Re: d-i with Dell Dimension XPS M200s
Martin Michlmayr wrote: > * Jimen Ching <[EMAIL PROTECTED]> [2004-07-24 13:27]: > > I recently received an old Dell M200s tower desktop. I tried to install > > Debian on it with beta4. But the kernel failed to detect the IDE > > controller, so I don't have any drives at the drive partition stage. > > > > According to the Dell website, the M200s uses the SMSC FDC37C93xFR IO > > controller. I did a search of the 2.4.23 kernel source tree and only > > found references to this chip in IrDA and the Alpha directory. I see no > > references of this chip in the IDE driver directory. But a search on > > google returns mailing list entries that imply people have installed > > According to http://lists.debian.org/debian-cd/2002/04/msg00339.html > John H. Robinson has such a machine, so let's CC him. No, the system belonged to a friend of mine. I will pass on the inquiry. -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: debian-installer - devfs
Bastian Blank wrote: > On Fri, Sep 27, 2002 at 10:27:17AM -0700, John H. Robinson, IV wrote: > > > > mingetty allocated the first terminal, no problem. any second and > > subsequent totaly failed. > > do you have devfsd installed and/or yes > use /dev/ttyX instead of /dev/console/X? i made no tweaks to the devfsd config files; i have no idea what mingetty/agetty are using. i think i'll take a look at that later this day, and get back. i have found even more interesting results (with lvm, actually) when you mount defvs onto /devices and leave /dev as a ``normal'' filesystem -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: debian-installer - devfs
Bastian Blank wrote: > On Fri, Sep 27, 2002 at 10:18:46AM -0700, John H. Robinson, IV wrote: > > there are a couple of applications that i am aware of that do not get > > along with devfs very well. (wmppp and mingetty being two of them). > > mingetty works fine with devfs, while getty wont. other software can be > fixed. i had just the opposite experience: mingetty allocated the first terminal, no problem. any second and subsequent totaly failed. agetty was able to allocate two or more terminals. -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: debian-installer - devfs
Bastian Blank wrote: > > devfs is marked as stable within 2.4.x, also there are problems with > not devfs-capable drivers (i don't find any the since several months). there are a couple of applications that i am aware of that do not get along with devfs very well. (wmppp and mingetty being two of them). this is not so important on the debian-installer, as it is on the /target system, where we may or may not be mounting devfs by default. -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: New boot-floppies for 3.0r1?
David Kimdon wrote: > > > > yes, it sounds as though we definately want an updated 2.4.18 i was forwarded the actual securityfocus links that described the vulnerabilities. the kernel-source-2.4.18 that we have currently fails to address these issues specifically. http://online.securityfocus.com/bid/4259 http://online.securityfocus.com/bid/5178 http://online.securityfocus.com/bid/5539 neither was debian specifically mentioned as being vulnerable (probably because at the time, we did not ship a default installable 2.4.y kernel). a 2.4.19-bf2.4 kernel was uploaded to stable, but it is stalled in queue/new suggestions? -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: New boot-floppies for 3.0r1?
Adam Di Carlo wrote: > > I think we're waiting for the problem to be fixed by the kernel > maintainer. We should use a patched 2.4.18 since it seems unlikely to > me that the archive maintainers will allow a new kernel (2.4.19) into > stable point releases. actually, i was asked to produce a 2.4.19bf kernel, which i have. if we need updated 2.4.18, i'll provide that. expect 2.4.18bf u/l by friday -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Woody Installation
Axel Schlicht wrote: > Karsten Merker wrote: > > > It is described in the installation manual, which is on your CD set > > and also available online at: > > http://www.debian.org/releases/stable/i386/ch-rescue-boot.en.html#s-install-cd ^^ ^^ chapterlangauge(english) > > Didn't look into it; thought ch stands for Chinese, and with no Chinese > fonts istalled, why bother to look into it? no worries :) -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#149278: cfdisk partition type selection and installation
On Fri, Jun 07, 2002 at 12:44:34AM -0500, Wilbur Killebrew wrote: > Package: installation > Version: 2.11n (Debian 3.0 "woody") > > 1. The partition type entry field gets filled with seemingly > random, usually inappropriate values. i'm not quite sure what you mean here, can you give an example of the ``random'' data, and what you would have expected to see? > 2. It is not possible to set a partition to Type 85, Linux > Extended (presumably ext2), linux native (0x83) is what you would use for an ext2 partition. linux extended is a portion of drive that would have other, logical, partitions placed inside. in practice, i have never seen this used, as the ``normal'' extended (0x05?) is used. > Sorry I'm not using the Debian bug reporting system, but I don't > have Debain Linux fully configured yet. I'm doing this on a SuSE > 7.1 box. no worries :) thanks for your help in improving debian gnu/linux! -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: grub splashimage support?
On Tue, May 21, 2002 at 09:43:58PM -0400, Justin Whitney wrote: > Do any debian packages support splashimage= support for grub? i did find some splashimage= patches: http://people.redhat.com/~katzj/grub/patches/grub-0.91-vga16.patch and http://people.redhat.com/~katzj/grub/patches/grub-0.91-splashimagehelp.patch i was able to apt-get source grub, wget the two patches, cd to the grub-0.91 dir, apply the two patches, and debuild w/o error. -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Best way to use base-config_1.33.18?
On Fri, May 10, 2002 at 09:26:09AM +0100, J M Cerqueira Esteves wrote: > * John H. Robinson, IV <[EMAIL PROTECTED]> [2002-05-09 22:47 +]: > > On Thu, May 09, 2002 at 07:34:30PM -0300, Jim Skea wrote: > > > > > > I'm posting this to debian-boot since it seems where most of the > > > discussion on the base-config problem is gong on. > > > > my question, for those with base-config 1.33.17, and no easy way to > > upgrade it, what is the workaround? > > What I've been doing is: > [snip of easy way of upgrading base-config] i'm more worried about the debian newbies that have or will get woody pre-release ISO's with the broken base-config. telling them ``download the new base-config, and toss it on a floppy'' is certianly one solution (and the one i would follow, even though my newest system does not _have_ a floppy drive . . .) i'm looking for the workaround for the person locked up in their bunker, cut off from civilisation, with only the (flawed) woody cdroms. is there such a one? -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Best way to use base-config_1.33.18?
On Thu, May 09, 2002 at 07:34:30PM -0300, Jim Skea wrote: > > I'm posting this to debian-boot since it seems where most of the > discussion on the base-config problem is gong on. my question, for those with base-config 1.33.17, and no easy way to upgrade it, what is the workaround? -john -- 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: > > We have made a test image (thanks go to Chris Lawrence) : > http://www.phy.olemiss.edu/debian-cd/ this image failed on a Dell Dimension XPS M200s however, i suspect hardware problems since after copying the floppy images onto floppies, reading from the CD was still spotty at best. we resorted to a net install, which worked flawlessly. every other system attempted (including a handfull of various laptops) it worked great. -john NB: i am not subscribed to debian-cd -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: screenshots from BFs?
On Sun, Apr 21, 2002 at 07:54:52PM -0400, Matt Zimmerman wrote: > 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. you can also run dbootstrap from a chroot, or from your main filesystem. you will miss the opening screens, and the language selector. and you will probably be limited to english as a language. but it is there, and it does work :) -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Please test this woody cd image
On Fri, Apr 12, 2002 at 11:35:31AM +1000, Drew Parsons wrote: > On Thu, Apr 11, 2002 at 01:38:14PM -0600, Erik Andersen wrote: > > On Thu Apr 11, 2002 at 08:31:55PM +1000, Anthony Towns wrote: > > > Seriously: everyone reading this mail, burn a copy of Raphael's test image > > > on a CD and try booting it in any computers you have handy. > ... > > I just tested it on all the bootable x86 systems in my house: > ... > > Toshiba 490CDT Satellite Pro Laptop: Works > > That's good news. What was your success booting the Potato CD on this > laptop? i booted it on my toshiba 330cds laptop,and it worked fine. i test booted into bf24 and idepci. and yes, this model worked with the 2.88 el torito images, too :) -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: 2.4 kernel as default boot kernel on CD #1 ??
On Fri, Apr 05, 2002 at 04:50:50PM -0800, David Kimdon wrote: > > cd1: idepci; cd2 scsi; cd3: bf2.4; cd4: vanilla > > Any seconds? seconded. -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: makiing boot-floppies in a clean chroot
On Sat, Mar 02, 2002 at 09:09:57AM +0900, Junichi Uekawa wrote: > > > > sid + whiptail + libnewt-utf8-0 libnewt-utf8-dev libnewt-utf8-pic > > is not happy. since whiptail wants libnewt0, which conflicts with the > > rest. > > whiptail-utf8 is the one required to be installed, not whiptail. i noticed whiptail-utf8 after posting, but some issues: cthulhu:/# apt-cache show modconf | grep ^Dep Depends: whiptail, modutils (>= 2.1.85-14) cthulhu:/# apt-cache show whiptail-utf8 | grep ^Prov Provides: whiptail-provider so by my reckoning, either modconf needs to depend upon whiptail-provider (possibly in addition to whiptail), or whiptail-utf8 needs to provide whiptail i can submitt bugs, but i'd prefer to know that i am not total off base here i had previously --forced in whiptail, and things build. i'm now thinking that it was a mistake, and am trying to undo that, but modconf is really not letting me. downgrading whiptail to 0.50.17-9.2, then installing whiptail-utf8 0.50.17-9.4 caused an error when the latter tried to overwrite /usr/share/man/man1/whiptail.1.gz HEY! when did modconf no longer become required? removing that solved all those problems... i'm strongly reconsidering make a brand new chroot. -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
makiing boot-floppies in a clean chroot
i feel dense :/ i have boot-floppies in a clean woody and sid chroot. they both pass make check with no problems. but for make, they both fail at the same place: i18n_low_space=true ./rootdisk.sh "" /archive/debian/download 3700 3.0 "" C "" I: internationalized mode enabled ... snip of uninteresting stuff ... I: determining set of required libraries E: the following required libraries weren't extracted: /lib/libslang.so.1 E: ./rootdisk.sh abort looking through the list of stuff in scripts/rootdisk/ it seems to pick up slang1 (which has /lib/libslang.so.1) in EXTRACT_LIST_non-i18n since this is an internationalised mode, i assume that it is picking up on EXTRACT_LIST_i18n:slang1a-utf8 EXTRACT_LIST_i386_non-i18n-mixed-build:slang1-utf8 and i guess that neither one of those has the /lib/libslang.so.1 symlink. how to fix, and am i the only one that is seeing these problems? and if so, _why_ am i the only one? -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: boot-floppies 3.0.20
On Fri, Mar 01, 2002 at 01:49:55AM +0900, Junichi Uekawa wrote: > > I've got an impression that make clean doesn't seem to clean up the > downloaded deb files ? no, but make distclean does -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Bug#134045: slang/newt/utf8 troubles
On Thu, Feb 21, 2002 at 11:03:19PM -0800, John H. Robinson, IV wrote: > > oh, i think i see it: comparing dpkg --get-selections in each chroot, i > see sid has this: slang1a-utf8 > > let me see if this is the case... no; it is not quite that simple. installing libnewt-utf8-0 removes libnewt-dev and libnewt0. the libnewt.so symlink has to be re-made by hand, adn b-f won't compile anyway (probably due to libnewt-dev not being installed) installing libnewt-dev and libnewt0 (which removes libnewt-utf8-0) prevents compiliation also. but i was able to get it to be quiet about the slang1-utf8 stuff (hollow victory, it turns out) so we have boot-floppies that are not compiling anywhere > -john 100% post consumer .sig --- /usr/local/src/b-f-woody/tmp/woody-sel Thu Feb 21 23:00:04 2002 +++ /usr/local/src/b-f-sid/tmp/sid-sel Thu Feb 21 23:54:29 2002 @@ -40,0 +41 @@ +emacsen-common install @@ -91 +92 @@ -libnewt-utf8-0 install +libnewt0 install @@ -165 +165,0 @@ -slang1-utf8install @@ -167,0 +168 @@ +slang1a-utf8 install
Re: Bug#134045: slang/newt/utf8 troubles
On Wed, Feb 20, 2002 at 11:34:05AM +0100, Eduard Bloch wrote: > David Kimdon wrote on Tue Feb 19, 2002 um 10:48:22PM: > > Can someone summarize what needs to be done to make boot-floppies > > buildable again? > > With the split of libslang, we need following programs which link against the > .UTF-8 version. > > - whiptail-utf8 Ask dancerj, he forked the newt package which links against >utf8-slang. > - nano-tiny-utf8 > - cfdisk-utf8 is there any work on this being done? in a clean sid chroot, with only build-essential, and build-dep boot-floppies, and trying to make check in current cvs sources, i cannot make it happy. checking variable dependancies didn't find slang1-utf8 didn't find libnewt-utf8-0 # apt-get install slang1-utf8 libnewt-utf8-0 ... Package slang1-utf8 has no available version, but exists in the database. ... E: Package slang1-utf8 has no installation candidate in a clean woody chroot, i was able to get make check to be happy, but a make produced: E: Couldn't download slang1-utf8 E: ./rootdisk.sh abort oh, i think i see it: comparing dpkg --get-selections in each chroot, i see sid has this: slang1a-utf8 let me see if this is the case... -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
3.0.19 bf2.4 install success
Architecture: i386 (Pentium II (Deschutes) 400.913) Disk: IDE (WDC WD400BB-00AUA1, ATA DISK drive) Video: nVidia Corporation Vanta [NV6] (rev 21) NIC: Intel Corp. 82557 [Ethernet Pro 100] (rev 5) using bf2.4 flavour, it installed nicely. using the framebuffer. the one complaint, when formatting the partition (yes, i made it one huge partition! - i'll resize it later) the scrolling of the bterm was exceptionally slow. i think it completed the actual format by the time it had printed the list of ``sueprblocks'' this was a very simple install, configured keyboard, installed the drivers from floppy, added the eepro100 driver, got the base system from net. this is the way an install _should_ go. once i have this system set up, i can do some more strange installs, and fire test the installer some more. (i am moving the amd k6-2/300 to another location, and replacing it with the p2/400. i prefer amd, but i did not make this purchase, so i get to live with it) i saw a message on d-d about a debian boot logo? any chance of that? -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: initrd extends beyond end of memory
On Thu, Jan 31, 2002 at 03:31:38PM -0300, [EMAIL PROTECTED] wrote: > > > You could try adding "ramdisk=8000" to that. > > Tried. Nothing. did you run memtest86 http://www.teresaudio.com/memtest86/ to prove that your memory really is good? -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: [i386] install report: 2.4.17-bf kernel (long)
On Mon, Jan 28, 2002 at 04:09:45PM +0100, Eduard Bloch wrote: > > > - Formatting the partitions: the screen is filled with details about the > > progress of the ReiserFS-Journal being written. This is confusing for some > > Well, on slow harddisks you would like to have more info about the > problem. > > To Jaqque: what's your opinion? i had added a -q flag, which supressed progress as it happened. this sped it up significantly for those with slow (read: serial) terminals. without patching mkreiserfs, there is no way to make it quieter. unless we were to redirect all output to /dev/null, but i am not certain that is the right way to go. i do not believe that the mkreiserfs information is significantly out of characteristic with the other (ext2/ext3) mkfs programs. -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
new libc6
ome congratulations may be in order, i am not sure i updated sid again this afternoon (around midnight GMT), and rebuilt the bootfloppies. it failed: I: single locale mode forced for flavor '', no FB console support I: ld.so on this architecture is ld-linux.so.2 E: expected to find /lib/libc-2.2.4.so; upgrade local libc to 2.2.4 E: ./rootdisk.sh abort make[1]: *** [root.bin] Error 255 make[1]: Leaving directory `/usr/src/boot-floppies/woody/boot-floppies' make: *** [build] Error 2 zsh: exit 2 make due to the new libc6 (2.2.5-1). the congratulations that may be in order is the ``b-f no longer builds against sid'' of course, i may be totaly out to lunch on that one. my question: are we supposed to build against libc-2.2.4 or -2.2.5? changelog indicates that -2.2.5 fixes the glob problem, so i guess that answers my question. i'd fix the sources, but it is _way_ past my bedtime ATM (past midnight local time :) -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Things we need from sid
On Fri, Jan 25, 2002 at 07:25:13AM +1100, Herbert Xu wrote: > > > > > > IMHO reiserfs simply isn't viable anymore. So I don't really care please explain - how is it ``not viable anymore''? replaced by ext3? the only thing ext3 has is journalling. reiserfs brings a lot more to the table than simply journalling. > > 2.4 uses reiserfs-3.6 as default, 2.2.20 does not (and apparently will > > never) support it. Imagine someone downloads the Debian stuff to a > > ReiserFS partition under SuSE/RedHat (where reiserfs-3.6 has been used > > for epochs) and tries to mount it. > > Please recall that we're discussing boot floppies here so existing file > systems do not matter. not true. the basedebs.tgz could have been downloaded by the SuSE alt-boot reiserfs partition. /usr/local could be shared between the Debian and Red Hat boots, which could easily be reiserfs. this is not a non-issue, but a real and valid one. the latter can be worked around by upgrading the kernel to 2.4.x after the install, then adding the /usr/local to fstab. the former cold only be worked around by putting the basedebs.tgz on a non-reiser partition. the same issue faced by those using xfs. -john disclaimer: i happily run 2.2.20. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: [i386] introducing a kernel 2.4 installation flavor
On Tue, Jan 15, 2002 at 01:02:12AM +0100, Eduard Bloch wrote: > Hello Adam, hello everyone. > > As promised on IRC, I have worked on a useable solution for using the > 2.4 kernels in our current boot floppies and I think I have one. provided we can meet the following criteria, i have no objections to the use of 2.4.x, and dropping of 2.2.x: 1) (mandatory) we are still able to use ext3/resierfs/xfs(if available) as the root partition. this would either mean building the install kernel with ext3/resierfs/xfs(if available) statictly in the kernel (unlikely) or include them as modules int he RAM disk (possible, but might take up more room on the root disk that we don't really have). 2) (optional, but nice) we are able to support a 2.2.x install by the end user dropping in a 2.2.x kernel by following the instructions for using a custom kernel basically, i don't want to see us lose our options to support multiple filesystems on the root partition. it would be great if one flavour could do all supported filesystems, and it looks like 2.4.x will be able to do that. -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Testing boot floppies 3.0.19-2001-12-30
On Sat, Jan 12, 2002 at 07:16:14PM -0500, Jeff Sheinberg wrote: > > I switched to the syslog VC to see what was happening here, and > the problem was that the mount command executed was, > > mount -t ext3 /dev/hda6 /instmnt > > which can never succeed without the proper kernel support. > I suggest that boot-floppies should try ext2 after ext3 fails, or > just try only ext2 for kernels that do not support ext3. i'm thinking that a ``mount -t auto'' may be in order. the only time this would fail, where ``mount -t '' would work is if the FS type is a module that is not currently loaded. comments? -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: 3.0.19 install issues
On Tue, Jan 08, 2002 at 02:32:41AM -0500, Daniel Jacobowitz wrote: > On Mon, Jan 07, 2002 at 02:46:21PM -0800, John H. Robinson, IV wrote: > > On Mon, Jan 07, 2002 at 04:07:35PM -0500, Daniel Jacobowitz wrote: > > > I just ran through a clean Woody installation on 3.0.19; here's my > > > observations from it. Hope these are useful. Please CC me on any comments; > > > I no longer read debian-boot. > > > > > > - Choosing "mounted" defaults to /instmnt but is actually RELATIVE TO > > > "/instmnt/". This results in a lot of confusion and some ugly doubled > > > slashes. > > > > you gotta be kidding me :( > > i thought i had fixed all that :( :( > > > > were you using fresh-from-the-cvs sources? > > I was using the set Adam built; did you fix it after 3.0.19 was > uploaded? i'm not sure exactly what you did, but the disks i built last night from fresh CVS sources worked the way they were supposed to. for ``hard disk : partition on the hard disk'' there is a note that says ``Note that the filesystem you just mounted is located in the directory, /instmnt, but you don't need to give that directory here.'' that may have been confusing. unless you were using 3.0.18 images, which would exhibit the old, broken behaviour. the one way to tell for sure: if you have mounted multiple partitions, and select ``mounted: already mounted filesystem'' and you are presented with a menu of all your currently mounted partitions then you have the new choose_medium.c behaviour. if you feel the documentation does not match the behaviour, please, i implore you, to try to come up with better wording. i am hesitant to because 1) i can't think of a better way and 2) i don't want to create more burden on the translators without a real benefit. -john i ``Initialize a Linux Partition'' /dev/hda1, and mounted it as /. i then ``Mount a Previously-Initialized Partition'' /dev/hda7 and mounted it on ``Other: Enter name of a different mount point.'' /try i then dropped to a shell, and manually mount /dev/hda6 /instmnt and mkdir /this_is_dumb; mount /dev/hda2 /th* so i got a menu that looked a bit like this: lqu Select Partition tqk x x x Please choose the partition where the Debian archive x x resides. x x x x lqqqk x x x /dev/hda6: /instmnt x x x x /dev/hda1: /targetx x x x /dev/hda2: /this_is_dumb x x x x /dev/hda7: /target/tryx x x mqqqj x x x x x mqqj -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: 3.0.19 install issues
On Mon, Jan 07, 2002 at 04:07:35PM -0500, Daniel Jacobowitz wrote: > I just ran through a clean Woody installation on 3.0.19; here's my > observations from it. Hope these are useful. Please CC me on any comments; > I no longer read debian-boot. > > - Choosing "mounted" defaults to /instmnt but is actually RELATIVE TO > "/instmnt/". This results in a lot of confusion and some ugly doubled > slashes. you gotta be kidding me :( i thought i had fixed all that :( :( were you using fresh-from-the-cvs sources? -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: problems making own boot-floppies for MegaRAID
On Wed, Jan 02, 2002 at 02:13:58PM +, Rory Campbell-Lange wrote: > > 2) I managed to make a 1.2M kernel which fitted on the floppy disk. > However the syslinux loader reports a 'corrupt kernel'. I noticed that > when I tried to cp /usr/src/source*2.2.19/vmlinux to /mnt/linux the > machine wouldn't let me - I had to rm /mnt/linux and then copy */vmlinux > to /mnt/linux. look in /usr/src/source*2.2.19/arch/i386/boot/ for your new kernel. it will be named bzImage -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: i'm stuck now.
On Mon, Dec 31, 2001 at 07:01:36PM -0700, Chris Tillman wrote: > > export USE_LANGUAGE_CHOOSER := false > > in config (since your stuff is not involved with either of those). nope. did not work, but with the help of dancerj on irc, we were able to determine that it was python-xml that was the culprit, see previous message. -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: i'm stuck now.
On Mon, Dec 31, 2001 at 04:54:32PM -0800, John H. Robinson, IV wrote: > On Mon, Dec 31, 2001 at 02:02:19AM -0800, John H. Robinson, IV wrote: > > > > ./ver2.py 1 i386 langs.xml > > processing "langs.xml", utf=1, arch=i386 > > opening langs.c > > Traceback (most recent call last): > > File "./ver2.py", line 138, in ? > > dumper (arch) (utf == '0', outfile, result) > > File "./dumper.py", line 101, in dump_c > > outfile.write ('"%s",\n' % e.convert (lang.name)) > > UnicodeError: ASCII encoding error: ordinal not in range(128) > > make: *** [langs.c] Error 1 > > i am suspecting it a a python problem. i was close, it is a python-xml problem. sid now has python-xml 0.7-1, but that does not work., downgrading to python-xml 0.6.6-7 fixed the problem. since i do not know enough about utf-8, or i18n, or anything else of that nature, i am very ill-equiped to write a meaningful bugreport. so i will not. though i might be inclined to add a build-conflict of python-xml >= 0.7 i'm hoping that someone much smarter than me can fix this, so such a drastic step need not be taken. -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: i'm stuck now.
On Mon, Dec 31, 2001 at 07:01:36PM -0700, Chris Tillman wrote: > > This being a UTF error, maybe it can be avoided by either > > make build not from the top level CVS directory. that is what i have been trying :/ i want to get a complete set of disks, and use the root.bin as a chroot. if that works well, actually boot to the images. > export USE_LANGUAGE_CHOOSER := false this might work. i will try this. -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: i'm stuck now.
On Mon, Dec 31, 2001 at 02:02:19AM -0800, John H. Robinson, IV wrote: > > ./ver2.py 1 i386 langs.xml > processing "langs.xml", utf=1, arch=i386 > opening langs.c > Traceback (most recent call last): > File "./ver2.py", line 138, in ? > dumper (arch) (utf == '0', outfile, result) > File "./dumper.py", line 101, in dump_c > outfile.write ('"%s",\n' % e.convert (lang.name)) > UnicodeError: ASCII encoding error: ordinal not in range(128) > make: *** [langs.c] Error 1 i am suspecting it a a python problem. i know nothing of python. i even replaced the ``./ver2.py $(LANG_GEN_FLAG) $(architecture) langs.xml'' with ``./pl.py i386 $<'' in utilities/dbootstrap/langs/Makefile to no avail (same type of error) ./pl.py i386 langs.xml Traceback (most recent call last): File "./pl.py", line 510, in ? app.dump (outfile) File "./pl.py", line 425, in dump output.write ('\t%s' % item.definition ()) File "./pl.py", line 188, in definition return '{ "%s", NULL, NULL, NULL, NULL, NULL }' % from_utf8 (self.charset, self.name) File "./pl.py", line 34, in from_utf8 return convert ('UTF-8', charset, what) UnicodeError: ASCII encoding error: ordinal not in range(128) i included the version of python* i have installed, (sid, updated Mon Dec 31 09:50:46 2001 UTC) ii python 2.1.1-7An interactive object-oriented scripting lan ii python-dev 2.1.1-7Header files and a static library for Python ii python-newt0.50.17-9 A newt module for Python. ii python-tk 2.1.1-7Tkinter - Writing Tk applications with Pytho ii python-xml 0.7-1 XML tools for Python [dummy package] ii python2-base 2.0.1-1An interactive object-oriented scripting lan ii python2.1 2.1.1-7An interactive object-oriented scripting lan ii python2.1-dev 2.1.1-7Header files and a static library for Python ii python2.1-tk 2.1.1-7Tkinter - Writing Tk applications with Pytho ii python2.1-xml 0.7-1 XML tools for Python (2.1.x) ii python2.1-xmlb 2.1.1-7XML support included in Python (v2.1) -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: i'm stuck now.
On Mon, Dec 31, 2001 at 01:19:39PM -0700, Chris Tillman wrote: > > > > i have no idea what to do about this, or where to look for it, or > > anything. so i am stuck. if anyone has any clues, please let me know. > > Perhaps you didn't do a > > make check did that, everything was good. > before building? Also be sure to cvs update... did that, too. even before posting :) -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
i'm stuck now.
i feel very good about my /instmnt patches. but i do wish to test them as best as possible first, so i did a make clean, and tried to rebuild. utilities/dbootstrap/lang is causing me problems: ./ver2.py 1 i386 langs.xml processing "langs.xml", utf=1, arch=i386 opening langs.c Traceback (most recent call last): File "./ver2.py", line 138, in ? dumper (arch) (utf == '0', outfile, result) File "./dumper.py", line 101, in dump_c outfile.write ('"%s",\n' % e.convert (lang.name)) UnicodeError: ASCII encoding error: ordinal not in range(128) make: *** [langs.c] Error 1 i have no idea what to do about this, or where to look for it, or anything. so i am stuck. if anyone has any clues, please let me know. i'm going to bed :) happy new year, everyone -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: problems making own boot-floppies for MegaRAID
On Sun, Dec 30, 2001 at 02:09:34AM +, Rory Campbell-Lange wrote: > > Problems: > 1) My kernel is 1.5M, larger than a floppy! Won't this be a problem? of you try to put in on a floppy, yes. > 2) Can I somehow mount the boot disk image without putting it on a > floppy and mounting that, to do the kernel image swap? (I'm trying > to do as much of this as possible away from the office, sshing in!). YES! you can do it. i can only speak for the GRUB bootloader, not LILO. if you use LILO, translate to suite your purpose. in your menu.lst, add something like the following: title Debian Bootfloppy test /boot/debian/linux root (hd0,1) kernel /boot/debian/linux vga=normal load_ramdisk=1 ramdisk_size=16384 disksize=1.44 initrd /boot/debian/root.bin /boot/debian/linux is your 1.5M kernel, and /boot/debian/root.bin is the root.bin image of the flavour you want to use. some flavours have more capabilities, like ext3 or reiserfs. make sure if you do that, you append a ``flavour='' to the kernel line. boot, and enjoy! > 3) Although I've installed the boot-floppies package (and deps) on my > potato server, I can't find the promised docs. look in the ``documentation'' directory. plety to go around. -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: choose_medium.c and the /instmnt issue
On Thu, Dec 27, 2001 at 05:08:35PM -0800, John H. Robinson, IV wrote: > > it seems that i have resolved, either through accident or design, most > of the /instmnt problems. > > the only issue left is the Live CD. and, of course, after putting the code down for a night, i came up with some possibly better ideas: 1) if someone mounts a partition on /instmnt, they should be able to access it via the menu. the ``mounted'' option did that. but that seems to conflict with the mindset that mounted = part of /target. we don't want to take that away, which my patches currently would. 2) the cdrom method will mount a cdrom, no worries about that. we can safely ignore it. 3) LiveCD overloads (or seems to) the use of /instmnt. if true, this was probably a hack in order to maximise utilage of existing methods. now the issues are: a) partitions mounted under /target, b) LiveCD, c) partitions hand mounted under /instmnt (via a shell) d) mounting a previously unmounted partition my proposal: a) add a new method, ``target'' that specifically looks under /target b) look to see if we are a LiveCD, add a method ``LiveCD'' or whatever, that looks under some #define'd variable, so we can change it at will to meet the whims of the LiveCD people. (with some trickery, and if the LiveCD people say they want this, we can override the compiled in default via some bootargs variables) c) have ``mounted'' only available if there is a partition mounted on /instmnt d) same as it ever was so b) and c) would be new, on demand, methods, similar in nature to the ``cdrom'' method. this seems to leave everything as robust as possible, while making things seem a bit more natural. problems: the ``hard disk'' method first looks to see if anything is mounted on /instmnt, and immediately umounts it. however, the select_not_mounted() never sees this newly umounted partition. if i go back to the main menu, then mack into ``hard disk'' (via install kernel and modules) then it will see the umounted partition. there is some other function that is resetting dbootstrap's idea of what partitions are mounted or not. if you know, please tell me which one it is, so i can see if i can insert it cleanly between the umount() and select_not_mounted() calls. otherwise, i will have to look on my own - and i hate searching for needles in haystacks! -john erratta: in discussing this with adam on irc (btw, my /nick is jaqque), he said that he did not like the ``target'' method idea. he thinks that having people wander about the real / (the RAM disk) is acceptable. justification: power users (the assumed majority of people installing debian) will be spawning a shell, and may mount something anywhere on the RAM disk. first-time installers will be using a CDROM, so would not be touching ``mounted'' anyway. my counter-argument: unless you go outside the installer (such as spawning a shell), the only places something will get mounted is /target or /insmnt. i don't want the ``mounted'' method to overload to mean /target or /instmnt, and a power user will be able to escape the /target or /instmnt jail by using ``/..'' as a directory. this is ugly, but it is functional, and i don't care about making things beautiful for people that are doing things they should not be, anyway! to add even more, i was given a boxes.c patch, that allows to escape the /intmnt or /target jail. no trickery. this should not confuse newbies too much, and gives power users the chance to wander all about the RAM disk willy nilly. (the drawback? it gives newbies the power to easily escape the /target or /instmnt, either by accident or design) thoughts? (i'm not going to think about this for a couple of hours, so i can let all this coalesce somewhat) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: choose_medium.c and the /instmnt issue
the diffs so far: Index: choose_medium.c === RCS file: /cvs/debian-boot/boot-floppies/utilities/dbootstrap/choose_medium.c,v retrieving revision 1.123 diff -u -r1.123 choose_medium.c --- choose_medium.c 2001/12/27 15:25:52 1.123 +++ choose_medium.c 2001/12/28 01:10:25 @@ -11,7 +11,7 @@ #include "../libfdisk/fdisk.h" /* for fdisk_partitions Move this code to another file??? */ #define MAX_IDE_FLOPPIES 4 -#define CM_MOUNTPOINT_DIR "/instmnt/" +#define CM_MOUNTPOINT_DIR "/instmnt" #ifdef _TESTING_ #define is_network_configured() 1 #define configure_network() 1 @@ -253,6 +253,8 @@ /* we've boot off a CD-ROM, live style */ strcpy(buffer2, ""); } else { + i++; /* if we have to manualy choose a dir, give us a warning + when we are wrong, even if it is our first time through */ status = enterDirBox(_("Choose Directory"), text, prefix, preventry, buffer2, PATH_MAX); if (status == DLG_CANCEL) @@ -263,7 +265,8 @@ free(preventry); preventry = strdup(buffer2); - snprintf(buffer, sizeof(buffer), "%s%s", prefix, buffer2); + snprintf(buffer, sizeof(buffer), "%s%s%s", prefix, + buffer2[0]=='/'?"":"/", buffer2); if (NAME_ISDIR(buffer, &statbuf)) break; else { @@ -527,7 +530,7 @@ break; case ARC_manually: snprintf(prtbuf, sizeof(prtbuf), -_("Please enter the name of the directory that contains the Archive files.\nThe installation medium is mounted below %s/ ."), +_("Please enter the name of the directory that contains the +Archive files."), mountpoint); ; status = enterDirBox (_("Enter the Archive directory"), prtbuf, mountpoint, "", buffer, PATH_MAX); @@ -959,11 +962,8 @@ CM_MOUNTPOINT_DIR); status = choose_archive_dir(prtbuf, CM_MOUNTPOINT_DIR); -if (status && status != DLG_CANCEL) -{ - if (! system("cat /proc/mounts | grep -q " CM_MOUNTPOINT_DIR)) - umount(CM_MOUNTPOINT_DIR); -} +if (! system("cat /proc/mounts | grep -q " CM_MOUNTPOINT_DIR)) + umount(CM_MOUNTPOINT_DIR); return status; } @@ -1321,9 +1321,9 @@ case MED_mounted: if (disqtype == problem_report ) status = choose_archive_dir( - _("Please choose the directory where you would like to save the problem report."), CM_MOUNTPOINT_DIR); + _("Please choose the directory where you would like to save the +problem report."), TARGET_PREFIX); else - status = choose_archive_dir( _("Please choose the directory where the Debian archive resides."), CM_MOUNTPOINT_DIR); + status = choose_archive_dir( _("Please choose the directory where the Debian +archive resides."), TARGET_PREFIX); if (status == DLG_CANCEL) return DLG_CANCEL; break; -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
choose_medium.c and the /instmnt issue
it seems that i have resolved, either through accident or design, most of the /instmnt problems. the only issue left is the Live CD. a normal CD boot is when it uses the Rock Ridge to boot to a 2.88 image. the Live CD is something beyond weird, that i cannot describe. to say the least, based upon my delvings into the code of choose_medium.c and util.c, it seems to have a /dists directory. either it will loopback mount the basedebs to /instmnt (unlikely) or it will have already populated /instmnt directory structure (more likely) i don't want to break the Live CD (if it exists) and i don't want the Live CD's assumptions to break the user expectation's and assumptions. i'd like to discuss with the Live CD maintainer exactly what the Live CD is, what it's assumtions are about its filesystem, and what we can do to make /instmnt and /target transparent to the user. as this is currently a debian-boot issue, (and i am not subscribed to debian-cd) i'd prefer if the discussion stayed on debian-boot. the synopsys: (the way it is) hard disk = unused partition on a drive, selected, and mounted onto /instmnt. if [cancel] is used, partition remains mounted mounted = unused partition on a drive that is already mounted onto /instmnt the way it _should_ be (and where Live CD breaks it) hard disk = unused partition on a drive, selected, and mounted onto /instmnt. unmounted at end, no matter what. mounted= the filesystem mounted under /target, which includes ``Initialize a Linux Partition'' and ``Mount a Previously-Initialized Partition'' mounted partitions Live CD breaks this, by assuming that mounted means /instmnt, which is counter-intuitive to how the users and the menuing system is set up (it is possible to select ``mounted'' even though there is _nothing_ mounted on /instmnt!) this affects: Install Kernel and Driver Modules Install the Base System Report a Problem -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#126589: boot-floppies: Install Base System from already mounted filesytem produces "Supplied Directory Does Not Exist" error.
On Thu, Dec 27, 2001 at 01:12:02AM -0800, John Wenger wrote: > Package: boot-floppies > Version: N/A; reported 2001-12-27 > Severity: critical > Justification: breaks the whole system > > The version is 3.0.19, obtained from http://sbih.org/debian/bf3.0.19/. this is un-official set contains still experimental patches to fix the /instmnt and /target issue. the fix obviously needs fixing :) -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: choose_medium() and choose_archive_dir()
On Fri, Dec 21, 2001 at 12:45:24PM -0700, Chris Tillman wrote: > On Thu, Dec 20, 2001 at 10:58:33PM -0800, John H. Robinson, IV wrote: > > > > should we use /target, or / ? > > my initial thought was /target, since the only other mounted partitions > > would be there (such as, re-using a /var partition, but putting the > > drivers.tgz and rescue.bin in there first). the comments seem to imply / > > I think even /instmnt is really /target/instmnt. Since the installer > often operates in the /target chroot, /target is really the only safe > choice, ISTM. Or maybe it's actually already inside the chroot here? no, /insmnt is /insmnt. no symlink. since i am working in this part of the code, i will try to find the ``cancel does not umount /instmnt'' bit, and fix that too. but i am not going to lose (much!) sleep over it. -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: AppleScript debian floppy maker script
this is more appropriate for debian-devel or debian-mentor On Fri, Dec 21, 2001 at 01:02:48AM -0600, Elizabeth Barham wrote: > Chris Tillman <[EMAIL PROTECTED]> writes: > > A few months ago I put together a Debian floppy-maker script for usage > > on powerpc. The script depends on the MacOS-provided Applescript > > 'shell' and the scriptable Disk Copy program. The point was made > > before, that since these MacOS programs are not free (they come with > > all MacOS distributions), a script depending on them would not be > > Debian-distributable. > > If, for example, you wrote a GPL program for AIX that relied upon > their version of cc, I don't see how anyone could object: "IBM's > version of cc isn't GPL'd!" the issue is more of: though the floppy-maker script is GPL'd, and DFSG compliant, it _requires_(depends upon) something that fails the DFSG. thus, it would have to go into contrib, and cannot be a part of Debian proper. -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
choose_medium() and choose_archive_dir()
i'm working on getting rid of the /instmnt messages for choose_archive_dir(). i have successfully gotten rid of it, and only some cosmetic issues remain. i will submit my patches, even if i do not get the cosmetic issues entirely worked out. the usability issue i have found is this: the comments for choose_archive_dir indicate that if we use `mounted', we should be able to look through the root filesystem (the ramdisk, and all other mounted partitions, including /target) wherease if we chose ``harddisk'' then we should be limited to under /instmnt. choose_medium.c:209 /* text - the message displayed that tells the user what the directory will be used for prefix - the default prefix, either "/" for things on the root filesystem or "/instmnt/" for things that we mounted for the user */ static int choose_archive_dir(char *text, char *prefix) however, in choose_medium, this comment is ignored, and we are forced to use /instmnt anyway: choose_medium.c:14 #define CM_MOUNTPOINT_DIR "/instmnt/" choose_mdeium.c:1321 case MED_mounted: if (disqtype == problem_report ) status = choose_archive_dir( _("Please choose the directory where you would like to save the problem report."), CM_MOUNTPOINT_DIR); else status = choose_archive_dir( _("Please choose the directory where the Debian archive resides."), CM_MOUNTPOINT_DIR); if (status == DLG_CANCEL) return DLG_CANCEL; break; so my question: forcing /instmnt seems wrong, since we umount /instmnt at the end of choose_archive_dir (the exception seems if the user hit cancel). should we use /target, or / ? my initial thought was /target, since the only other mounted partitions would be there (such as, re-using a /var partition, but putting the drivers.tgz and rescue.bin in there first). the comments seem to imply / if i don't get any arguments, i am going to use /target -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Bug#123974: FWD: Bug#123974: post-boot network install fails because dhcp was not run
On Fri, Dec 21, 2001 at 12:09:24AM -0500, Joey Hess wrote: > John H. Robinson, IV wrote: > > last time i did a _full_ install of woody onto a laptop, baseconfig was > > started _before_ the pcmcia card services were started. so getting > > packages over the net failed. > > base-config is run from inittab, at the same time it runs the gettys. > This should not be before pcmcia or any other init scripts have run, but > after them. based upon my test-install of 3.0.18, this is correct. pcmcia is started before base-config. -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
3.0.18 - success!
i just did a test install of 3.0.18 architecture: i386 flavour: reiserfs console: serial there are still some problems with the mounting of non-ext2 partitions (i have a fix for that), and the /instmnt problem (i have a partial fix for that) another minor annoyance, more of a user issue, and a base-config issue, is this little gem: in base-config's intro: If you want to revisit this setup process at a later date, just run /usr/sbin/base-config. however, it marks itself for deletion, and will remove itself The following packages will be REMOVED: base-config console-data* console-tools* console-tools-libs* pcmcia-cs* and the user will be none the wiser (yes, it does say it in black and white, right there, but a complete newbie is not going to notice it) -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Booting my laptop
On Tue, Dec 18, 2001 at 10:18:20PM -0500, Colin Walters wrote: > > The way to generate basedebs.tgz is by using "debootstrap", like: > > cd /tmp > mkdir woody > debootstrap --download-only woody woody http://http.us.debian.org debootstrap --download-only woody woody http://http.us.debian.org/debian > cd woody > tar czvf basedebs.tgz var -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: 3.0.18 success (at last)
On Tue, Dec 18, 2001 at 03:17:26PM -0500, Adam Di Carlo wrote: > > Ok, good enough for me. I've retagged 3.0.18 boot-floppies. Uploading > source and i386 shortly (will take 2-3 hours). i have built the i386 from a tagged checkout, you can get the release directory from: http://sbih.org/debian/release/index.html suitable for making your own, personal mirror! -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#123974: FWD: Bug#123974: post-boot network install fails because dhcp was not run
On Tue, Dec 18, 2001 at 03:06:11PM -0500, Adam Di Carlo wrote: > "Tom Wzietek" <[EMAIL PROTECTED]> writes: > > > My point is: it should be started automatically. I'm referring to the > > first time logon after base installation, when base-config is run on > > the first console. > > My point is: It *is* started automatically. For everyone but you. My > job is to find how, why are you different? What went wrong? if i remember correctly, he was talking about the PCMCIA network. last time i did a _full_ install of woody onto a laptop, baseconfig was started _before_ the pcmcia card services were started. so getting packages over the net failed. after swapping to another VC, /etc/init.d/pcmcia start, everything worked great. i do not know if that is still the case, but might be worthy of looking into. -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
trying to get rid of /instmnt
in trying to get rid of the /instmnt from the mounting of a harddrive partition for installing of the base system, i was able to get rid of it from one dialog box, but not all. the problem is more subtle than i initially suspected. i am going to work on it some more later tonight or tommorow, but if someone has some time before then, and they can get it, please do! -john Index: choose_medium.c === RCS file: /cvs/debian-boot/boot-floppies/utilities/dbootstrap/choose_medium.c,v retrieving revision 1.118 diff -u -r1.118 choose_medium.c --- choose_medium.c 2001/11/03 15:28:01 1.118 +++ choose_medium.c 2001/12/16 20:31:41 @@ -241,10 +241,6 @@ /* keep user entry between calls to choose_archive_dir */ if (!preventry) { - /* CM_MOUNTPOINT_DIR is default if it's mounted */ - if (!system("cat /proc/mounts | grep -q " CM_MOUNTPOINT_DIR)) - preventry = strdup(CM_MOUNTPOINT_DIR); - else preventry = strdup(""); }
Re: 3.0.18 status confirmation
On Sun, Dec 16, 2001 at 12:00:33AM -0800, John H. Robinson, IV wrote: > On Sat, Dec 15, 2001 at 03:27:41PM -0700, Chris Tillman wrote: > > are showing. Here's what they say: > > > > init started: Busybox v. 0.60.3-pre (2001.12.07-03:53+) > > Bummer, could not run '/etc/init.d/rcS': Permission denied > > grep: /proc/cmdline: No such file or directory > > ah, damn. my fault. i'll fix if not already fixed. false alarm, not really my fault. but i removed the ``grep: /proc/cmdline: No such file or directory'' error message. -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: 3.0.18 status confirmation
On Sat, Dec 15, 2001 at 03:27:41PM -0700, Chris Tillman wrote: > are showing. Here's what they say: > > init started: Busybox v. 0.60.3-pre (2001.12.07-03:53+) > Bummer, could not run '/etc/init.d/rcS': Permission denied > grep: /proc/cmdline: No such file or directory ah, damn. my fault. i'll fix if not already fixed. -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Bug#123661: boot-floppies: 3.0.18(cvs) serial console unusable on i18n enabled flavours (i386)
On Fri, Dec 14, 2001 at 01:11:41AM -0500, Adam Di Carlo wrote: > "John H . Robinson IV" <[EMAIL PROTECTED]> writes: > > > it is caused by the spawning of bterm in the /sbin/udbootstrap script. > > So we need some way to bypass bterm when booting with a serial console > right? Is there a way to do it in shell script ? Any clues? fixed in cvs, changelog has note about this bug closure. -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: reiserfs notes
On Fri, Dec 14, 2001 at 01:37:08AM -0500, Adam Di Carlo wrote: > "John H. Robinson, IV" <[EMAIL PROTECTED]> writes: > > > 1) no mkfs.reiserfs in the root.bin > > scripts/rootdisk/SMALL_BASE_LIST_i386_reiserfs lists 'sbin/mkreiserfs'. > Isn't that the right thing? If not, please check in a fix. the scripts/rootdisk/SMALL_BASE_LIST_i386_reiserfs is correct. > > 2) Install Kernel and Driver Modules -> > > harddisk : partition on the hard disk -> > > /dev/hda6: Linux native (happens to be reiserfs) -> > > Mount Failed (The partition was not mounted successfully.) > > I would suggest booting with 'debug' option, trying the actual 'mount' > arguments as reported in the system log when you use debug. the problem is interesting: the code looks at the partition number (82) and then reverse maps that to ``ext2'' and mounts with that -t option. since ext3, ext2, xfs, and resierfs all share that same code, i wrote a little routine that uses the magic number code from partition_config.c, added it to util.c, and used that if the partition is ``ext2''. it works for reiserfs, anyway. * code not checked in yet, want to test with ext2 > > 3) ftp.us.debian.org seems to be out of date, the modules installed were > >2.2.19-reiserfs > > Huh, that's wierd. Seems like a build problem -- try 'make distclean' > before building. will do. -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
default directory in choose_archive_dir()
code snippet: utilities/dbootstrap/choose_medium.c 216 static int choose_archive_dir(char *text, char *prefix) 217 { ...snip... 243 if (!preventry) { 244 /* CM_MOUNTPOINT_DIR is default if it's mounted */ 245 if (!system("cat /proc/mounts | grep -q " CM_MOUNTPOINT_DIR)) 246 preventry = strdup(CM_MOUNTPOINT_DIR); 247 else 248 preventry = strdup(""); 249 } this is completely useless when you are installing from a harrdrive partition. worse yet, it is misleading. the ``/instmnt'' is placed on the selection, with a note that says that it is not even required. very few hard drive partitions would have such a directory, and when you simply hit return, it comes up with with the same screen, but the entry removed. worse yet, if you hit tab to the <...> an error, and feeds you right back to the dialog box, with the same (wrong!) default. i'm right now leaning towards removing that check altogether, and giving the user a blank entry line however, there might be a legitimate reason for that default _to_ be there (say when installing from cdrom, or nfs, or whathave you) i simply do not know. if no one can say ``yes! this is important!'' i will remove it. -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
insufficient build deps
this is yet another note, but i am not sure how to fix it. this is an atypical build process, so this is why i can find it % make clean % make resc2880reiserfs.bin ... snip! ... I: including font for bterm cp: cannot stat `utilities/bogl/unifont-reduced.bgf': No such file or directory E: ./rootdisk.sh abort make: *** [rootreiserfs.bin] Error 1 a ``make -C utilities/bogl unifont-reduced.bgf'' fixes the problem. not sure whose depedency it really is, so i am unsure where to actually add that bit. in other words: i cannot fix it at this time, hopefully someone else can. priority: really really low -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#123661: boot-floppies: 3.0.18(cvs) serial console unusable on i18n enabled flavours (i386)
some additional information: it is caused by the spawning of bterm in the /sbin/udbootstrap script. i changed the script to: #!/bin/sh if (/dev/null && ! /usr/bin/tty | grep -q ttyS ; then export LC_CTYPE=C@utf-8 exec /usr/bin/bterm -f /unifont-reduced.bgf /sbin/dbootstrap fi exec /sbin/dbootstrap based upon the model in /etc/init.d/check_fb.sh. it worked in a chroot, but failed when booted. when i commented out the entire if .. fi clause, it worked fine. -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#123661: boot-floppies: 3.0.18(cvs) serial console unusable on i18n enabled flavours (i386)
Package: boot-floppies Version: N/A; reported 2001-12-12 Severity: important when the kernel is passed the console=ttyS0 option, the kernel boot messages are properly sent to the console. however, the language choose shows up on the first virtual console, and the keyboard is non-responsive. the /boot/debian/kernel and /boot/debian/root.bin were taken from resc2880reiserfs.bin from /boot/grub/menu.lst: title Debian Bootfloppy test (serial console) /boot/debian/linux kernel (hd0,1)/boot/debian/linux console=ttyS0,115200n8 vga=normal load_ramdisk=1 ramdisk_size=16384 disksize=1.44 flavor=reiserfs initrd (hd0,1)/boot/debian/root.bin -- System Information Debian Release: 3.0 Architecture: i386 Kernel: Linux osiris 2.2.19 #1 Thu May 3 21:53:55 PDT 2001 i586 Locale: LANG=C, LC_CTYPE=C -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: cvs commit to boot-floppies by jaqque
On Tue, Dec 11, 2001 at 04:59:53PM -0800, Debian Boot CVS Master wrote: > > fixed the EXTACT_LIST and SMALL_BASE_LIST to support varying > flavours within an architecture > > required for reiserfs and udma100-ext3 as a secondary effect, rootreiserfs.bin and rootudma100-ext3.bin are now both too large to fit on a 1.44M floppy. i can make rootreiserfs.bin fit by removing reiserfsck and resize_reiserfs, but this severly limits the use of the rescue disk -rw-r--r--1 root src 1475798 Dec 11 18:09 rootreiserfs.bin (that is with resize_reiserfs, but reiserfsck missing) a mere 1238 over quota. does anyone have any ideas what we can do, to at least save resize_reiserfs? i want to save that one, because i can imagine someone using the rescue disk to resize their root partition, instead of simply booting to single user mode. if we can't do it, we can't do it. the flavour will still work. reiserfsck is simply way too large (182172 bytes) to try and save, and reiserfs journals, so no need to fsck anyway, right? ;) -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
reiserfs notes
i am doing this mostly for documentation. i should probably file bugreports, all issues listed affect the reiserfs flavour: 1) no mkfs.reiserfs in the root.bin unable to create a reiserfs partition (this is very defeatest of the whole flavour) (could be an artifact of my abreiviated build process) 2) Install Kernel and Driver Modules -> harddisk : partition on the hard disk -> /dev/hda6: Linux native (happens to be reiserfs) -> Mount Failed (The partition was not mounted successfully.) i assume it tried to mount as ext2, because when i drop to a shell, i am able to mount it succesfully: # mount /dev/hda6 /mnt will probably have to try this again, on a ``real'' console to find the problem, serial console's error reporting is a bit sparse. 3) ftp.us.debian.org seems to be out of date, the modules installed were 2.2.19-reiserfs -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: for 2.2.20/i386: reiserfs pcmcia modules needed, udma100-ext3 has vga16 fb again. (was: Re: fallback to single language when fbcon not available)
On Fri, Dec 07, 2001 at 04:35:39PM +0100, Eduard Bloch wrote: > > BTW, the recently uploaded version of kernel-image-2.2.20-udma100-ext3 > has vga16fb again. If you switch to the new kernel version, you can > enable it in rootdisk.sh. AFAICS the pcmcia-modules-2.2.20-reiserfs > packages is the only thing that prevents us from switching. i just uploaded a kernel-2.2.20-reiserfs with working pcmcia modules, and they are in unstable now. the new -resierfs kernel supports the frame buffer's requested in Bug#122475. -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Flashing screen with cvs bootdisks
On Sat, Dec 01, 2001 at 10:59:29PM +, Philip Blundell wrote: > In message <[EMAIL PROTECTED]>, "John H. Robinson, IV" writes: > >i'm testing the installer now. should it have asked about language? > > No, the reiserfs flavour isn't i18n enabled. The only ones that will > give you a choice of language are compact, idepci and udma100-ext3. ah, very good. that will require an actual on-site reboot. i'll try to test monday. will it be international over serial? is it supposed to be? -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Flashing screen with cvs bootdisks
On Sat, Dec 01, 2001 at 11:14:54AM +, Philip Blundell wrote: > In message <[EMAIL PROTECTED]>, "John H. Robinson, IV" writes: > >there is an etc/messages.trm. i did do a make clean before building. > > > >this was the reiserfs flavour, over serial. > > Right, yeah, I think I see the problem. > > Can you try this patch to dbootstrap and see if it helps? did the trick. i'm testing the installer now. should it have asked about language? environment: chrooted on /mnt, proc mounted on /mnt/proc and a loopback mount on /mnt/target, with /etc/resolv.conf copied to /mnt/etc/resolv.conf -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Flashing screen with cvs bootdisks
On Fri, Nov 30, 2001 at 04:46:15PM -0800, John H. Robinson, IV wrote: > > i am rebuilding, after a make distclean (not sure what that is going to > buy me, other than a lot more network usage, as .deb's get downloaded) same error. i even tried the root.bin version. i verified that /etc/messages.trm exists. (note: after you gunzip the image, and mount it, make sure you mount proc also. esp. if you are on serial console / ssh. otherwise i am assuming that all display is sent to the first VC. i ended up having to kill the linuxrc process, which rebooted the whole system. whups!) any idea where to start looking for the problem? i'm not very i18n aware. -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Flashing screen with cvs bootdisks
On Sat, Dec 01, 2001 at 12:05:30AM +, Philip Blundell wrote: > In message <[EMAIL PROTECTED]>, "John H. Robinson, IV" writes: > > Problem > > > > An error occured while loading application messages. > > > >that is further than i got before. > > If you haven't done "make clean" recently, try that and then rebuild. > Otherwise, you may need to do a bit of debugging. What .trm files do > you have on the root disk? there is an etc/messages.trm. i did do a make clean before building. this was the reiserfs flavour, over serial. i am rebuilding, after a make distclean (not sure what that is going to buy me, other than a lot more network usage, as .deb's get downloaded) -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Flashing screen with cvs bootdisks
On Fri, Nov 30, 2001 at 02:55:35PM -0800, John H. Robinson, IV wrote: > On Fri, Nov 30, 2001 at 10:35:03PM +, Philip Blundell wrote: > > > > I think it's the -pic packages that you have to watch out for particularly. > > Make sure that libnewt-utf8-pic is installed, and libnewt-pic is not. Having > > both of them is probably bad news. > > libnewt-pic install > libnewt-utf8-picinstall > > whups. i'll fix that, and try to rebuild,and let you know how it goes. Problem An error occured while loading application messages. that is further than i got before. -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Flashing screen with cvs bootdisks
On Fri, Nov 30, 2001 at 10:35:03PM +, Philip Blundell wrote: > > I think it's the -pic packages that you have to watch out for particularly. > Make sure that libnewt-utf8-pic is installed, and libnewt-pic is not. Having > both of them is probably bad news. libnewt-pic install libnewt-utf8-picinstall whups. i'll fix that, and try to rebuild,and let you know how it goes. -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Flashing screen with cvs bootdisks
On Fri, Nov 30, 2001 at 07:41:51PM +, Philip Blundell wrote: > > Are you running on a machine with fbcon enabled? (If you see a penguin on the >screen during bootup, you have fbcon.) > > Do the vanilla or reiserfs flavours work any better? i tried with reiserfs, and i get the flashing also. i386, reiserfs flavour, console and serial. -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: last gasp to achieve b-f i18n (was Re: LC build failing)
On Sat, Nov 17, 2001 at 02:00:08AM -0500, Adam Di Carlo wrote: > > i186 has 1027 available on root.bin 1440 vanilla. Other flavors seem > similar, although reiserfs is signifcantly smaller. no framebuffer support (had to take it out, in order to get the reiserfs bit in monolithically) -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: [PATCH] set sticky bit when creating /var/tmp mount-point
On Tue, Nov 13, 2001 at 03:06:38PM -0900, Ethan Benson wrote: > > your missing the point: /tmp and /var/tmp are being treated differently. with no good reason. thus i submit: fix /tmp, or fix /var/tmp, so they both act the same way. consitency is the path to least surprise. -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Capturing install screenshots
On Fri, Nov 02, 2001 at 05:47:42PM -0500, Carpenter, Dean wrote: > > I need to produce a screen-by-screen document detailing the woody/testing > install process for internal use here. i found that running dbootstrap in an xterm is sufficient. you may want to do it in a chroot, and take othre reasonable precautions. the only parts you miss are the boot: screen, and the inital message from adam di carlo. -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: cvs commit to boot-floppies/debian by dwhedon
On Fri, Sep 07, 2001 at 04:06:29PM -0700, Debian Boot CVS Master wrote: > > Make tftpboot images for vanilla, compact, ide and idepci (do we want them all, > reiser too?). it makes my testing easier, in the lab i had set up, yes -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: build failure on ia32
On Tue, Sep 04, 2001 at 05:47:48PM -0600, Erik Andersen wrote: > > [andersen@slag gcc-2.95-2.95.4.ds5]$ dpkg -l reiserfsprogs | tail -n1 > ii reiserfsprogs 3.x.0j-6 User-level tools for ReiserFS filesystems > [andersen@slag gcc-2.95-2.95.4.ds5]$ du -hc /sbin/mkreiserfs /sbin/reiserfsck >/sbin/resize_reiserfs > 92k /sbin/mkreiserfs > 184k/sbin/reiserfsck > 96k /sbin/resize_reiserfs > 372ktotal i am not avers to reiserfsck going away, since it is 1) about useless and 2) usually not required. (perhaps a way to make a good rescue disk with these utilities? hmm, must think about that) resize_reiserfsck? it can probably go away also. not a great need for that at install time. (at rescue time? again, probably not) not going to help the pps busybox bloat though. -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: can't detect NIC
On Fri, Aug 31, 2001 at 12:04:04PM -0400, Susan G. Kleinmann wrote: > It looks like this problem is related to pcmcia-cs and not to the boot-floppies > work at all. i've seen that, too. i'd recommend installing the 3.1.22potato pcmcia-cs, and if that works, hold the package there echo pcmcia-cs hold| dpkg --set-selectons -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: can't detect NIC
On Wed, Aug 29, 2001 at 09:52:52PM -0400, Susan G. Kleinmann wrote: > > Susan G. Kleinmann wrote: > > > I'm tyring to install the i386 disks from 'testing' on a Toshiba 1805-S203, > > > to which I've added a Linksys "Combo" ethernet card. > I guess I'm a little confused here. Do you mean that I should find a > very old version of pcmcia-source (3.1.22 rather than the current 3.1.28), > and then compile that (with some kernel or other)? this is what i did, since i knew that the 3.1.22 modules worked for my PCMCIA NIC: installed kernel-source-2.2.19and pcmcia-source into a woody chroot. unpacked them both i took the /boot/config-2.2.19-reiserfs, copied it into /usr/src/kernel-source-2.2.19/.config, cd'd into /usr/src/kernel-source-2.2.19, applies the reiserfs patch (you can skip that part), and did a make-kpkg modules i then installed that .deb, took the /lib/modules/2.2.19/pcmcia directory and tar'd it up, and cp'd the tarbal onto an MSDOS formatted floppy. once inside the installer, right after the install kernel and modules, i rm -rf;d /target/lib/modules/2.2.19, mounted the floppy with the pcmcia tar file, and untared it into /target/lib/modules/2.2.19 then ran a depmod -a i then continued on with Configure the PCMCIA modules. most of it was dependent upon 1) a working woody chroot and 2) i knew that the 3.1.22 modules worked. > but I was not able to mount root.bin or root1440.bin the same way: right. they are both gzipped. if you ungzipped them first, then you could have mounted them. > I gave cardmgr the option "-v", and then looked at the output of dmesg. > It looks as if the system is trying to put the card at IRQ 11. odd. what else was using IRQ 11? /proc/interrupts will tell you > How can I force cardmgr to pick up pcnet_cs instead of axnet_cs? 3.1.22 does not even have an axnet_cs > If success here is just a matter of forcing cardmgr to try irq 5, then > how do I do that? dunno, i am afraid. > At this point, I don't see a way to install Debian, short of using > the potato floppies. Any alternative ideas appreciated. woody bootfloppies are still experimental. in some cases (PCMCIA) there may be required additional steps to get them to work. this seems to be one of them. if you cannot rebuild the 3.1.22 PCMCIA modules for whatever reason, i would recommend using the potato disks, then upgrading to woody. best of luck! -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: can't detect NIC
On Tue, Aug 28, 2001 at 11:03:16AM -0400, Susan G. Kleinmann wrote: > I'm tyring to install the i386 disks from 'testing' on a Toshiba 1805-S203, > to which I've added a Linksys "Combo" ethernet card. you didn't mention which flavour (if any) you were using. > I guess that somehow the system doesn't know that eth0 should be > identified with the pcnet_cs driver. How can I tell it? i found something similar with my Xircom CreditCard Ethernet CE3B-100BTX card - some versions (3.1.28) of pcmcia-card-services detect it as a memory card. however, i found that compiling the 3.1.22 modules, and using it with the 3.1.28 cardmgr (as is found on bf-3.0.13 and greater) provides satisfactory results. to gain even more feedback as to what is going on, either start a shell, or go over to VC2 and start the shell there, and check the output of dmesg. -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: boot-floppie 3.0.13 success, of sorts
On Tue, Aug 28, 2001 at 12:09:22AM -0400, Colin Walters wrote: > "John H. Robinson, IV" <[EMAIL PROTECTED]> writes: > > > major frownies: (again, who to bug? netbase?) > > smtp is probably exim, which should be secure enough (although I > prefer Postfix personally). The others are from netkit-inetd. i knew smtp was exim, that was not the concern. i will bug netkit-inetd, thank you. -john who prefers qmail -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: boot-floppie 3.0.13 success, of sorts
On Mon, Aug 27, 2001 at 08:00:16PM -0700, John H. Robinson, IV wrote: > > and now i have the Congratulations, you have successfully installed > Debian! screen. *whew!* major frownies: (again, who to bug? netbase?) [jhriv@chao:~]% nmap cthulhu Starting nmap V. 2.53 by [EMAIL PROTECTED] ( www.insecure.org/nmap/ ) Interesting ports on cthulhu.ucsd.edu (132.239.50.6): (The 1519 ports scanned but not shown below are in state: closed) Port State Service 9/tcp opendiscard 13/tcp opendaytime 25/tcp opensmtp 37/tcp opentime Nmap run completed -- 1 IP address (1 host up) scanned in 1 second -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
boot-floppie 3.0.13 success, of sorts
Version: 3.0.13 Flavour: reiserfs Distribution: woody Hardware: Toshiba Satellite 330CDS The pcmcia modules are totaly hosed. the only way to make the pcmcia work is to compile them seperately, and after the modules have been installed, copy in the good modules, and re-run depmod -a (this is not strictly a boot floppy problem, however) making the user wait to acknowledge the empty exim.conf is plain mean. it serves no use. configuring LILO caused a bunch of spewage onto VC 1. hitting ^C in VC1 while it seemed that debootstrap has froze (it was http.us.debian.org being extremely slow) caused the installer to restart. at least the network settings were kept :) it seems that when i specified a transceiver type for the PCMCIA nic, it did not take, until i tried to reconfigure PCMCIA. then the network came up just fine as PCMCIA was restarted. and now i have the Congratulations, you have successfully installed Debian! screen. *whew!* -john MD5 passwords are not the default?! uh oh.pcmcia was not started, no net. Apt Configuration does not like that. a little /etc/init.d/pcmcia start luv fixed that right up. wishlist bug? against which package? -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: successful i386/reiserfs install
On Mon, Aug 27, 2001 at 12:54:34PM +0300, Riku Voipio wrote: > > 3. asked, if i wanted to mount root with -notail, althought >i had a ext2 /boot partition, this could be probably >autodected. the -notail option is useful for more than just /boot. it also speeds up access, at a cost of space. this is why it is good for partitions that will hold mail/news spools. -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: ext3 - support
On Fri, Aug 17, 2001 at 01:01:38AM -0800, Ethan Benson wrote: > > one question, do these patches patch debootstrap to include the > relevant filesystem's -progs package? eg xfsprogs, reiserfsprogs > etc. if not the first boot may fail miserably when the fsck wrapper > can't find /sbin/fsck.{xfs,reiserfs} (in xfs' case its a copy of > /bin/true, i don't know about reiser). for reiser, the fsck pass is zero, so it should not be fsck'd at all. -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: serial console install - boot-floppies woody 3.0
On Wed, Jul 25, 2001 at 11:29:54PM -0400, Matt Zimmerman wrote: > > "John H. Robinson, IV" <[EMAIL PROTECTED]> writes: > > > > > it looks like update_console_info() in > > > ./utilities/dbootstrap/baseconfig.c is the place to fix this. > > > > Hmm, are you suggesting the console args should be parsed from > > /proc/cmdline and then we frob around in inittab to set this? that's exactly what i did. if the speed is not set, then it defults to 9600, as both the kernel and grub do. > > Is there any more elegant way of doing it? possibly, but my reasoning works like this: 1) we parse the comand args anyway 2) if the kernel defaults to 9600, then it would _have_ to be told to use something else, so the cmdline is pretty definitive. > > Would it be possible to work through a patch or, if not, file a bug > > for it so we don't forget? no need; already fixed. > Unless, of course, TIOCGSERIAL works on /dev/console if it happens to be > connected to a serial device. I don't have a system to test. if this is more elegant, i have a system i can test. (i386, connected via telix/minicom) but as this is already fixed, with a minmum of code added, i don't see much harm. -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: ResierFS in boot-floppies (was: Debian "reiserfsprogs" and installation)
On Thu, Jul 19, 2001 at 03:53:18PM -0600, Ed Boraas wrote: > > Just out of curiosity, are we planning to ship 2.2.19 on woody's floppies, > or (already-reiserfs-enabled) 2.4.x? i386 will be 2.2.19 -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: ResierFS in boot-floppies (was: Debian "reiserfsprogs" and installation)
On Thu, Jul 19, 2001 at 08:10:17AM -0600, Ed Boraas wrote: > What are our plans for boot-floppies, wrt supporting a ReiserFS root > partition? At the very least, would it be feasible to include mkreiserfs on > the boot set? I know that there's some code in boot-floppies to ask about > formatting using ReiserFS, but I'm not sure if it's enabled. i have a reiserfs kernel sitting in incoming, waiting on the ftp-masters to okay it. i have that kernel in an apt-get'able archive at deb http://people.debian.org/~jaqque woody main contrib non-free or you can use wget http://people.debian.org/~jaqque/dists/woody/main/binary-i386/kernel-image-2.2.19-reiserfs_2.2.19-1_i386.deb all the reiserfs code is in the cvs tree, the only thing not checked in yet are the makefile type things to build the reiserfs images. i was waiting on the kernel to make it into unstable, but since i have it publicly available through other means, i may go ahead and check in the changes today. -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
serial console install - boot-floppies woody 3.0
doing a serial console install on i386 architecture, serial console, i noticed upon reboot that everything went smoothly until base-config completed, and the getty was spawned. T0:23:respawn:/sbin/getty -L ttyS0 9600 vt102 my console was using 115200 bps, not 9600. the ``console=ttyS0,115200n8'' was passed to the kernel at boot time, so the serial speed was in /proc/cmdline at install and reboot time. it looks like update_console_info() in ./utilities/dbootstrap/baseconfig.c is the place to fix this. -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: New Install (easydeb)
On Tue, Jul 17, 2001 at 04:47:09PM +1000, Robert Browne wrote: > Hi, > > Have a floppy image called easydeb.img, happy to download it to > debian so you can have look. Just tell me where. sounds interesting. do you have a link for the sources and image? i assume that the installer is licensed under the GPL. -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
boot-floppies 2.2.23
maybe i am missing something, but i built a stable chroot, apt-got boot-floppies, and am trying to build them. however, i am running into a chicken/egg problem. for i386, the kernel is 2.2.19pre17. however, there is no vanilla kernel, and if i bump it up to 2.2.19, there are no PCMCIA modules. i have no desire to build the vanilla and friends kernel/pcmcia modules. are they hiding somewhere that i am not aware of? or should i simply pull the source from the CVS server, and if so, which tag? potato? -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: reiserfs (was Re: merge-floppy help request)
On Fri, Jun 22, 2001 at 01:52:34PM -0400, Adam Di Carlo wrote: > > MaX in the FaX <[EMAIL PROTECTED]> writes: > > > > >the guy who builds the > > > reiser version doesn't provide that? Why not? > > > > I think because the goal, is to have only a floppy disk set, not a > > bootable cdrom. The problem is that the email address of the guy that > > has provided this set, is not avaible, so i cannot communicate directly > > whit them. [EMAIL PROTECTED] there are no CDROM images because i don't have a cdrom burner, so i have no way to test such an image. > > Do you know a hacked version of boot-floppies.xx.deb with support of > > reiserfs, during the install phase? if exist, where? > > No, sorry. http://chao.ucsd.edu/debian/boot-floppies/ > > If possible for 2.2.19 kernel version. > > Yes, I understand the boot-floppies could easily be changed to support > this in the mainstream. > > Does anyone know of any issues which would block our providing a > "resierfs" version? 1) reiserfs flavored kernel or 2) reiserfs patched kernel somehwere. if we used 2.4.x kernels which have the reiserfs ``patches'' already integrated, this would negate both of those. in the current CVS for woody, there is support for reiserfs. however, the resierfs eabled code looks to see if flavor=reiserfs is passes as a boot argument. this can be readily changed, if we somehow make another flavor with reiserfs patches already integrated. -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#99652: dialog boxes confusing - active choice not clear
On Sat, Jun 02, 2001 at 09:48:12PM -0400, Joey Hess wrote: > > Screenshot: http://kitenet.net/~joey/tmp/whiptail.png (expires in 2 > weeks or so) but how does that look on a b/w term? -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#99652: dialog boxes confusing - active choice not clear
On Sat, Jun 02, 2001 at 12:28:50PM -0700, David Whedon wrote: > > An easy option which may or may not be any better is to not use the 'compact' > version of buttons for yes/no boxes. An example of this can be found here: much nicer! that looks like it would work well on black and white terms too (so long as they support reverse video) -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
reiserfs+woody
reiserfs eneabled installer for woody is still a possibility. since it looks like that 2.2.19 (or a variant thereof) shall be the default kernel for at least one reiserfs supported architecture, it will be ultimately up to the kernel maintainer as to which route to go: 1) yet-another-flavor 2) patch the default kernel i can go either way. they both have their merits, and their drawbacks. i tend towards the yet-another-flavour because i frown upon a patched kernel. i believe the base kernel should be what you get from kernel.org however, i am easy and am very willing to #2 work, if all others affected do not mind. i was kind of hoping to see the 2.4.x kernel be the default across the board, so this would be a moot point. -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: sponsor for QEmacs maybe
On Fri, Apr 27, 2001 at 10:24:50PM +0100, Edmund GRIMLEY EVANS wrote: > QEmacs is a tiny full-screen editor with UTF-8 support, a hex mode and > the ability to edit huge files efficiently: > > http://www-stud.enst.fr/~bellard/qemacs/ have you contacted the author? has he given his blessing to QEmacs being pacakged? i ask because i came across this: Download Current snapshot (please do not redistribute since it is not finished): qemacs-0.2.tar.gz -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
kernel for woody
we have a 2.2.x kernel for about half the architectures, and a 2.4.x kernel for the rest. are we going to have a 2.4.x kernel for all our architectures when the woody floppies are released? what's the likelyhood of supporting grabbing the ramdisk over nfs? (thinking specifically of i386 architecture - this may be too niche of a market to widely support, and i'll accept that argument) -john alpha 2.2.19pre13 arm 2.4.3 hppa 2.4.0 i386 2.2.19 ia64 2.4.0 m68k 2.2.17 mips 2.4.2 powerpc 2.2.19 powerpc (apus) 2.2.10 powerpc (pcmcia) 2.2.19-pmac sparc 2.2.19 -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Question regarding /etc/apt/sources.list and boot-floppies
On Wed, Apr 18, 2001 at 03:42:16PM -0400, James D Strandboge wrote: > > Currently I have it for woody, and just grab what I need from sid and put it in my > $(ftp_archive)/local directory, which seems to work ok. that is exactly what i have done, and is probably the safer way to go. otherwise, all the packages will come from unstable vice testing (where there is a different one in each) -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: debian 2.3
On Fri, Apr 13, 2001 at 05:28:22PM -0700, Sam Powers wrote: > On Fri, Apr 13, 2001 at 08:20:12PM -0400, Adam Di Carlo wrote: > > "Dwayne C. Litzenberger" <[EMAIL PROTECTED]> writes: > > > > If so, will the installation program support reiserfs, a parted > > > > frontend, will it support reiserfs as a default install method, > > > > Well, absolutely not the default, but it's a possible option. If we > > go with grub, I think that doesn't get along with ReiserFS. Is it > > necessary to use kernel 2.4.x if you want ReiserFS? > > GRUB groks reiserfs.. but I think a reiserfs root should be very > discouraged. It's only like 100mb for me usually, it doesn't take long > to fsck at all. ReiserFS has the speed where it counts. i dunno. i use 100% reiserfs. but i guess i could be what is called ``daring'' as far as that goes, i can commit my ReiserFS additions to debootstrap at any time. since i cannot get a boot disk to build, i have been hesitant to actually make the commits. > Are ReiserFS folks keeping the 2.2 series up to date as well as 2.4? i dunno. i have not seen the reiserfs patch for the 2.2.19 kernel in Debian yet (it is on the namesys.com site though) > For what it's worth, on i386, GRUB should DEFINATLY be used over lilo. > It's like 20 times cooler. And all the trendy distros are doing it >;) -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: debian 2.3
On Fri, Apr 13, 2001 at 02:54:01AM -0400, Adam Di Carlo wrote: > [suggestion to add telnet/netcat to boot-floppies] > > I guess I would really shelf this for now... or just take suggestions > and leave them in todo for now. busybox 0.51 (others also?) includes netcat and telnet. i guess its a moot point now? -john now if we turn them on or not, is yet another question. busybox 0.51 with everything turned on is 284316 bytes. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: FWIW : vi in busybox announced
On Fri, Apr 13, 2001 at 04:51:16PM +0200, Thierry Laronde wrote: > FWIW, on the 10th of april has been released a version of Busybox including > a 22k vi. Perhaps, in the future, the end of the small editors wars ;) :) but how close to a ``real'' vi is it? /me goes to find out . . . -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
suggested method for building woody floppies?
thanks to the informative error messages, i have determined that the shared library problem was due to the fact that i was trying to build the boot-floppies on a woody system, while the boot disks were pulling down unstable .deb's. the libc versions were different, which was causing the problem. i cleaned the build tree, cleaned the archive cache, rem'd out the unstable deb URI iin the sources.list, and it failed to get console-data. so should we get the console-data from unstable, and put it in the local cache, or should we be building in an unstable chroot? -john -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
add -f -f to mkreiserfs
Package: reiserfsprogs Version: 3.x.0d-2 Severity: wishlist this is for the woody bootfloppies to be able to install a reiserfs partition. this patch will allow mkreiserfs to be called with multiple -f flags to prevent asking any questions. great for non-interactive use diff -Naur reiserfsprogs-3.0.20001019-old/mkreiserfs/mkreiserfs.c reiserfsprogs-3.0.20001019/mkreiserfs/mkreiserfs.c --- reiserfsprogs-3.0.20001019-old/mkreiserfs/mkreiserfs.c Sat Oct 14 01:33:52 2000 +++ reiserfsprogs-3.0.20001019/mkreiserfs/mkreiserfs.c Sun Mar 18 18:54:29 2001 @@ -371,7 +371,9 @@ { case 'f' : /* force if file is not a block device or fs is mounted. Confirm still required */ - force = 1; + /* force >1 is an unconditional force; no + confirmation required */ + force += 1; break; case 'h': @@ -414,9 +416,16 @@ printf ("mkreiserfs: '%s' contains a mounted file system\n", device_name); if (!force) exit (1); - printf ("Forced to continue, but please confirm (y/n)"); - if (getchar () != 'y') - exit (1); + switch (force) { + case 1: + printf ("Forced to continue, but please confirm (y/n)"); + if (getchar () != 'y') + exit (1); + break; + default: + printf ("Forced to continue.\n"); + break; + } } dev = open (device_name, O_RDWR); @@ -453,10 +462,13 @@ report (device_name); -printf ("ATTENTION: YOU SHOULD REBOOT AFTER FDISK!\n\tALL DATA WILL BE LOST ON '%s'! (y/n)", device_name); -c = getchar (); -if (c != 'y' && c != 'Y') - die ("mkreiserfs: Disk was not formatted"); +if (force < 2) +{ +printf ("ATTENTION: YOU SHOULD REBOOT AFTER FDISK!\n\tALL DATA WILL BE +LOST ON '%s'! (y/n)", device_name); +c = getchar (); +if (c != 'y' && c != 'Y') + die ("mkreiserfs: Disk was not formatted"); +} invalidate_other_formats (dev); write_super_and_root_blocks (); -- System Information Debian Release: testing/unstable Kernel Version: Linux osiris 2.2.18 [classified] Sat Jan 6 11:19:04 PST 2001 i586 unknown Versions of the packages reiserfsprogs depends on: ii libc6 2.2.2-1GNU C Library: Shared libraries and Timezone -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Impressions on using the debian potato installation Cds
On Fri, Mar 09, 2001 at 12:14:38PM +0800, Dan Jacobson wrote: > sure hope non subscribers can mail to this group, and hope one day to > be able to read it via nntp. also the name "boot" doesn't 100% match > the description "installation' ... the next generation installer, debian-installer, should more match your expectations. the name boot-floppies is mostly historical, and in a pinch the install disks can be used as an emergency boot disk. > why so many submenus with no way to back out, only one choice, > "" most of those are informational only. no questions being asked, no need to confuse with a ``yes or no'' > I use chrony on my other linux distribution. please ask the user if > he agrees to writing the software clock to the hardware clock before > doing this in the rc. scripts! the first time we notice it it is too > late. during the install, you are asked if the real time clock is set to localtime or GMT. it does not actually set the hardware clock. > could at least tell user '35/44 steps thru installation' or something > so we have some idea on screen. this is a good idea. i like it. > XF86setup: big disaster for me this is after the install, and into package installation. the installer just gives you a very minimal system (called ``base'') then hands you over to the real package management tools. > please when asking the user about 'initializing' disks, use the word > 'format' as that better implies the danger. for hard drives, formatting is a low-level procedure done usually at the factory. it is unfortunate that one company's misconception between formatting and creating a filesystem (``initializing'') should affect such a large number of people :( so although the term ``initializing'' may be confusing for some, it is an acurate term. there are plenty of warning that data will be destroyed. this is similarly true for floppies. but some tools will actually both format and lay a filesystem at the same time. the tools provided with debian do not, however. > why not when giving the list of partitions to format, mention which > ones have already been formatted. you mean initialized. you format a drive, and initialize a partition. any partition that you have already mounted will not be listed. any partition that is not of the proper type will not be listed. i hope that the next generation installer, debian-installer, can handle this a bit more gracefully. this is a good idea. > ALT F2 shell window of ramdisk: I set set -o emacs but still couldn't > use ^P the shell on the boot-floppies is ash, not bash. ash has no sense of history. i beleive the shell warns you of this when you start it. > module selecting: yucky agreed. however, this can usually be skipped (the documentation to this affect seems to be rather sketchy) > when debian gets larger, perhaps do the installation that doesn't need > answering questions in the background at the same time as that that > does? the next generation installer, debian-installer, should be much more accomodation to this. > many questions aren't clear about what will happen if one hits > RET only. the default value is highlighted. (but i will agree that trying to determine if the highlight is blue or red can be difficult) > v19 emacs the default?! how about a newer default emacs19, emacs20, and xemacs21 are all available in the potato (stable) distribution. > before writing any files on the hard disk, please ask about the time zone > or something. the installation leaves files that seem 8 hrs younger > than now [I'm a Taiwan localtime user ] time stamps are written in GMT. if you correctly answered the question about if the RTC is set to GMT or localtime, then the timestamps should be correct. > red/blue for default choice might not always work for perhaps > color blind on your yes no questions. a black/white screen is also available. one of the first questions you are asked is color or b/w. (default is color) > I'm actually glad that x windows isn't a requirement to run your > installer script. me too. > depending on a separate installation instructions document isn't > feasible well, unless you tell a user how to see it at the same > time on ALT F4 etc. windows. this is a space consideration. the entire installer and all support files have to be able to fit on a single 1.44M floppy. this make adding all the things we would like (such as online documentation) very difficult, especially when you consider translations for all text. the next generation insaller, debian-installer should address this better. > entering passwds: no echo *** chars: bad no * echos is very good, and a UNIX standard for years. so you really want to tell the shoulder surfer exactly how long your password is with a slight glance? > no message on how to redo screwed up parts of the install from the shell in the install itself (before the reboot) you can go back in the menu and correct almost any step you might have made a mistake on. -john PGP