04/08 14:49:07 1.117 +++ inst-methods.sgml 2002/04/11 02:23:01 -993,6 +993,7 If you get ttSIOCSRARP: Invalid argument/tt you probably need to load the rarp kernel module or else recompile the kernel to support RARP. Try ttmodprobe rarp/tt and then try the rarp command again. +Newer kernels use an rarpd user program rather than a module. This should be in a separate paragraph, like the part about SunOS, because the method is completely different. The 'rarp' command is not used, and none of the preceding text applies. How about something like this: -- - mdz --- inst-methods.sgml.old Wed Apr 10 22:57:41 2002 +++ inst-methods.sgml Wed Apr 10 22:58:20 2002 -1020,19 +1020,25 OpenBoot tt.enet-addr/tt command, or ]] boot into ``Rescue'' mode (e.g., from the rescue floppy) and use the command tt/sbin/ifconfig eth0/tt. -p -In GNU/Linux you need to populate the kernel's RARP table. To do this -execute +p +On systems using a Linux 2.2.x kernel, you need to populate the kernel's +RARP table. To do this execute example /sbin/rarp -s varclient-hostname/var varclient-enet-addr/var /usr/sbin/arp -s varclient-ip/var varclient-enet-addr/var /example -p +p If you get ttSIOCSRARP: Invalid argument/tt you probably need to load the rarp kernel module or else recompile the kernel to support RARP. Try ttmodprobe rarp/tt and then try the rarp command again. + +p +On systems using a Linux 2.4.x kernel, there is no rarp module, and you +should instead use the rarpd program. The procedure is similar to that used +under SunOS below. + p Under SunOS, you need to ensure that the Ethernet hardware address for the client is listed in the ``ethers'' database (either in the
Re: Please test this woody cd image
On Wed, Apr 10, 2002 at 07:07:04PM -0600, Erik Andersen wrote: Other distros (such as slackware and mandrake) are already using isolinux in their installers. I don't hear too many people complaining that those distro fail to boot from CD... Do you follow those distributions' user and support mailing lists? I certainly don't, so I've no idea who can or can't install them. What I do hear quite often is the only thing that would install was Debian [and NetBSD]. A search on lists.debian.org shows: http://lists.debian.org/debian-boot/2001/debian-boot-200111/msg00016.html http://lists.debian.org/debian-boot/2001/debian-boot-200108/msg00258.html http://lists.debian.org/debian-boot/2001/debian-boot-200108/msg00260.html and nothing else ever. The couple of messages in 200108 are all musing and no code or testing. Last August would have been a good time to start experimenting with this if it was intended for woody. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Please test this woody cd image
On Wed, Apr 10, 2002 at 05:39:36PM -0400, Michael Stone wrote: On Wed, Apr 10, 2002 at 04:45:08PM -0400, Matt Zimmerman wrote: It seems to work for a large number of users, and that is its only job, is it not? Well, no. :) idepci is known to fail for people with scsi and new ide hardware, so it's not really the best choice either. There probably is no best choice, but a system with a menu where you can choose a choice is probably better than any arbitrary default. I have installed many SCSI systems (including the one that I'm using right now) with potato CD #1, which I assume has a similar configuration. I've done the same with several IDE systems, but I'm not in touch with IDE stuff enough to know whether it is considered recent. I absolutely agree that a choice on CD 1 would be superior, especially since it would make it possible to have the choice of a 2.4 kernel while only using one CD. But I think it would be even better to make a high-quality release release according to aj's tentative schedule, rather than a minimally-tested release (possibly much) later. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: frame-buffer on vanilla?
On Sat, Apr 06, 2002 at 02:36:35PM +0200, Eduard Bloch wrote: Anthony Towns wrote on Sat Apr 06, 2002 um 10:14:16PM: Eduard, _no_. What part of that don't you understand? The wtf no without any GOOD reasons part. We want to actually make a release is good enough reason for me. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
sparc tftpboot failure
On an Ultra1, 384MB boot-floppies 3.0.20-2002-03-07: 2b3000 TILO Selecting sun4u kernel... Memory Address not Aligned I also tried potato's tftpboot.img, which does boot, but dbootstrap hangs after downloading rescue.bin. All I can do is kill dbootstrap. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Solved: sparc tftpboot failure
On Wed, Apr 03, 2002 at 12:17:59PM -0500, Matt Zimmerman wrote: On an Ultra1, 384MB boot-floppies 3.0.20-2002-03-07: 2b3000 TILO Selecting sun4u kernel... Memory Address not Aligned I built my own tftpboot.img using boot-floppies 3.0.21 source, and everything works. I do not know whether this was a bug in 3.0.20 that is now fixed, or if something went wrong when the 3.0.20 images were built. Anyway, since there is no official 3.0.21 sparc boot-floppies build, should I do one and upload it? -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Solved: sparc tftpboot failure
On Wed, Apr 03, 2002 at 03:03:29PM -0500, Ben Collins wrote: No, no, no, no and no. There is a serious bug with the 2.4.x kernel images. I have to fix that. The iptables thing that I just saw, or something else? Not to mention that I want some serious testing that not just anyone can perform. What works for you will not work for everyone. Well, then, when you have something ready to test, let me know. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#141102: Release notes should mention rarpd
Package: boot-floppies The release notes (I happened to be looking at SPARC, but I think that section is generic) only seem to mention the kernel RARP server, which is more or less obsolete. They should mention that rarpd is used instead, with newer kernels. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
sparc (sun4u) installation report
On Wed, Apr 03, 2002 at 02:45:30PM -0500, Matt Zimmerman wrote: I built my own tftpboot.img using boot-floppies 3.0.21 source, and everything works. I do not know whether this was a bug in 3.0.20 that is now fixed, or if something went wrong when the 3.0.20 images were built. The only other mishap during installation was that the 'Configure device drivers' step did not work, presumably because I was using a different kernel for the installation. This was non-critical for my installation, since I didn't need any additional drivers, and I just skipped that step. Everything else went according to plan, and the system was up and running woody without a problem once I was able to boot the installer. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Bug#139770: Problem with Installing Debian on a Powerbook 190cs
On Sun, Mar 24, 2002 at 09:31:37PM -0600, Christian T. Steigies wrote: you mean hda=noprobe ? Is that documented somewhere, I found nothing after a quick search. Documentation/ide.txt /usr/share/doc/kernel-doc-x.y.z/Documentation/ide.txt.gz -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: upgrading makedev
On Tue, Mar 12, 2002 at 08:03:30AM -0800, Matt Kraai wrote: How can I make this work, without upgrading my entire system to Sid? Do I need to add something to my sources.list? You need to add unstable to sources.list, otherwise how is apt going to find the proper packages from unstable? You'll also want to add something like: APT::Default-Release testing; to apt.conf to avoid upgrading everything to unstable. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Missing filesystems for 2.4 kernel on testing cds
On Sun, Mar 10, 2002 at 03:14:49AM -0500, Colin Walters wrote: I've downloaded and tried the unofficial woody cd (just disc 1) from powerpc.trasno.net (downloaded on March 1). Well, you are the first person I've heard from who has actually tested installing woody on a PowerPC machine from CD. I have done this, using a previous version of the netinst CDs, and it worked fine. I didn't use NFS, but I was able to install the kernel and such from the CD. I'm pretty sure that I even posted a report to debian-boot at the time. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: minimal cd firewall
On Thu, Feb 21, 2002 at 10:17:54AM -0500, Blanco Alejandro-EAB005 wrote: i'm behind a firewall (http/ftp/socks proxies with authentication). can i use the minimal cd to install debian? When you say minimal cd, do you mean the unofficial netinst images? You can install through an authenticated HTTP proxy with the official CD and floppies, and probably with the netinst images as well. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: 3.0.19-bf2.4 install report
On Sun, Feb 17, 2002 at 11:21:54PM +1100, Glenn McGrath wrote: On Sun, 17 Feb 2002 23:53:44 +1300 (NZDT) Philip Charles [EMAIL PROTECTED] wrote: Why not use fping to check the reply speed of all the mirrors, and choose the fastest as the suggested mirror? In my case the closest mirror is probably San Francisco, a mere 10,000 km (6,000 miles) away. My isp's satellite uplink is there. In other words, let people choose their mirrors. Have you tried netselect-apt, i would be interested to know if it doesnt find the best mirror. I have used it several times, and often it does not find the best mirror. Bear in mind that the mirror with the lowest round-trip latency, topological distance, or geographical distance is not always (or even usually) the one with the greatest end-to-end bandwidth available, and hence not always (or even usually) the one which will provide the fastest package downloads. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Testing boot-floppies on iMac
Yesterday, I got an iMac (NewWorld) to play with at work, and decided to install woody on it. It has no floppy drive, so the only bootable woody installation media that I could find were ISO images at: http://people.debian.org/~ieure/netinst/ which I burned onto CD-R. Unfortunately, at the time there were only images from last December[0] labeled 3.0pre, which contained at least one major known bug (the bad mac-fdisk) which stopped my installation in its tracks). I was able to continue the installation by bringing up the network (which worked flawlessly via DHCP) and wget'ing a fixed mac-fdisk from the 3.0.19 floppy images. I couldn't figure out how to tag a partition as swap, this being my first PowerPC installation. the mac-fdisk help was not helpful). I continued without one, and have since learned (from http://penguinppc.org/projects/yaboot/doc/mac-fdisk-basics.shtml linked from the release notes) that it should be named 'swap'. I assume this would have worked; if so, it should be mentioned directly in the release notes. In tasksel, I selected X window system, desktop environment, C and C++ and Python. During package configuration, I was bitten by pydb #133520, which is fixed and should get into woody eventually. If it doesn't, tasksel #129520 removes pydb from the task anyway. After all packages were installed, I set about trying to get X working, and (after using the debconf prompts to get an initial configuration) co-workers and I fiddled with it for some time before discovering that we had to specify the PCI BusID in XF86Config-4, and google for the correct monitor sync information. I don't know why I had to specify the BusID, as it detected the video card (an ATI) at the correct PCI bus/slot/func during initialization. I had a Device section specifying Driver ati, but until I added the BusID parameter, it said that it couldn't find a matching device section for the card. Once the X server was working, I was very confused to find that no window manager had been installed, nor had xterm. Fortunately, gnome-terminal was there from the desktop environment task. I assume this was tasksel #129217. To summarize: - boot-floppies 3.0.19 has a fix for the mac-fdisk problem - pydb 1.01-5 fixes #133520 (urgency=low) - tasksel 1.15 fixes #129217 and #129520 (urgency=low) The new tasksel will make it into woody in 2 days according to update_excuses.html, and boot-floppies 3.0.19 is already current, so my problems seem to be already fixed and propagating. The situation with tasksel is confusing, though; how does the delay help, when tasksel will only really be tested during new installations? Also, how can I help to test future PowerPC boot-floppies? Is there some way to network-boot this beast, or do I have to continue to use CD-Rs? [0] There seem to be 3.0.19 ISOs at that URL as of today. Thanks to Ian Eure for providing these so that I could get up and running with woody -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Testing boot-floppies on iMac
On Sun, Feb 17, 2002 at 01:12:26AM +0100, Michel D?nzer wrote: On Sam, 2002-02-16 at 21:27, Matt Zimmerman wrote: initialization. I had a Device section specifying Driver ati, but until I added the BusID parameter, it said that it couldn't find a matching device section for the card. A device section without a bus ID is normally assigend to the primary adapter; unfortunately, the primary adapter isn't recognized reliably yet on Macs. Providing the bus ID is the safe bet, and there's a debconf question for it now. Wonderful, so that's fixed as well. Also, how can I help to test future PowerPC boot-floppies? Is there some way to network-boot this beast, or do I have to continue to use CD-Rs? You can boot the installer kernel with the ramdisk from HD with yaboot, or am I missing something? Yes, I suppose I could do that. Thanks. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Why does dbootstrap not make a core file?
On Thu, Feb 14, 2002 at 09:29:09AM -0700, Mark Zimmerman wrote: connection and plenty of mounted partitions. I tranferred an ftp client and the curses and readline libraries to the system via floppy and tried to run ftp. Unfortunately, it said: ftp: ftp/tcp: unknown service Copying in /etc/services should fix that. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Things we need from sid
On Thu, Jan 24, 2002 at 08:42:29PM +0100, Eduard Bloch wrote: Not providing an Ext3 option as-default would damage Debian's reputation. If we are worried about reputations, the best thing to do is to get a release out the door with updated versions of common user packages. Users who want a release that they can install and use are much more numerous than users who cannot convert to ext3 post-install. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Cannot mount previously-initialized ext3 filesystem
I'm not bothering to file a real bug about this unless someone tells me otherwise, because it's an odd case and we all want to get a stable release out the door. But I'm sending a message here so that if it happens to be trivial to fix, someone will do it. I was testing s390 installation under Hercules. Because the installer does not seem to currently support FBA disks, and that is what I must use due to a bug with ECKD disks, I had to initialize and mount my root filesystem manually. This would not be a problem, except that dbootstrap doesn't recognize that I have done so. If I: 1. mke2fs device 2. mount on /target 3. Mount a previously-initialized partition dbootstrap (as expected) tells me that it can't find any suitable partitions, but then after continue, notices that I have mounted it and moves on. If instead, I: 1. mke2fs -j device 2. mount on /target 3. Mount a previously-initalized partition dbootstrap leaves me at the partition a hard disk step with no way to continue. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Cannot mount previously-initialized ext3 filesystem
On Mon, Jan 21, 2002 at 09:04:13PM -0500, Matt Zimmerman wrote: I'm not bothering to file a real bug about this unless someone tells me otherwise, because it's an odd case and we all want to get a stable release out the door. But I'm sending a message here so that if it happens to be trivial to fix, someone will do it. I was testing s390 installation under Hercules. Because the installer does not seem to currently support FBA disks, and that is what I must use due to a bug with ECKD disks, I had to initialize and mount my root filesystem manually. This would not be a problem, except that dbootstrap doesn't recognize that I have done so. Err, never mind. After rebooting and trying again, I have a /dev/dasd/0120/part1 which is now recognized as a useable partition on the FBA device. I don't know why it didn't work before, honestly. So now I don't have to mess with mounting it by hand, and this doesn't matter. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Debian Linus and Macs
On Sun, Jan 20, 2002 at 06:29:33PM -0500, Harry Moses wrote: In the Debian home pages you mention several architectures, none of which seem to indicate that a download is included for the Mac. Perhaps the Power PC includes the Mac. I have an IMac system 9.2. Can the Debian download work on my computer? Yes, the iMac uses a PowerPC processor. See: http://penguinppc.org/projects/hw/ For information about Linux/PowerPC supported hardware. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Flags corruption
On Sat, Dec 29, 2001 at 09:27:53PM +0100, Antoni Bella wrote: On Saturday 29 December 2001 20:52, Junichi Uekawa wrote: In Sat, 29 Dec 2001 20:47:03 +0100 Antoni cum veritate scripsit : I,m catalan traductor and found malformated flags in inst-methods.sgml; prgn/Make Debian Floppy/, -- prgn/Make Debian Floppy/prgn, And your point is? I thought it was a valid sgml. I do not know format SGML. I repeat in the shape of question: it is this correctly? Yes, the SGML syntax above is correct. It is an abbreviated form. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: 3.0.18 testing needed
On Sat, Dec 15, 2001 at 01:32:44PM -0500, Adam Di Carlo wrote: The primary things to test for: - no more flashing screen or nasty segfaults when dbootstrap starts I get the same flashing screen (segfaulting dbootstrap), with the same segfault as before: (gdb) bt #0 0x4007915f in malloc () from /lib/libc.so.6 #1 0x400787e4 in malloc () from /lib/libc.so.6 #2 0x4009c3c7 in ttyname () from /lib/libc.so.6 #3 0x4001c4d9 in newtButton () from /lib/libnewt.so.0.50 #4 0x4001d75b in newtFormRun () from /lib/libnewt.so.0.50 #5 0x4001d511 in newtRunForm () from /lib/libnewt.so.0.50 #6 0x0804d4c2 in strcpy () Cannot access memory at address 0x16 Running on i386 under UML. (If you do get the flashing screen on i386, please fall back to trying another version, URL:http://people.debian.org/~dwhedon/dists/woody/main/disks-i386/current/) I get exactly the same behavior with this one. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: 3.0.18 testing needed
On Sat, Dec 15, 2001 at 10:52:12AM -0800, David Kimdon wrote: trying another version, URL:http://people.debian.org/~dwhedon/dists/woody/main/disks-i386/current/) I get exactly the same behavior with this one. FWIW I just installed two systems using the idepci/reiserfs flavors built out of current cvs, perhaps UML is causing the bad behavior? The only bad behavior I observed is documented in #124117 This is a different issue from #124117; it is the problem where dbootstrap segfaults immediately at startup. The 3.0.17-2001-11-18 root.bin works fine under UML. This problem appeared in later builds, and seemed to affect others as well. If I replace libslang and libm on root.bin with the ones from my unstable system, everything works. It certainly doesn't seem like a UML problem. Perhaps we aren't using the same disks? d41d8cd98f00b204e9800998ecf8427e 3.0.17-2001-11-18/root.bin 94d3f2e5675d641ebec98ba934f755fb 3.0.18-2001-12-14/root.bin 2fd4d93d5134c1a4afc76cf7cefa62e8 dwhedon/root.bin -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Debian Won't Install! (System Locks Up)
On Wed, Dec 12, 2001 at 05:53:41PM -0800, Mark Seven Smith wrote: I can get through the installation process, by formatting my hard drive ahead of time, and then just hitting ENTER at each prompt. BUT THE REAL PROBLEM COMES when I am supposed to choose modules. I do not understand this part, but in the book, Learning Debian GNU/Linux, at this section, they say that I should choose THESE MODULES to install: If this is really the part which causes the real problem, just skip it. This is something that can be done after installation, if you need it at all. Just accept the default modules. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
b-f testing and development with user-mode linux
I just recently packaged User-mode Linux for Debian (http://user-mode-linux.sourceforge.net/). During that time, it struck me as being incredibly useful for several Debian-related tasks. One of these was development and testing of boot-floppies/debian-installer. For those unfamiliar with the project, User-mode Linux is a port of the Linux kernel to Linux userland. It allows you to run a Linux kernel as a user process under a normal Linux kernel. I have done a complete installation under UML using root.bin from boot-floppies 3.0.17-2001-11-18 and a UML kernel[0]. I think that this could prove useful to many of you in your debugging efforts, as it makes it possible to, e.g., run gdb, easily copy files back and forth to the host system using hostfs, and have access to tools like strace and gdb without installing them on a separate partition or cramming them onto installation media. You can do all of this with only one system, and without requiring root access. Here's a screenshot to demonstrate: http://people.debian.org/~mdz/uml-bootfloppies.jpeg The package is called user-mode-linux, and it's now in unstable. It currently only works on i386, but it is theoretically possible to port it to certain other architectures, and upstream is very interested in finding people who have the necessary arch-specific knowledge and can do the work. Some work has already been done on PowerPC, but it is currently stalled. Naturally, it can't test all aspects of the installation process, but I imagine that the more generic bits can be tested more easily this way. Let me know what you think. [0] Well, it almost completely works. There are a couple of caveats: 1. The HD autodetection doesn't find the virtual disks, which use the ubd device. This is easy to work around by mounting the device by hand and running MAKEDEV ubd in /target/dev before rebooting. UML has a mechanism whereby the ubd devices can be accessed via major device 3 to look like IDE disks, but processes which try to access these devices hang indefinitely. This is probably a UML bug, and it would be great to find it and fix it. 2. /sbin/probe from pcmcia-cs seems to hang the system. This can be worked around by disabling the PCMCIA controller in dbootstrap. This is also probably a UML bug, and it would also be great to find it and fix it. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: forget LANG_CHOOSER for 3.0.18 (was Re: 3.0.18 testing)
On Mon, Dec 10, 2001 at 11:07:01AM +, [EMAIL PROTECTED] wrote: In message [EMAIL PROTECTED], Adam Di Carlo writes: This seems like a reasonable set of stuff for i18n builds, but it causes the segmentation faults. I got a static strace on the rootdisks and this is the strace ending: Yes, something is _still_ going wrong with libnewt. I don't understand what, yet, but I'm looking at it. I'd expect that libnewt.so on the root disk should be exactly the same as /usr/lib/libnewt.so on my build system (in absence of library reduction) yet this is not the case. And if I copy the one from the host into the root image, the segfault goes away. Did you check whether they are actually different? My system's libnewt.so has exactly the same md5sum as the one on root.bin. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: b-f testing and development with user-mode linux
On Mon, Dec 10, 2001 at 01:03:44PM +0100, Petter Reinholdtsen wrote: [Matt Zimmerman] Here's a screenshot to demonstrate: http://people.debian.org/~mdz/uml-bootfloppies.jpeg How to you start uml to get this working? Can you make it boot from the install CD? Well, you have to use a patched kernel, so you can't use the exact kernel from the rescue disk or the CD. You can use the same root filesystem, though. What I do is use the standard 2.4.16-2um UML kernel from the user-mode-linux package and give it root.bin as its root filesystem. I add on a filesystem to install on, and one containing some auxiliary utilities (like the modules for the UML kernel), and I start it like this: linux con=xterm eth0=tuntap,,,aaa.bbb.ccc.ddd \ ubd0=install.ext2 ubd1=root.ext2 ubd2=utils.ext2 \ root=/dev/ubd/1 devfs=nomount The eth0 bit is for TUN/TAP virtual networking, which is optional if you supply the .debs and such from some other source. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Problems with network install of rescue.bin
Using boot-floppies 3.0.17-2001-11-18 on i386 to install testing. When trying to download rescue.bin over the network, dbootstrap used the path: http://my.mirror/debian/images-1.44/rescue.bin That is, just appending images-1.44/rescue.bin to the archive URL. Earlier in the process, of course, I had selected 'testing', and debootstrap successfully downloaded and installed the base .debs from the correct location. To work around the problem, in the URL dialog, I replaced: http://my.mirror/debian (the default value) with http://my.mirror/debian/dists/woody/main/disks-i386/current which allowed rescue.bin and drivers.tgz to be downloaded correctly. After it was downloaded, I got an error: Mount Failed Unable to mount the Rescue Floppy. You may have inserted the wrong floppy. Please try again. Continue and then: Select Floppy Disk Drive Please select the floppy drive you will use to read the Rescue Floppy. /dev/fd0 : first floppy drive /dev/fd1 : second floppy drive Cancel Of course, I don't want either of these; I want to use the image I just downloaded. Are these known problems? -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Problems with network install of rescue.bin
On Sun, Dec 09, 2001 at 04:46:43AM -0500, Matt Zimmerman wrote: Using boot-floppies 3.0.17-2001-11-18 on i386 to install testing. When trying to download rescue.bin over the network, dbootstrap used the path: http://my.mirror/debian/images-1.44/rescue.bin That is, just appending images-1.44/rescue.bin to the archive URL. Earlier in the process, of course, I had selected 'testing', and debootstrap successfully downloaded and installed the base .debs from the correct location. In case it wasn't apparent, I happened to install the base system BEFORE installing the kernel and modules. So it seems that the download URL was set to point to the top of the archive tree, and then that (incorrect) path was used when I went back to installing the kernel. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Problems with network install of rescue.bin
On Sun, Dec 09, 2001 at 01:30:49PM +, Philip Blundell wrote: In message [EMAIL PROTECTED], Matt Zimmerman writes: In case it wasn't apparent, I happened to install the base system BEFORE installing the kernel and modules. So it seems that the download URL was set to point to the top of the archive tree, and then that (incorrect) path was used when I went back to installing the kernel. Incidentally, this is basically the same issue as #122633 and #69155. Hmm, I didn't realize this was a requested feature and not a bug. ;-) The bug title is actually misleading; I don't care whether it remembers my mirror selection. The problem is that, if you do things in this order, the default is wrong and doesn't work. Any idea about the problem with extracting them once they've been downloaded? -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: forget LANG_CHOOSER for 3.0.18 (was Re: 3.0.18 testing)
On Sun, Dec 09, 2001 at 09:54:14PM -0800, David Kimdon wrote: ifeq ($(LC),true) -CFLAGS += -DUSE_LANGUAGE_CHOOSER +CFLAGS += -DUSE_LANGUAGE_CHOOSE -DUTF8 ^ Is this intentional? -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Please commit to help the translators (was: What is IPL ?)
On Wed, Dec 05, 2001 at 10:27:09AM +0100, Martin Quinson wrote: === #. TRANS: IPL (Initial Program Load) means boot in mainframe context and #. DASD can be viewed as a hard disk. Please don't translate these terms #: bootconfig.c:1557 [...] This is the regular way for the programmer to say something to the translators without direct interaction. (for the curious, it works because in the makefile, xgettext is called with the argument --add-comments=TRANS) I'm not very sure of what DASD means, so I let a S390 correct my explanation and commit that... The acronym stands for Direct-Access Storage Device. It refers to a class of secondary storage devices typically used on mainframe systems. For translation purposes, it is interchangeable with disk or storage array. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: What is IPL ?
On Tue, Dec 04, 2001 at 08:34:17AM +0100, Martin Quinson wrote: I'm the french translator of dbootstrap, and I can't translate what I don't understand, so I have to ask it here : What does the following messages mean ? #: bootconfig.c:1555 msgid Running ZIPL to make the kernel able to IPL from the DASD... #: bootconfig.c:2582 msgid ZIPL could not be installed. You are not able to IPL this Linux system right now. Please see the log file and manually repair the problem. Don't you think it would be an improvement to redo these sentences in a more understandable way ? ;) IPL stands for Initial Program Load, and is another name for what we microcomputer types call booting. The acronym IPL is very common in the mainframe industry, culture and documentation, and should be preserved verbatim. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Translation statistics
On Sun, Nov 18, 2001 at 09:15:18PM +0100, Claus Hindsgaul wrote: Code date translated up-to-date strings (out of 769) ca 2001-10-15 774 How can there be 774 up-to-date strings out of 769? Are strings which are no longer used considered up-to-date? -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: apt behaviour with satellite
On Sat, Nov 10, 2001 at 07:32:16PM -0500, John Davidson wrote: Apt now behaves like wget when installing woody using the rieserfs 3.0.15 bootfloppies. This means that it no longer works with my one-way satellite connection. Prior to the releease of 3.0.14 bootfloppies this was not a problem and I would regularly get 40KB speeds. Now I get less than 2KB??! Things improve to about 3.5KB when I use my modem without satellite. This is the exact performance I get with wget. Has apt been changed since mid-Aug?? If so I will report a critical bug. If you do, I'll be sure to downgrade it if nobody else gets to it first. critical makes unrelated software on the system (or the whole system) break, or causes serious data loss, or introduces a security hole on systems where you install the package. This bug obviously does not break the whole system, or even unrelated software. If you could provide some information about what is different, at the source code or protocol level, to cause this change in behavior, someone could help you track down the problem (which is probably not a bug in APT or wget). -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Installation under Hercules?
On Wed, Sep 12, 2001 at 02:24:47PM +0200, Stefan Gybas wrote: On Tue, Sep 11, 2001 at 09:32:15PM -0400, Matt Zimmerman wrote: illegal operation: 0001 Very strange, but this looks like a bug in Hercules to me. I agree that the bug is probably in Hercules. I had hoped that someone else here might have investigated the problem further. I can no longer reproduce the problem reliably; sometimes, the format will complete successfully for a 600-cylinder volume, sometimes not. I am also continuing to track down the update-alternatives problem that I and others have experienced under Hercules. -- - mdz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]