Bug#682211: wheezy installation report: some rough edges
Subject: installation-reports: d630 Package: installation-reports Severity: normal (sorry for lack of logs, sending from different machine) -- Package-specific info: Boot method: Image version: http://cdimage.debian.org/cdimage/wheezy_di_alpha1/i386/iso-cd/debian-wheezy-DI-a1-i386-netinst.iso 12-May-2012 02:45 Date: Machine: Dell Latitude D630 Partitions: Base System Installation Checklist: [O] = OK, [E] = Error (please elaborate below), [ ] = didn't try it Initial boot: [O] Detect network card:[O] Configure network: [E] Detect CD: [O] Load installer modules: [O] Clock/timezone setup: [ ] User/password setup:[O] Detect hard drives: [O] Partition hard drives: [E] Install base system:[O] Install tasks: [E] Install boot loader:[O] Overall install:[O] Comments/Problems: * If I leave the "domain" field empty in the hostname selection screen, resolv.conf expectedly only gets a "nameserver" line, and the resolver does not resolve my (non-fully-qualified) proxy name. Although it is normal, I have to switch to tty4, where the log just tell that the mirror does not carry wheezy (sic). * I do not remember any prompt telling me to check time, but TZ is correctly set * when partitionning (assisted lvm), double clicking on a partition to see details, then back to main screen, the partition is still selected and clicking on "continue" unexpectedly just enters the same partition-editing dialogs again * only a single task to be selected in the installer (standard system utilities), probably why no desktop env has been installed (selected XFCE) * selecting task-french, task-french-desktop, task-xfce triggers conflict between - myspell-fr and myspell-fr-gut - hunspell-en-us and myspell-en-us * I would have expected http_proxy to be propagated to /etc/environment, not just to apt.conf (although for a laptop it is probably better not to do that) Those seem to be problems of the package versions on install CD, but fixed in current wheezy: * no extended package descriptions are downloaded by apt on first update after install, only on second update * aptitude with french l10n does not translate "yes/no" to "oui/non" in confirmation dialogs, but does it only accepts "o" instead of "y" (not any more after upgrade ?) -- To UNSUBSCRIBE, email to debian-boot-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20120720120159.6c7f6...@chalon.bertin.fr
Bug#588773: debootstrap: cannot use fakechroot variant
Package: debootstrap Version: 1.0.23, 1.0.12 * --variant=fakechroot is mentionned in the manpage, but no example command-line is provided, so the user is left with trial-and-error (which is annoyingly long when using a remote mirror) * the only way in which I could launch it is: $ fakeroot fakechroot debootstrap --variant=fakechroot lenny /tmp/lenny-test But then it fails with: W: Failure trying to run: chroot /tmp/lenny-test dpkg --force-depends --install /var/cache/apt/archives/base-files_5lenny6_i386.deb /var/cache/apt/archives/base-passwd_3.5.20_i386.deb * re-running the same command without removing the WIP chroot gives no error, but then fails silently, which is bad since it could leave the user thinking that "this time it worked": I: Validating zlib1g I: Chosen extractor for .deb packages: dpkg-deb I: Extracting libacl1... I: Extracting libattr1... I: Extracting base-files... local...@chalon:~$ echo $? 2 -- Yann Dirson - Bertin Technologies -- To UNSUBSCRIBE, email to debian-boot-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20100712094549.2c8b1...@chalon.bertin.fr
Bug#543149: console-setup: perl errors during installation
> On Monday 31 August 2009, Yann Dirson wrote: >> This seems to be similar to #480533, which was apparently fixed by >> moving more stuff to perl-base. > > I very much doubt it is related as the version of perl that fixed #480533 > is ancient (included in stable). So if that change had fixed it, then I > would never have seen the problem when I did. > > It may be that *a* solution to this issue could be to move yet more stuff > to perl-base, but I doubt that is desirable. > > IMO console-setup should, if at all possible, make sure that it only uses > functions available in current perl-base. Neither extending perl-base, > nor having console-setup depend on perl are desirable solutions. If I understand correctly what happened with #480533, it was a utf8-related module that was already part of perl-base, which suddenly gained dependencies it did not have prior to that. I thought we could be hitting the same problem again. -- To UNSUBSCRIBE, email to debian-boot-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#543149:
This seems to be similar to #480533, which was apparently fixed by moving more stuff to perl-base. -- To UNSUBSCRIBE, email to debian-boot-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#538397: console-setup: complains about unsupported ctrl:nocaps on upgrade
On Sat, Jul 25, 2009 at 09:30:10PM +0300, Anton Zinoviev wrote: > On Sat, Jul 25, 2009 at 03:50:31PM +0200, Yann Dirson wrote: > > > > I use ctrl:nocaps since ages, and only now (upgrading from 1.36) does > > console-setup start to complain about it. > > What do you mean by "complain"? I tried manualy the ckbcomp utility and > ctrl:nocaps seems to work. On upgrade, I got a debconf question saying it was not supported, and asking if I wanted to use it nevertheless. -- To UNSUBSCRIBE, email to debian-boot-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#538397: console-setup: complains about unsupported ctrl:nocaps on upgrade
Package: console-setup Version: 1.44 Severity: normal I use ctrl:nocaps since ages, and only now (upgrading from 1.36) does console-setup start to complain about it. I could not find a reason for this from a quick search in the changelog. I just hope it will continue to work in X despite console-setup not knowing about it :) -- System Information: Debian Release: squeeze/sid APT prefers testing APT policy: (990, 'testing'), (500, 'oldstable'), (500, 'unstable'), (101, 'experimental') Architecture: i386 (i686) Kernel: Linux 2.6.28.10-smp-smp-2-g0fee61b (SMP w/2 CPU cores) Locale: LANG=C, LC_CTYPE=french (charmap=ISO-8859-1) Shell: /bin/sh linked to /bin/bash Versions of packages console-setup depends on: ii console-terminus 4.28-1 Fixed-width fonts for fast reading ii debconf [debconf-2.0] 1.5.27 Debian configuration management sy ii xkb-data 1.6-1 X Keyboard Extension (XKB) configu Versions of packages console-setup recommends: ii console-tools1:0.2.3dbs-65.1 Linux console and font utilities Versions of packages console-setup suggests: ii locales 2.9-12 GNU C Library: National Language ( ii lsb-base 3.2-22 Linux Standard Base 3.2 init scrip -- debconf information: * console-setup/variant: France console-setup/unsupported_options: true console-setup/modelcode: pc105 console-setup/use_system_font: console-setup/fontsize: 16 console-setup/unsupported_layout: true console-setup/layoutcode: fr console-setup/codesetcode: Uni1 console-setup/altgr: No AltGr key * console-setup/ttys: /dev/tty[1-6] * console-setup/codeset: . Combined - Latin; Slavic Cyrillic; Hebrew; basic Arabic console-setup/toggle: No toggling * console-setup/fontface: VGA * console-setup/fontsize-text: 16 console-setup/compose: No compose key debian-installer/console-setup-udeb/title: console-setup/switch: No temporary switch console-setup/unsupported_config_layout: true * console-setup/charmap: UTF-8 console-setup/optionscode: ctrl:nocaps console-setup/unsupported_config_options: true * console-setup/layout: France console-setup/variantcode: * console-setup/model: Generic 105-key (Intl) PC console-setup/fontsize-fb: 16 -- To UNSUBSCRIBE, email to debian-boot-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#230440: FWD: Re: INSTALL REPORT with 2004-01-09 netinst, failed
On Tue, May 25, 2004 at 02:17:50PM +0100, Martin Michlmayr wrote: > > As I attempt to do more not-so-uncommon things (with the 2004-01-03 > > snapshot)... > > This is very old now can you please try a new image from > http://gluck.debian.org/cdimage/testing/sarge_d-i/ and see if that > works better? Yes, I've planned to test the latest beta, I'll let you know. BTW, this @alcove email will expire soon, please use the @d.o one for further contact. > > - I initially have a HD with a single ext3 partition + swap (from a > > previous RH install). If I repartition by deleting the ext3 and > > repartitionning that space in 2, on the partition setup screen I > > notice the 1st partition is already listed as ext3, and the 2nd not. > > Then I realize it comes from the previous (RH) install. I had not > > realized that when just selecting a single partition, when I thought > > "ext3" was just a default. > > The partitioning tool recognizes the existing file system on that > partition and shows it. The 2nd partition didn't have the start of a > new partition... maybe the partitioning tool should overwrite the > beginning of a partion when creating one. Hmm, no, that wouldn't work > because creating partitions is first only done in memory. Then partition objects in memory should probably be initialized as having no fs on them, instead of looking up the disk. Or, maybe better, they could be initialized from disk only if the fs size matches (or maybe is smaller than) the new partition, and maybe offer to resize2fs to the partition size if it was smaller than the new partition ? > > Now if my original single-partition-over-RH had gone throught the > > reboot stage, I may surely have ended up with a strange install. And > > Well, you have to mark the partition as to be formated. I've assimilated this by now ;), but for some reason it did not seem intuitive to me when doing this first install with d-i. I'll look at the latest version, and see if I still feel some needed improvements in this area. > > - Now when I decide to split the existing partition in 2, but do not > > realize I must explicitely select "format as ext3", since my partition > > already appears as having ext3 selected (I thought), and go on, I'm > > brought back to the main menu on the next step, and requesting to > > install the base system brings me back there again almost instantly. > > Oh, hold on. So you used the new tool. So you didn't select "format > as ext3". But what did you select instead? "dont touch" ? Because > if you choose nothing at all, you shouldn't have been able to assign a > mount point at all. I suppose I used the default choice, thinking the partitionner was by default suggesting to put an ext3 fs on the 1st partition I created (which could have been a reasonable suggestion), whereas it was telling me something different - see above. > > - Selecting "French" as the keyboard layout after reboot, I do not > > have time to see the keymap name, and subsequent error message, before > > the next dialog screen. And I end up with a querty-looking layout. > > Can you check if this has been fixed? I think it has, and beta3 was correct here. Regards, -- Yann Dirson <[EMAIL PROTECTED]> http://www.alcove.com/ Senior Free-Software Consultant Consultant senior en Logiciels Libres Debian developer ([EMAIL PROTECTED])Développeur Debian -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#241647: base-config: wrong dialog ordering at user creation
On Fri, Apr 02, 2004 at 06:35:25AM -0800, Matt Kraai wrote: > On Fri, Apr 02, 2004 at 10:00:41AM +0200, Yann Dirson wrote: > > When asking questions to create the non-root user, the question for the > > user's name appears before the explanation saying that a non-root user is to > > be created. Even an experienced user can wonder whether the question is > > about the name of the root user. > > I suspect this is bug 240607, which has since been fixed. Would > you please retest with a later image and, if it is indeed fixed, > close this bug? Hm, only yesterday's build is more recent, could this have been fixed between the 20040331 and 20040401 builds ? -- Yann Dirson<[EMAIL PROTECTED]> |Why make M$-Bill richer & richer ? Debian-related: <[EMAIL PROTECTED]> | Support Debian GNU/Linux: Pro:<[EMAIL PROTECTED]> | Freedom, Power, Stability, Gratuity http://ydirson.free.fr/| Check <http://www.debian.org/> -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#241697: base-config: "standalone" choice in exim4-config returns to main menu
Package: base-config Version: 20040331 d-i snapshot Tags: d-i After an install in "normal" mode, if I select a "standalone server" setup in exim4 config dialog, I'm brought back to the base-config main menu, with "configure MTA" selected. If then I select this (obvious ?) choice, I'm asked questions at medium priority, give default answer to all, and I'm then left back to main menu, with next step selected (so I assume this time it had worked). This may be an exim4 bug instead, you know better :) -- Yann Dirson<[EMAIL PROTECTED]> |Why make M$-Bill richer & richer ? Debian-related: <[EMAIL PROTECTED]> | Support Debian GNU/Linux: Pro:<[EMAIL PROTECTED]> | Freedom, Power, Stability, Gratuity http://ydirson.free.fr/| Check <http://www.debian.org/> -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#241647: base-config: wrong dialog ordering at user creation
Package: base-config Version: 20040331 d-i snapshot Tags: d-i important When asking questions to create the non-root user, the question for the user's name appears before the explanation saying that a non-root user is to be created. Even an experienced user can wonder whether the question is about the name of the root user. -- Yann Dirson <[EMAIL PROTECTED]> http://www.alcove.com/ Senior Free-Software Consultant Consultant senior en Logiciels Libres Debian developer ([EMAIL PROTECTED])Développeur Debian -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#241644: d-i: bad french title for last dialog
Package: debian-installer Version: 20040331 Tags: d-i One of the last boxes says "Terminer l'installation", while it would much probably be "Fin de l'installation" or "Finalisation de l'installation". -- Yann Dirson<[EMAIL PROTECTED]> |Why make M$-Bill richer & richer ? Debian-related: <[EMAIL PROTECTED]> | Support Debian GNU/Linux: Pro:<[EMAIL PROTECTED]> | Freedom, Power, Stability, Gratuity http://ydirson.free.fr/| Check <http://www.debian.org/> -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#241478: partman: fatal error : OK - cancel ?
On Thu, Apr 01, 2004 at 08:53:27PM -0500, Joey Hess wrote: > Yann Dirson wrote: > > Package: debian-installer > > Version: 20040331 > > Tags: d-i > > Since you know the package is partman, please file the bug report on > partman. Hm, I should have checked further really - I had just noticed that there were partman items currently assigned to d-i... -- Yann Dirson<[EMAIL PROTECTED]> |Why make M$-Bill richer & richer ? Debian-related: <[EMAIL PROTECTED]> | Support Debian GNU/Linux: Pro:<[EMAIL PROTECTED]> | Freedom, Power, Stability, Gratuity http://ydirson.free.fr/| Check <http://www.debian.org/> -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#241434: d-i: multiple paths for keyboard selector, "back" broken
On Fri, Apr 02, 2004 at 07:08:23AM +0200, Christian Perrier wrote: > Quoting Joey Hess ([EMAIL PROTECTED]): > > > Having back go to a menu from which the user can pick any prior step, or > > get more control, is a reasonable compromise. There are probably better > > ones, like renaming the back button to "cancel" and having it always go to > > the menu, or something. > > Having it called "Main menu" and always bring users to the main menu > would be indeed an interesting compromise, yes. Sure, that would be less confusing than the current use of "back". But if we can still go back inside a menu item, it may be worth to keep the 2 notions, and only provide a "back" button when we know we can honor it ? > Also, if we find a way to do this, we should try to avoid lowering the > debconf priority to mediumso this is indeed tricky as trying to go > back to main-menu with priority=high will never display the menu..:-) What about giving it a high priority, but ensuring it is not in the direct path for normal install ? -- Yann Dirson<[EMAIL PROTECTED]> |Why make M$-Bill richer & richer ? Debian-related: <[EMAIL PROTECTED]> | Support Debian GNU/Linux: Pro:<[EMAIL PROTECTED]> | Freedom, Power, Stability, Gratuity http://ydirson.free.fr/| Check <http://www.debian.org/> -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#241478: partman: fatal error : OK - cancel ?
On Thu, Apr 01, 2004 at 05:49:43PM +0200, Christian Perrier wrote: > No. Auto-partitioning is only possible for bigger disks. I don't > remember which size is the limit. Well, then the choice should not be proposed to the user. And why would a 200MB disk be too small for autopartitionning ? IIRC the base install only takes around 150MB, isn't it ? It could simply be formatted as a single-partition one without swap - what I didn indeed, 128MB RAM should be enough I suppose (could not check yet, hit a qemu segfault, under investigation). > > And below that I get 2 buttons : "back" and "continue". In what are those > > supposed to be different ? Reminds me of the old Windows dialog proposing > > "OK" and "Cancel" when a fatal error occured - no offense meant :) > > Uh. That would mean the errors dialogs get the two buttons which is > definitely not what expected... > > However, I'm not really sure which package you *really* > used...because, you seem to have been hit by some bad handling of the > "Back" buttons. After the segfault I restarted installation, being careful not to hit any "back" button, and when I asked for autopartitionning, I don't remember seing a different dialog than the 1st time. But it's easily reproduceable: - install qemu - dd if=/dev/zero of=hda count=409600 - make sure you have TUN/TAP available to your host kernel, and /dev/net/tun created (see other bugreport) - qemu -cdrom sarge-netinst.img -hda hda -boot d -m 128 -- Yann Dirson<[EMAIL PROTECTED]> |Why make M$-Bill richer & richer ? Debian-related: <[EMAIL PROTECTED]> | Support Debian GNU/Linux: Pro:<[EMAIL PROTECTED]> | Freedom, Power, Stability, Gratuity http://ydirson.free.fr/| Check <http://www.debian.org/> -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#241494: partman: cannot reuse existing partition
Package: debian-installer Version: 20040331 Tags: d-i During manual partionning, when trying to reuse the single-partition (but reformatting it) on my HD image from qemu, I can select the partition, but when I select the "file system type" item I'm brought back to the manual-patitionning menu. If then I select the line describing the disk, I'm brought back to main d-i menu instead. I expected some menu allowing me to destroy all partitions. Note: even with an initial patition table on the disk, I could not make the auto-partitionning work. Workaround: If I first delete the partition, things proceed as expected. -- Yann Dirson<[EMAIL PROTECTED]> |Why make M$-Bill richer & richer ? Debian-related: <[EMAIL PROTECTED]> | Support Debian GNU/Linux: Pro:<[EMAIL PROTECTED]> | Freedom, Power, Stability, Gratuity http://ydirson.free.fr/| Check <http://www.debian.org/> -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#241491: d-i: cannot install without a network card
Package: debian-installer Version: 20040331 Tags: d-i If there is no network card in the machine, using the netinst image (which should be able to install the base system without a network access, as I could check), refuses to get past the network driver selection dialog, at least in qemu. I had to start qemu in such a way that it has a network interface, even if it's not connected to a usable network, to go further. This will be annoying to install on some embedded systems such as PDAs. -- Yann Dirson<[EMAIL PROTECTED]> |Why make M$-Bill richer & richer ? Debian-related: <[EMAIL PROTECTED]> | Support Debian GNU/Linux: Pro:<[EMAIL PROTECTED]> | Freedom, Power, Stability, Gratuity http://ydirson.free.fr/| Check <http://www.debian.org/> -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#241490: d-i: manual IP conf may suggest self IP as gateway
Package: debian-installer Version: 20040331 Severity: minor Tags: d-i If I enter an IP ending with .1, the suggested IPs for gw and ns end up being the same as the machine's IP. Surely not accurate to the gateway. -- Yann Dirson<[EMAIL PROTECTED]> |Why make M$-Bill richer & richer ? Debian-related: <[EMAIL PROTECTED]> | Support Debian GNU/Linux: Pro:<[EMAIL PROTECTED]> | Freedom, Power, Stability, Gratuity http://ydirson.free.fr/| Check <http://www.debian.org/> -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#241478: partman: fatal error : OK - cancel ?
Package: debian-installer Version: 20040331 Tags: d-i Requesting autopartitionning of my blank 200Mb HD (qemu disk image created with dd if=/dev/zero), I get a red screen (in english, whereas the install is otherwise in french), telling that automatic partitionning failed, probably because there is not enough space. This in itself look suspect. And below that I get 2 buttons : "back" and "continue". In what are those supposed to be different ? Reminds me of the old Windows dialog proposing "OK" and "Cancel" when a fatal error occured - no offense meant :) -- Yann Dirson<[EMAIL PROTECTED]> |Why make M$-Bill richer & richer ? Debian-related: <[EMAIL PROTECTED]> | Support Debian GNU/Linux: Pro:<[EMAIL PROTECTED]> | Freedom, Power, Stability, Gratuity http://ydirson.free.fr/| Check <http://www.debian.org/> -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#241476: partman: automatic/manual switch not consistent with d-i
Package: debian-installer Version: 20030331 Tags: d-i When I get the 1st partman choice-list ("entire drive" or "manual"), then select "manual", realize as Joe A. User that was not a good choice, I expect "back" to bring me to these simple alternatives I saw on previous screen. Unfortunately, I'm back to main menu (even more complicated !), and when I select "partition the HD", I'm thrown back into the "manual" stuff. There I have to carefully read all items, and infer that "automatically partition the HD" is what I want. IMHO it would be much more consistent and intuitive to get rid of this "automatically partition the HD" item, and rely on "back" for this feature. -- Yann Dirson<[EMAIL PROTECTED]> |Why make M$-Bill richer & richer ? Debian-related: <[EMAIL PROTECTED]> | Support Debian GNU/Linux: Pro:<[EMAIL PROTECTED]> | Freedom, Power, Stability, Gratuity http://ydirson.free.fr/| Check <http://www.debian.org/> -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#241467: d-i: too many things happen when wrong network driver is selected
Package: debian-installer Version: 20040331 i386 Context: installing in qemu, and using a trail-and-error approach to configure the network card (emulated non-pci ne2000). Whenever I select a wrong driver (or wrong io/irq parameters), the driver can't init as shown in the logs. But then the installer goes on autodetecting hardware, including attempts to load vesafb and friends. Whenever I try another _network_ driver I have to wait for the installer to redo the same unrelated probing again. -- Yann Dirson<[EMAIL PROTECTED]> |Why make M$-Bill richer & richer ? Debian-related: <[EMAIL PROTECTED]> | Support Debian GNU/Linux: Pro:<[EMAIL PROTECTED]> | Freedom, Power, Stability, Gratuity http://ydirson.free.fr/| Check <http://www.debian.org/> -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#241434: d-i: multiple paths for keyboard selector, "back" broken
On Thu, Apr 01, 2004 at 02:32:00PM +0200, Christian Perrier wrote: > This is because priority has been lowered to low. The main-menu > only shows up at medium priority. Same for keyboard type choice. This sounds not too bad. > Hitting "Back" at the keyboard choice screen automatically lowered the > debconf priority. This is mostly a consequence of the general design > of the installer, IIRC. But this sounds quite strange: 1. just lowering the priority (as opposed to "setting to medium" may have the consequence you describe below 2. even then, why should "back" bring me to a menu I had not seen before ? AFAICT it should bring me to the previous choice-list, period. A novice user may have doubts on what he previously selected, may realize that the previous choice was not the one he wanted, etc. Dropping him into the menu is definitely an error IMHO. Especially, if what you describe is what hapenned, I first switched to "medium" when hitting "back" in the layout list, and only to "high" after seeing the "AT vs. USB" one. Then the latter must be priority "medium", and I should probably have seen it when hitting "back" (which would have been highly counter-intuitive as well). > Thus, you were on main menu at medium priority. > > Choosing back lowered the priority again.and you end up with > installing at "low" priority and be prompted on about everything s/low/high/ I suppose ? Puzzled, -- Yann Dirson<[EMAIL PROTECTED]> |Why make M$-Bill richer & richer ? Debian-related: <[EMAIL PROTECTED]> | Support Debian GNU/Linux: Pro:<[EMAIL PROTECTED]> | Freedom, Power, Stability, Gratuity http://ydirson.free.fr/| Check <http://www.debian.org/> -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#241434: d-i: multiple paths for keyboard selector, "back" broken
Well, there is more, although I'm not sure it's a different problem: After this, I'm sent back to the main menu instead of proceeding to the next step. And shortly after I'm presented with the list of modules to attempt to load. I'm quite sure that should not be the experience of the average user who just hit enter at the boot prompt... -- Yann Dirson<[EMAIL PROTECTED]> |Why make M$-Bill richer & richer ? Debian-related: <[EMAIL PROTECTED]> | Support Debian GNU/Linux: Pro:<[EMAIL PROTECTED]> | Freedom, Power, Stability, Gratuity http://ydirson.free.fr/| Check <http://www.debian.org/> -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#241434: d-i: multiple paths for keyboard selector, "back" broken
Package: debian-installer Version: beta3 (Mar 15 22:04 sarge-i386-netinst.iso) Tags: d-i When I select fr_FR as language, and then select "Go back" (which brings me to main menu instead of going back), then select the item to choose my keyboard layout, I'm presented a new dialog about keyboard type (AT vs. USB). When I select AT I now get the list I was expecting. If now I hit "back" again I (presumably correctly) get the AT vs. USB list. Those 2 issues seem closely related. -- Yann Dirson<[EMAIL PROTECTED]> |Why make M$-Bill richer & richer ? Debian-related: <[EMAIL PROTECTED]> | Support Debian GNU/Linux: Pro:<[EMAIL PROTECTED]> | Freedom, Power, Stability, Gratuity http://ydirson.free.fr/| Check <http://www.debian.org/> -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#221908: initrd image name
The generated lilo.conf also points to /initrd.img. However, at least after installing a new custom-built kernel-image, the links are named /initrd and /initrd.old, so lilo still complains on that. After having complained about the devfs issues, that is. I did not see that mentionned here, but since devfs is not mounted by default, lilo refuses to run. Not counting that the device names in lilo.conf are non-devfs ones... "Obviously", mounting devfs and (temporarily) editing lilo.conf, just the time to reboot and edit lilo.conf back to normal (non-devfs), and things are OK again... Regards, -- Yann Dirson <[EMAIL PROTECTED]> http://www.alcove.com/ Technical support managerResponsable de l'assistance technique Senior Free-Software Consultant Consultant senior en Logiciels Libres Debian developer ([EMAIL PROTECTED])Développeur Debian -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Missing packages on initial install
After completing my install using the Jan 3rd snapshot, I found the following crucial packages missing, which I suppose may be caused by their "optional" priority: - hotplug This one is the prefered way to handle USB stuff, and without it my machine with a USB mouse, well... did not have a usable mouse. Given that USB devices are more and more ubiquitous, and that it's non-trivial to make them work without hotplug (you know, hid, mousedev, etc), and that when installing hotplug afterward you have {un,re}plug your USB devices, or to reboot, to have them seen, it would seem reasonable to have it there by default, and installed early. Not having it available when it is referenced in /proc/sys/kernel/hotplug by default would be like not having modutils installed by default... - gpm Maybe less crucial (esp. as the mouse was not configured ;), but quite standard in fact... Maybe it's not included because it has not been debconfized ? - tree Even less crucial, but I believe it is quite standard as well now. What's to be done about that, assuming most people agree with this ? Requesting priorities to be raised in the override file ? Or is there some d-i-specific action to be done ? Regards, -- Yann Dirson <[EMAIL PROTECTED]> http://www.alcove.com/ Technical support managerResponsable de l'assistance technique Senior Free-Software Consultant Consultant senior en Logiciels Libres Debian developer ([EMAIL PROTECTED])Développeur Debian -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: INSTALL REPORT with 2004-01-09 netinst, failed
As I attempt to do more not-so-uncommon things (with the 2004-01-03 snapshot)... - I initially have a HD with a single ext3 partition + swap (from a previous RH install). If I repartition by deleting the ext3 and repartitionning that space in 2, on the partition setup screen I notice the 1st partition is already listed as ext3, and the 2nd not. Then I realize it comes from the previous (RH) install. I had not realized that when just selecting a single partition, when I thought "ext3" was just a default. Now if my original single-partition-over-RH had gone throught the reboot stage, I may surely have ended up with a strange install. And indeed the nc binary I had noticed, as well as the __libc_stack_end issues probably derive from that. -> Maybe the partition-setup screen could emphasize this issue ? -> How about forcing system partitions (/, /usr(/*)?, /var(/*)?) to be formatted, or at least require confirmation ? -> Is it really a good idea to use those programs just installed before rebooting ? - Now when I decide to split the existing partition in 2, but do not realize I must explicitely select "format as ext3", since my partition already appears as having ext3 selected (I thought), and go on, I'm brought back to the main menu on the next step, and requesting to install the base system brings me back there again almost instantly. Only by looking at syslog I finally notice a "read beyond end of device" attempt, which finally enlighten me on what's going on. -> the user should be notified of such errors by an error box. -> What about forcing the user to reformat such partitions, which we can decide must be invalid because they just reuse an existing block ? Or, probably better, when destroying a partition, we (cfdisk ?) could ask the user whether to nuke the 1st sectors, to avoid further misdetection. - Once base packages are unpacked, while setting up lilo, I'm now presented with a textbox containing the devfs name for the disk. If I want to install lilo in another location, it looks like I have to know the devfs path, whereas in the old installer I could choose in a list, which is much more suited to people not knowing devfs BTW, is it wise at all to use devfs, which is tagged as OBSOLETE in 2.6 !? - Selecting "French" as the keyboard layout after reboot, I do not have time to see the keymap name, and subsequent error message, before the next dialog screen. And I end up with a querty-looking layout. Looking at base-config.log, it appears that the selected layout is mac-usb-fr (!). That reminds me of a long-gone issue in console-tools debconf settings. Is it possible that some old config-scripts/templates are kept around ? Just after that, I get "KDGKBENT at index 128 in table 8: invalid argument", "failed to dump keymap" and "not loading keymap" messages, which may explain why I still have a qwerty layout. In the meantime, I was asked to enter the root password... and did that using the qwerty layout, which is sure to give me an unretrievable password when I change back to the correct layout (ie. fr-pc). If then I attempt to "loadkeys fr-pc" it works fine, but if I "install-keymap fr-pc" I get the same error messages. - When asked to select (again) an http mirror for downloading packages, 1) my previous selection is not the default, maybe because of 2) the country names being listed in english, despite all other text being correctly in french. However, the proxy information was correctly memorized. Regards, -- Yann Dirson <[EMAIL PROTECTED]> http://www.alcove.com/ Technical support managerResponsable de l'assistance technique Senior Free-Software Consultant Consultant senior en Logiciels Libres Debian developer ([EMAIL PROTECTED])Développeur Debian -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: INSTALL REPORT with 2004-01-09 netinst, failed
Ah, I forgot to fill some boxes On Mon, Jan 12, 2004 at 12:46:01PM +0100, ydirson wrote: > Memory: 192MB > Root Size/partition table: single ~9GB root, 512M swap And to mention yet other problems: - the items in the initial language menu all start with "choose this" or a translation of it, which makes it impossible to quickly jump to a particular language with one keystroke, as was possible in old installer - in the list of country to select mirrors from (and possibly other lists), when using french as a language, the coutries are not sorted according to their translation, but appear to be sorted by their english name, which is quite awkward - in the step where one assigns fstype/mountpoints to partitions, when one selects a swap partition, and just validates current choice for fstype, the swap partiton is assigned a mountpoint matching the last mountpoint assigned (eg. "/", which then causes an error because of duplicate / being found) -- Yann Dirson <[EMAIL PROTECTED]> http://www.alcove.com/ Technical support managerResponsable de l'assistance technique Senior Free-Software Consultant Consultant senior en Logiciels Libres Debian developer ([EMAIL PROTECTED])Développeur Debian -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: INSTALL REPORT with 2004-01-09 netinst, failed
On Mon, Jan 12, 2004 at 12:46:01PM +0100, ydirson wrote: > Comments/Problems: > > Partition hard drives: > > if I just follow the steps lineraly, no problem. At first, I had hit > cancel at some point before (have to check and reproduce), and the HD > selection menu was incomplete (did not mention the probbedname of the > drive, only the devfs name, and no "terminate" choice). When > selecting the drive, all I could see was the blue background - could > Ctrl-C out of it, but could not achieve partitionning at all. I could narrow the problem, here are the steps to reproduce. You'll need a broken mirror for this to work (at least a mirror with not all of the files required to download the initial packages): - when 1st asked to select a mirror, enter informations manually to point to your local incomplete mirror (FWIW there is a proxy between the test machine and the mirror). - when the download step fails, go back to mirror selection (note: we have here Z choices "go back" and "continue", both of which lead into the main menu to the next step, despite the error). - then select a valid one (I use ftp2.de.d.o) - download completes fine, and the list of discs to partition is presented, featuring the problem described above I could find nothing obviously alarming on this subject in syslog, although there are many "broken dependency" reported all around. I saved the syslog through nc (hopefully installed by a previous run on the target HD, kudos to you for having put it here ;) Syslog filtered for "partition" reads the following - full syslog available on request: [...] Jan 12 13:02:25 (none) user.debug main-menu[227]: DEBUG: search for package resolving mounted-partitions, dependency from base-installer Jan 12 13:02:25 (none) user.info main-menu[227]: INFO: broken dependency: debootstrap-udeb to mounted-partitions Jan 12 13:02:25 (none) user.info main-menu[227]: INFO: Falling back to the package description for di-utils-partitioner [...] Jan 12 13:02:25 (none) user.debug main-menu[227]: DEBUG: Menu item 'di-utils-partitioner' selected Jan 12 13:02:25 (none) user.debug main-menu[227]: DEBUG: configure di-utils-partitioner, status: 2 Jan 12 13:02:25 (none) user.warn main-menu[227]: WARNING **: Unable to set title for di-utils-partitioner. HTH, -- Yann Dirson <[EMAIL PROTECTED]> http://www.alcove.com/ Technical support managerResponsable de l'assistance technique Senior Free-Software Consultant Consultant senior en Logiciels Libres Debian developer ([EMAIL PROTECTED])Développeur Debian -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
INSTALL REPORT with 2004-01-09 netinst, failed
INSTALL REPORT Debian-installer-version: http://gluck.debian.org/~manty/testing/netinst/i386/20040109/sarge-i386-netinst.iso uname -a: Date: 2004-01-12 Method: boot off CD, "net" boot, proxied http, sarge install Machine: custom Processor: celeron 350 Memory: XXX Root Device: IDE (hda1) Root Size/partition table: single 9GB root, XXX swap #Output of lspci: Base System Installation Checklist: Initial boot worked:[O] Configure network HW: [O] Config network: [O] Detect CD: [ ] Load installer modules: [ ] Detect hard drives: [O] Partition hard drives: [E] Create file systems:[O] Mount partitions: [ ] Install base system:[E] Install boot loader:[ ] Reboot: [ ] [O] = OK, [E] = Error (please elaborate below), [ ] = didn't try it Comments/Problems: Partition hard drives: if I just follow the steps lineraly, no problem. At first, I had hit cancel at some point before (have to check and reproduce), and the HD selection menu was incomplete (did not mention the probbedname of the drive, only the devfs name, and no "terminate" choice). When selecting the drive, all I could see was the blue background - could Ctrl-C out of it, but could not achieve partitionning at all. Install base system: undefined symbol __libc_stack_end from "mount" when run to mount /target/proc in the chroot. The same happens with /bin/sh when running "chroot /target" in VC#2. This symbol was IIRC exported by old glibc dropped when the cleaned up the exported symbols from undocumented/internal ones, so I suppose the glibc on the CD is a post-cleanup one, while the problematic binaries (mount, sh) were built on an old glibc. Now why did they get this symbol ? I know libgc had it, but that was just because its build procedure explicitely tested for it and used it if available. And the current /bin/mount I have on another sarge box does not have this symbol. -- Yann Dirson <[EMAIL PROTECTED]> http://www.alcove.com/ Technical support managerResponsable de l'assistance technique Senior Free-Software Consultant Consultant senior en Logiciels Libres Debian developer ([EMAIL PROTECTED])Développeur Debian -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Bug#183453: e2fsprogs-bf: Please provide a lintian override for man pages
On Tue, Mar 04, 2003 at 04:06:06PM -0800, Micah Anderson wrote: > http://qa.debian.org/man-pages.html is generated based on the regular > lintian reports at http://lintian.debian.org/. Listed on this page are > all the manpages that normally would accompany e2fsprogs-bf. However, > the man pages have been removed from this package because the package > is intended to be used on a floppy, so anything extra goes. The > lintian report gives an error, but this is a lintian false positive. > > If you were to add a lintian override for this error, it would > disappear from the QA web pages automatically. Simply add one by > supplying a /usr/share/llintian/overrides/e2fsprogs-bf file as > detailed in http://lintian.debian.org/manual/ch2.html#s2.4 Since it's a generic rule that "*-bf" packages should not contain manpages, I'm of the opinion that a new rule should be added to lintian, rather than adding overrides to all *-bf packages. Further more, since I don't think the new debian-installer is using those *-bf packages, but uses udeb's instead, those *-bf packages may be completely obsolete already. Did you discuss the issue first with the lintian maintainer and the bf team ? Regards, -- Yann Dirson<[EMAIL PROTECTED]> |Why make M$-Bill richer & richer ? Debian-related: <[EMAIL PROTECTED]> | Support Debian GNU/Linux: Pro:<[EMAIL PROTECTED]> | Freedom, Power, Stability, Gratuity http://ydirson.free.fr/| Check <http://www.debian.org/> -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#161999: boot-floppies: 3.0r0 install can end with no /etc/host{s,name}
On Tue, Sep 24, 2002 at 11:34:12AM +0200, Eduard Bloch wrote: > /etc/hostname may not be written if you did not run "Configure the > network" nor the "Configure hostname" steps successfully. I did run the "configure hostname step", at least. I don't have much time tonight, will try to test things tomorrow. Regards, -- Yann Dirson<[EMAIL PROTECTED]> |Why make M$-Bill richer & richer ? Debian-related: <[EMAIL PROTECTED]> | Support Debian GNU/Linux: Pro:<[EMAIL PROTECTED]> | Freedom, Power, Stability, Gratuity http://ydirson.free.fr/| Check <http://www.debian.org/> -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#161999: boot-floppies: 3.0r0 install can end with no /etc/host{s,name}
Package: boot-floppies Version: Debian 3.0r0 Severity: serious When installing woody from CD on a laptop yesterday, the installation ended without an /etc/hosts nor an /etc/hostname. Hm, maybe there are others, but I did not noticed yet. Not sure it is the source of the problem, but I had asked for a DHCP setup, while not connected to a network (something that I thought would be reasonable to ask, but for some reason it appeared that was not allowed), and after DHCP failure I had to skip the "network config" step. -- System Information Debian Release: 3.0 Architecture: i386 Kernel: Linux vitel 2.4.19-station #1 Tue Aug 6 11:36:58 CEST 2002 i686 Locale: LANG=C, LC_CTYPE= -- Yann Dirson <[EMAIL PROTECTED]> http://www.alcove.com/ Technical support managerResponsable de l'assistance technique Senior Free-Software Consultant Consultant senior en Logiciels Libres Debian developer ([EMAIL PROTECTED])Développeur Debian -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#161998: boot-floppies: wrong fr translations
Package: boot-floppies Version: Debian 3.0r0 Severity: important X-Debbugs-Cc: [EMAIL PROTECTED] I noticed a couple of translation errors in the french l10n, one of which is WILL make user understand the contrary of reality. -> when formatting an ext3 fs, the message says that ext3 is "seulement pour noyaux 2.2", which means "for 2.2 only", which is obviously wrong. -> when installing the base system, the dialog title says "installer le système de base" ("install base system"). This should be "installation du système de base" ("installing base system"). -> in the "make system bootable" help text, the 3 choices are wrongly referenced as "suivant", "suivant", "autre choix 1", instead of "suivant", "autre choix", "autre choix 1". BTW, shouldn't the 3rd be "autre choix 2" ? -- System Information Debian Release: 3.0 Architecture: i386 Kernel: Linux vitel 2.4.19-station #1 Tue Aug 6 11:36:58 CEST 2002 i686 Locale: LANG=C, LC_CTYPE= -- Yann Dirson <[EMAIL PROTECTED]> http://www.alcove.com/ Technical support managerResponsable de l'assistance technique Senior Free-Software Consultant Consultant senior en Logiciels Libres Debian developer ([EMAIL PROTECTED])Développeur Debian -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
sarge, console-tools, debconf or not
On Mon, Jul 22, 2002 at 03:43:55PM +0400, Wartan Hachaturow wrote: > On Mon, Jul 22, 2002 at 01:13:54PM +0200, Alastair McKinstry wrote: > > > So do you have any plans about what to do for console tools in the next > > Debian release? Do you need help ? > > Actually, I've asked people at debian-devel for the things they want > in console-tools-v2. Short answer is: Nobody cares. Or they are just happy with the current way it's done :) > So, I'm feeling free to implement what I like ;) > The basic idea is as follows: > > 1. Split console-data on multiple packages. > >That is, each package would contain keymaps.list (and, possibly, >fonts.list), describing included keymaps in a way similar to >Packages.gz. >This list would be scanned upon the build, and the packages like >console-data-cyrillic, console-data-ethiopic, etc. would be >created. > > 2. Throw away or minimize debconf usage. > >As you may see, current console-common code is a nightmare and too >tightly integrated with debconf. That leads to numerous bugs (due >to broken databases at some people's stations, etc., etc.). >My idea is to throw away debconf, and implement something like >default keymap for each console-data package, and show people a >debconf note on what to run to change a keymap (a special utility >like consoletoolsconfig could be written for that matter). >For those that really like to have debconf for the >I-want-to-have-my-cluster-machines-all-the-same purpose, there's an >idea to implement something like make-kpkg -- people would build >their own, custom console packages, and install them wherever they >want -- that system can also be used while building "distro" >packages. > > These are two major ideas. Of course, that would lead to a complete rewrite > of the current console configuration subsystem -- but that's A Good Thing > (tm), imo ;) I completely agree with the 1st point, but in a world where all Debian packages are moving towards debconf, the 2nd really looks like a step backwards. Despite the problems caused when building the debconf script, I got a number of people happy with the thing. And if we have multiple ways of handling configuration, we just shift the nightmare to people writing installers (hence my CC to debian-boot list and FAI maintainer). My analysis of why it is a nightmare is quite related with the fact it's too tightly integrated with debconf: I started to write the code on an ad-hoc basis - some sort of RAD - and progressively started to abstract things. What I'd have liked to see is this abstraction process turn into some sort of hierarchical-select debconf widget. That is, effectively completely de-coupling the handling of the tree-like classification from the console stuff proper, and in the end having some sort of "Type: hierarchical-select" available in debconf. Unfortunately, my attempts to convince the debconf maintainer we should go this way were vain - I do not recall reading a word of encouragement from him, nor any suggestion as to how to do it another way. Yet I don't see how it could get done properly within debconf otherwise. What I'd suggest would be first to check with Joey Hess his opinions on the subject - I know he would be of valuable help if convinced - before going another way. Opinions, anyone ? -- Yann Dirson<[EMAIL PROTECTED]> |Why make M$-Bill richer & richer ? Debian-related: <[EMAIL PROTECTED]> | Support Debian GNU/Linux: Pro:<[EMAIL PROTECTED]> | Freedom, Power, Stability, Gratuity http://ydirson.free.fr/| Check <http://www.debian.org/> -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
pcmcia_core module missing from woody installer ?
Hi, Looks like the pcmcia_core module is missing from the woody installer ? Regards, -- Yann Dirson <[EMAIL PROTECTED]> http://www.alcove.com/ Technical support managerResponsable de l'assistance technique Senior Free-Software Consultant Consultant senior en Logiciels Libres Debian developer ([EMAIL PROTECTED])Développeur Debian -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: e2fsprogs 1.26
On Mon, Feb 04, 2002 at 12:56:40PM -0500, Ben Collins wrote: > Seems it would just break things, more than solve them. Hm, maybe. Even more, my yesterday attempt showed that there are at least one versionned dep against e2fsprogs, from sysvinit. So until we have support for versionned provides (it's in CVS only, right ?), it's not possible to do it that way. Let's reconsider simply upgrading the main package: AJ wrote: > Am I correct in thinking that e2fsprogs doesn't provide any shared > libraries, or anything else that might cause packages built with the > new e2fsprogs to not work on systems with the old e2fsprogs? Assuming > that's the case... > > I don't think it should be a problem if you just upload e2fsprogs 1.26 > (same package, new version). It'll take 20 days of unstable users beating > on it before it's let into woody assuming you set the urgency to "low", > which should be enough time to notice any severe problems with it. Are your concerns related to the possibility that the packages build-depending on e2fsprogs libs would not be rebuilt ? That's 14 source packages (autoinstall-i386 bonobo delo dump ext2resize geas gnome-utils mc parted quik red-carpet silo xfsdump xfsprogs), 13 unique maintainers inclusing 'Debian QA Group' (quik is orphaned). Most of them have build-depend on uuid-dev, which AFAIK has not changed. That leaves 6 packages really needing rebuild (delo dump gnome-utils mc quik silo). IMHO, that'd be possible to achieve. Green light ? Red light ? -- Yann Dirson <[EMAIL PROTECTED]> http://www.alcove.com/ Free-Software EngineerIngénieur Logiciel-Libre Free-Software time manager Responsable du temps Informatique-Libre Debian GNU/Linux developper <[EMAIL PROTECTED]> -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: e2fsprogs 1.26
On Mon, Feb 04, 2002 at 12:01:12PM -0500, Ben Collins wrote: > > So I suppose it won't be OK, then... better provide an alternative > > package ? > > Better to not provide an alternative. Woops, wrong wording :) Sure ! Replaces/Conflicts/Provides -- Yann Dirson <[EMAIL PROTECTED]> http://www.alcove.com/ Free-Software EngineerIngénieur Logiciel-Libre Free-Software time manager Responsable du temps Informatique-Libre Debian GNU/Linux developper <[EMAIL PROTECTED]> -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: e2fsprogs 1.26
On Tue, Feb 05, 2002 at 02:07:43AM +1000, Anthony Towns wrote: > On Mon, Feb 04, 2002 at 10:39:25AM +0100, Yann Dirson wrote: > > e2fsprogs 1.26 is out, including fixes for one but tagged important, > > ext3-related fixes, reiserfs support in fsck, and more complete > > support for filesystems in >2GB files. > > > > I suppose it is too late to get this one in base, so I can package it > > in an e2fsprogs-1.26 package if needed. However, it would probably be > > better if it could go in base... What do boot-floppies and release > > manager think ? > > Am I correct in thinking that e2fsprogs doesn't provide any shared > libraries wrong - packages depending on the libs, from memory: dump, mc, some undeletion progs, maybe other things. > Assuming that's the case... So I suppose it won't be OK, then... better provide an alternative package ? -- Yann Dirson <[EMAIL PROTECTED]> http://www.alcove.com/ Free-Software EngineerIngénieur Logiciel-Libre Free-Software time manager Responsable du temps Informatique-Libre Debian GNU/Linux developper <[EMAIL PROTECTED]> -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
e2fsprogs 1.26
e2fsprogs 1.26 is out, including fixes for one but tagged important, ext3-related fixes, reiserfs support in fsck, and more complete support for filesystems in >2GB files. I suppose it is too late to get this one in base, so I can package it in an e2fsprogs-1.26 package if needed. However, it would probably be better if it could go in base... What do boot-floppies and release manager think ? Full list of changes at http://sourceforge.net/project/shownotes.php?release_id=73260 -- Yann Dirson <[EMAIL PROTECTED]> http://www.alcove.com/ Free-Software EngineerIngénieur Logiciel-Libre Free-Software time manager Responsable du temps Informatique-Libre Debian GNU/Linux developper <[EMAIL PROTECTED]> -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: more space spared in e2fsprogs-bf
On Fri, Jun 22, 2001 at 12:41:55AM -0400, Adam Di Carlo wrote: > Yann Dirson <[EMAIL PROTECTED]> writes: > > > Unless you have objections, I'll build e2fsprogs-bf with -Os. > > Probably many other packages would benefit from this flag as well... > > Sounds good to me. > > Oh, I heard a rumor this package is having trouble building in > non-i386 arches. Hopefully you can look at addressing that quickly, > because I can't really adopt this package in boot-floppies until it's > built for all arches. Ted has been given an account on vore to look at it. -- Yann Dirson<[EMAIL PROTECTED]> |Why make M$-Bill richer & richer ? debian-email: <[EMAIL PROTECTED]> | Support Debian GNU/Linux: | Cheaper, more Powerful, more Stable ! http://ydirson.free.fr/ | Check <http://www.debian.org/> -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
more space spared in e2fsprogs-bf
Hi all, I wanted to checked what "gcc -Os" could do to help us in our Holy Cruisade for Disk Space. Here are my results for i386: e2fsprogs-1.21+1.22-WIP-0620/debian[682]$ du -sb e2fsprogs-bf-* 364544 e2fsprogs-bf-O2 356352 e2fsprogs-bf-Os e2fsprogs-1.21+1.22-WIP-0620/debian[683]$ find e2fsprogs-bf-O2 -type f | xargs cat | wc -c 499620 e2fsprogs-1.21+1.22-WIP-0620/debian[684]$ find e2fsprogs-bf-Os -type f | xargs cat | wc -c 496904 /me still wonders in which files there are holes, that could account for the fact that the raw files are larger than the space that they, their symlinks, and their dir tree take on my 4k-blocks filesystem, but clearly there are a couple of KBs to take here. Unless you have objections, I'll build e2fsprogs-bf with -Os. Probably many other packages would benefit from this flag as well... Regards, -- Yann Dirson<[EMAIL PROTECTED]> |Why make M$-Bill richer & richer ? debian-email: <[EMAIL PROTECTED]> | Support Debian GNU/Linux: | Cheaper, more Powerful, more Stable ! http://ydirson.free.fr/ | Check <http://www.debian.org/> -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Patch not included in e2fsprogs 1.21.
On Tue, Jun 19, 2001 at 01:41:53PM -0400, Adam Di Carlo wrote: > I don't really see the problem with having an e2fsprogs-bf package, > just for root.bin, which no one should ever normally install... Well, my fear was that people would install e2fsprogs-bf on their system, but now that I think of it, e2fsprogs is Essential, so it can't be replaced by non-Essential e2fsprogs-bf, so life is cool. Will do that. Regards, -- Yann Dirson<[EMAIL PROTECTED]> |Why make M$-Bill richer & richer ? debian-email: <[EMAIL PROTECTED]> | Support Debian GNU/Linux: | Cheaper, more Powerful, more Stable ! http://ydirson.free.fr/ | Check <http://www.debian.org/> -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Patch not included in e2fsprogs 1.21.
On Sat, Jun 16, 2001 at 02:28:08PM +0100, Philip Blundell wrote: > Actually, for boot-floppies' purposes it would be slightly better to build > the shared library with non-PIC objects. You may not have done it > intentionally but it sounds like you got the right result. :-) Hm... let me know when you're sure :) As for the packaging of those libs and of the binaries that were reduced, I'm not sure what would be best. IIRC most stuff included in boot-floppies is taken right from real packages installed on the machine where they are built. The e2fsprogs-bf I've built contains stuff in e2fsprogs, so mush conflict with him to be installed, which 1) is a pain for people building boot-floppies, 2) creates a package that noone should ever installed except for building boot-floppies. Would it be OK to put them in some /usr/lib/boot-floppies so that they won't interfere with what is installed ? Or to put that in a tar.gz in the pool, where the boot-gloppies built will know to find it ? Or wouldn't it be best to have boot-floppies "apt-get source e2fsprogs" and run "debian/rules build-bf" instead ? It would require some tweeking so that the bf package does not get built and uploaded into the dist, but it may be worth it. Regards, -- Yann Dirson<[EMAIL PROTECTED]> |Why make M$-Bill richer & richer ? debian-email: <[EMAIL PROTECTED]> | Support Debian GNU/Linux: | Cheaper, more Powerful, more Stable ! http://ydirson.free.fr/ | Check <http://www.debian.org/> -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Patch not included in e2fsprogs 1.21.
[cc-ing boot-floppies team to let them know] On Fri, Jun 15, 2001 at 07:44:43PM -0400, [EMAIL PROTECTED] wrote: > I've rejected this patch, for two reasons. > > (1) It's not needed any more since you can now configure e2fsprogs to > generate a "lite" version of libext2fs.so.2.4. OK I'll replace that with with an e2fsprogs-bf package. > (2) The patch is busted (not correct). The object files being placed > into the libext2fs_pic.a file are the non-shareable .o files, not > the PIC .o files. You need to cd into the elfshared subdirectory to > get at the PIC files, and the Makefile fragment doesn't do that. Woops... I should really read how vars are used before assuming they hold a full path to files... :} > I've left the _pic rules in the debian subdirectory, but in my opinion > those should replaced with rules that do a separate build in a second > directory using the --disable-swapfs, --disable-debugfs, et. al. flags. You might as well remove the refs to *pic from the control and rules files, or the build will fail because of the missing pic targets in the Makefile's. Regards, -- Yann Dirson<[EMAIL PROTECTED]> |Why make M$-Bill richer & richer ? debian-email: <[EMAIL PROTECTED]> | Support Debian GNU/Linux: | Cheaper, more Powerful, more Stable ! http://ydirson.free.fr/ | Check <http://www.debian.org/> -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: E2fsprogs 1.20 released
On Thu, Jun 14, 2001 at 03:49:30AM -0400, Theodore Tso wrote: > The e2fsprogs 1.21-WIP-0614 release which I just uploaded has some new > configure options; if you compile it: > configure --enable-elf-shlibs --disable-swapfs --disable-imager --disable-resizer >--disable-debugfs '--with-ccopts=-O2 -fomit-frame-pointer' Gasp, I had not realized that, the 1st 1.21-WIP-0614 package doesn't make use of this. > Would this be workable for you? I can have the pics libs and an e2fsck.bf binary built reduced, and shipped in the e2fslibs-pic package with not much work. Anyway the only breakage that can arise here by using the reduced libs will be an undefined symbol when a non-supported program will start, right ? My understanding of how the shared libs work lets me think there should not be other problems, but well... -- Yann Dirson<[EMAIL PROTECTED]> |Why make M$-Bill richer & richer ? debian-email: <[EMAIL PROTECTED]> | Support Debian GNU/Linux: | Cheaper, more Powerful, more Stable ! http://ydirson.free.fr/ | Check <http://www.debian.org/> -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: E2fsprogs 1.20 released
[CCing the boot-floppies team] On Thu, Jun 07, 2001 at 08:49:54AM -0400, Theodore Tso wrote: > On Thu, Jun 07, 2001 at 10:24:54AM +0200, Yann Dirson wrote: > > > > FWIW, I've added support for pic libraries to the elf-lib Makefile, on > > request of the boot-floopies guys, and was waiting for their feedback to > > forward it to you (quite trivial addition, and without interactions with > > the other things anyway). I'm still curious to see if they'll manage to > > reduce the shared libs size with this - in case you don't know, it's an > > ar archive build with the same .o files as the shared lib, used to build > > a stripped-down shared lib with only the needed parts for the tools in > > the boot-floppies. My guess is that mke2fs and e2fsck use most of the > > shared libs anyway, so size reduction would not be so great - do you > > confirm ? > > As long as they don't care about needing debugfs, they can probably > save 12k, maybe as much as 15k. How desperate for space are they? > > Is the boot floppy also supposed to double as a rescue floppy? If so, > removing debugfs might not be such a hot idea. IIRC it's the case. > The other way of doing this if they're really desparate for space > would be to make a new configure option, --enable-small-subset which > built a version of e2fsprogs that had been cut down for use on > installation floppies. This allows further cuts, such as not > supporting byte-swapping filesystems (needed to convert very old > Linux-PPC filesystems to the standard ext2 byte order; doing this > would save approximately 3k from e2fsck and libext2fs). This would > also not build debugfs and resize2fs, and automatically contain the > right list of functions that could be removed from libext2fs (as well > as using #ifdef to remove functions with .o files that can't be used). > I could remove all use of inline functions from e2fsck, which will slow > it down some, but make it quite a bit smaller. > > So if they're truly desparate for space, it may be possible to save a > some more by making much more invasive changes to e2fsprogs. > > If they're in a big hurry, I'd advise doing this as a new debian > package with a diff file that modified lib/ext2fs/Makefile.in to > remove certain files, removed debugfs and resize2fs from the list of > directories to be built, etc. Probably it can be handled in one source package, using a modified version of DBS (the thing used to build X and such). I was looking for an excuse to make a jam-based version of DBS, cool :) > If they're going to be code slipping for another six months anyway > :-), I can look into making the mods into e2fsprogs in a "clean" way, > i.e., using configure options to control how it's built. investigating... plans for base system appear to be: - fix release-critical bugs before July 1st - hard freeze July 20th -- Yann Dirson<[EMAIL PROTECTED]> |Why make M$-Bill richer & richer ? debian-email: <[EMAIL PROTECTED]> | Support Debian GNU/Linux: | Cheaper, more Powerful, more Stable ! http://ydirson.free.fr/ | Check <http://www.debian.org/> -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Bug#99285: would be nice to produce a -pic version of the libraries
On Wed, May 30, 2001 at 06:44:40PM -0700, David Whedon wrote: > After surveying other packages that have reduced libraries as well as looking > into mklibs.sh itself I made a patch for libdetect to make a pic lib. I don't > produce anything other than the *_pic.a and mklibs.sh is happy. > > http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=85638&repeatmerged=yes > > > from the patch: > + # This builds the pic library > + ar cqv $(topdir)/src/libdetect_pic.a $(topdir)/src/*.lo Hmm... that's .lo files... I suppose it doesn't make a difference to put .o files where libtool is not in use :) For e2fsprogs the minimal patch will be in upstream sources - you guys will tell me if the saving are worth sending the change upstream. Regards, -- Yann Dirson<[EMAIL PROTECTED]> |Why make M$-Bill richer & richer ? debian-email: <[EMAIL PROTECTED]> | Support Debian GNU/Linux: | Cheaper, more Powerful, more Stable ! http://ydirson.free.fr/ | Check <http://www.debian.org/> -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Bug#99285: would be nice to produce a -pic version of the libraries
On Wed, May 30, 2001 at 02:55:55PM -0400, Adam Di Carlo wrote: > Would it be possible to create a new package, say, e2fslibs-pic and > comerr-pic ? We use these libraries in the rootdisk of the boot > floppies, and we are unable to reduce libext2fs.so, libe2p.so, and > libcom_err.so since there are no PIC archives available for them. Hm... I wonder how much we could gain - I guess most of those libs are used by mke2fs and e2fsck alone... Anyway, I just had a look at the Policy doc for minimal guidelines, but did not find anything there (or elsewhere). Is there a common way of doing things, or does each package make its own cooking ? Eg, glibc has some other files in addition to the pic lib... -- Yann Dirson<[EMAIL PROTECTED]> |Why make M$-Bill richer & richer ? debian-email: <[EMAIL PROTECTED]> | Support Debian GNU/Linux: | Cheaper, more Powerful, more Stable ! http://ydirson.free.fr/ | Check <http://www.debian.org/> -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#75106: dbootstrap: wrong choice for partition for LILO bootsector
On Mon, Dec 04, 2000 at 11:12:18AM -0500, Adam Di Carlo wrote: > Ouch. Yes. That's not unlike what I said when I realized what I was going to ask the installer to do :) > I have no clue why this was happening. For some reason it's > thinking that /dev/hdc1 is your root partition. > > Is it really feasible to have /boot and / be different partitions? > Doesn't /boot contain the kernels? That's it. It's even in the suggested mountpoints in the potato install. The reason most probably is to make sure the installed kernels are accessible from the bios at boot-time - that's why I put /boot at the beginning of the drive, anyway. And yes, it's really feasible. I installed at least 3 potato boxes with this setup, with each time this quite annoying choice: install lilo either in the mbr, or in a place where it won't be able to boot :) Regards, -- Yann Dirson<[EMAIL PROTECTED]> |Why make M$-Bill richer & richer ? debian-email: <[EMAIL PROTECTED]> | Support Debian GNU/Linux: | Cheaper, more Powerful, more Stable ! http://ydirson.free.fr/ | Check <http://www.debian.org/> -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#75106: dbootstrap: wrong choice for partition for LILO bootsector
Package: boot-floppies Version: N/A; potato RC3 Severity: normal I did an install from my RC3 floppies this week - still do not have r1 CDs, waiting for r2 to be there :} - and went into the following problem. It was not a blocking problem for me, but I guess it may be to others. Partionning: /dev/hda unused (but has old OS and std M$ mbr) /dev/hdc1 /boot /dev/hdc2 swap /dev/hdc3 / /dev/hdc4 /local As answers to dbootstrap, as 1st partition to mke2fs I selected hdc3, mounted it as root, and then mke2fs'd the others. When it came to LILO setup, I only had 2 choices for LILO location: hda MBR - well, OK, but I would have prefered hdc MBR, to keep the old one. Any reason why it's not there ? My bios does support booting from ide1 ! I tested it successfully ! hdc1- er... yes. Bad guess. I suspect this can be due to a wrong assumption like "/ is the first ext2", but I suspect many people still put /boot first ! Strange that it did not picked up hdc3 here. -- System Information Debian Release: woody Architecture: i386 Kernel: Linux bylbo 2.2.17 #1 Sun Sep 10 19:21:52 CEST 2000 i586 -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#68373: fresh install: some dependency problems ?
On Sat, Aug 12, 2000 at 07:01:38PM -0400, Adam Di Carlo wrote: > Hmm. This isn't a boot-floppies bug. It's either a bug in the > individual packages you mentioned (i.e., missing deps) or else an apt > bug. I guess so. Maybe some chroot environment could be used to track potential dependancy issues. Maybe I'll find some time to do this - but maybe not ;) -- Yann Dirson<[EMAIL PROTECTED]> |Why make M$-Bill richer & richer ? debian-email: <[EMAIL PROTECTED]> | Support Debian GNU/Linux: | Cheaper, more Powerful, more Stable ! http://ydirson.free.fr/ | Check <http://www.debian.org/> -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#68373: fresh install: some dependency problems ?
Package: boot-floppies Version: 2.2.16 Severity: normal Another small issue I had was that after a first pass at installing packages, apt told me it had problems unpacking some packages, and that it could not install or configure (was not clear, but I'd say it was a configuration problem, as only package names were given, not full package filenames) the following: gmc, rstart, rstartd. It then asked me to press enter, went on configuring packages, and the 3 ones there went OK. It then told me some package installs went wrong and suggested to retry, which I accepted... just to be told no more packages were to be installed :) I had chosen "simple" as install choice and was asked for task packages. I selected the following ones (listing is from 'ls -d /usr/doc/task*'): task-c-dev task-gnome-desktop task-x-window-system task-devel-common task-gnome-gamestask-x-window-system-core (indeed task-gnome-desktop and task-x-window-system-core were automatically added) I'll keep this test install partition around for a while. I'm not sure how I can track the installation order, which would have probably helped to spot what looks like missing deps. Maybe the timestamps on dpkg/info/*.list files will be enough, but if anyone has a better idea, I'll take it. Uh, I forgot: nice work, guys ! :) -- System Information Debian Release: 2.2 Architecture: i386 Kernel: Linux bylbo 2.2.14 #1 sam avr 22 03:24:31 CEST 2000 i586 Versions of packages boot-floppies depends on: hi ash 0.3.5-11 NetBSD /bin/sh ii bison 1:1.28-5 A parser generator that is compati ii cslatex 1.2.1LaTeX for Czech/Slovak typesetting ii cspsfonts 0.2.1Support for Czech and Slovak Type1 hi debhelper 2.0.86 helper programs for debian/rules ii debiandoc-sgml 1.1.42 DebianDoc SGML DTD and formatting ii dpkg-awk1.0.0Gawk script to parse /var/lib/dpkg ii dpkg-dev1.6.14 Package building tools for Debian ii file3.28-1 Determines file type using "magic" ii flex2.5.4a-6 A fast lexical analyzer generator. ii gawk1:3.0.4-4GNU awk, a pattern scanning and pr hi gcc 1:2.95.2-13 The GNU C compiler. hi gettext 0.10.35-13 GNU Internationalization utilities ii libc6-pic [glibc-pic] 2.1.3-10 GNU C Library: PIC archive library ii libgd1g-dev 1.7.3-0.1GD Graphics Library (development v hi libi18n-langtags-perl 0.12-6 Perl module for dealing with RFC17 hi libnewt-dev 0.50-7 Developer's toolkit for newt windo ii libpaperg 1.0.3-13 Library for handling paper charact hi libpopt-dev 1.4-1.1 lib for parsing cmdline parameters hi libwww-perl 5.47-1 WWW client/server library for Perl ii m4 1.4-10 a macro processing language hi make3.78.1-8 The GNU version of the "make" util ii makedev 2.3.1-44 Creates special device files in /d hi man-db 2.3.16-1 Display the on-line manual. ii perl-5.005 [perl5] 5.005.03-7.1 Larry Wall's Practical Extracting ii pointerize 0.4 Internationalization utilities, ba ii recode 3.5-1Character set conversion utility. hi rsync 2.3.2-1.2fast remote file copy program (lik hi slang1-pic 1.3.9-1 The S-Lang programming library, sh ii tetex-bin 1.0.7-2 teTeX binary files ii tetex-extra 1.0.2-2 extra teTeX library files ii zlib1g-dev 1:1.1.3-7compression library - development -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#68372: fresh install: xdm does not start
Package: boot-floppies Version: 2.2.16 Severity: normal Just had my first install of potato TC3. No major problems. The main one was that xdm did not start at the end of the install procedure - I was left at the consoel prompt. The xdm package (and X server) was properly configured, and xdm started as expected after rebooting. Probably confusing for the newcomer, though. -- System Information Debian Release: 2.2 Architecture: i386 Kernel: Linux bylbo 2.2.14 #1 sam avr 22 03:24:31 CEST 2000 i586 Versions of packages boot-floppies depends on: hi ash 0.3.5-11 NetBSD /bin/sh ii bison 1:1.28-5 A parser generator that is compati ii cslatex 1.2.1LaTeX for Czech/Slovak typesetting ii cspsfonts 0.2.1Support for Czech and Slovak Type1 hi debhelper 2.0.86 helper programs for debian/rules ii debiandoc-sgml 1.1.42 DebianDoc SGML DTD and formatting ii dpkg-awk1.0.0Gawk script to parse /var/lib/dpkg ii dpkg-dev1.6.14 Package building tools for Debian ii file3.28-1 Determines file type using "magic" ii flex2.5.4a-6 A fast lexical analyzer generator. ii gawk1:3.0.4-4GNU awk, a pattern scanning and pr hi gcc 1:2.95.2-13 The GNU C compiler. hi gettext 0.10.35-13 GNU Internationalization utilities ii libc6-pic [glibc-pic] 2.1.3-10 GNU C Library: PIC archive library ii libgd1g-dev 1.7.3-0.1GD Graphics Library (development v hi libi18n-langtags-perl 0.12-6 Perl module for dealing with RFC17 hi libnewt-dev 0.50-7 Developer's toolkit for newt windo ii libpaperg 1.0.3-13 Library for handling paper charact hi libpopt-dev 1.4-1.1 lib for parsing cmdline parameters hi libwww-perl 5.47-1 WWW client/server library for Perl ii m4 1.4-10 a macro processing language hi make3.78.1-8 The GNU version of the "make" util ii makedev 2.3.1-44 Creates special device files in /d hi man-db 2.3.16-1 Display the on-line manual. ii perl-5.005 [perl5] 5.005.03-7.1 Larry Wall's Practical Extracting ii pointerize 0.4 Internationalization utilities, ba ii recode 3.5-1Character set conversion utility. hi rsync 2.3.2-1.2fast remote file copy program (lik hi slang1-pic 1.3.9-1 The S-Lang programming library, sh ii tetex-bin 1.0.7-2 teTeX binary files ii tetex-extra 1.0.2-2 extra teTeX library files ii zlib1g-dev 1:1.1.3-7compression library - development -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]