Re: junk showing up in /emul/ia32-linux
Friends - My original post said: > I have a really nice amd64 sid machine, which has run > [...] > I have never run a 32-bit binary on this computer, and > don't see any reason to ... ever. If it's not built from > published, 64-bit clean source code, it doesn't belong on Bernd Petrovitsch wrote: > You apparently do not need OpenOffice.org. Nope. It's not 64-bit clean source code. Besides, I'm an old TeX-head, and TeX works just fine. Jo Shields wrote: > or lilo/grub That's a strange special case, since I guess the BIOS jumps to the boot sector code in 32-bit mode. I suppose in the long run I could "fix" that problem by running LinuxBIOS. How did grub manage to get built before the days of /emul/ia32-linux? And that need can hardly explain libg2c0-dev libgfortran0-dev. Unless someone rewrote grub in Fortran when I wasn't looking. ;-) Will multi-arch make the situation cleaner or dirtier for people like me who only want support for one arch variant? - Larry signature.asc Description: Digital signature
junk showing up in /emul/ia32-linux
Friends - I have a really nice amd64 sid machine, which has run solidly for over a year now. I really appreciate all the really hard work behind the scenes that keeps Debian unique and useful. I have never run a 32-bit binary on this computer, and don't see any reason to ... ever. If it's not built from published, 64-bit clean source code, it doesn't belong on my computer. Recently files have started showing up in /emul/ia32-linux. Offending packages are libg2c0-dev, fakeroot, and libgfortran0-dev. I tried submitting bug reports for the first two: #341786 and #341788. The latter got merged with the inverse bug (amd64 junk showing up on i386 machines), #323285, and tagged "minor" and "wontfix". Shall I rm -rf /emul after every apt-get upgrade? Shouldn't there be a way to separate out compatibility cruft, either at the package or installer level? - Larry signature.asc Description: Digital signature
Re: eeyes
On Mon, November 7, 2005 10:47 pm, Olivier Bornet said: > I often use eeyes > (http://packages.debian.org/oldstable/graphics/eeyes). > Is someone knowing if there is a amd64 package for it ? (or even if it > is replaced by another tool on the Debian repositories). Reply by Dean Hamstead: > check out gqview The intended Gnome standard is now "eog" (Eye of Gnome). I personally use "qiv" (Quick Image Viewer). Both are supported Debian packages. - Larry signature.asc Description: Digital signature
libgnomeui-common upgrade asks to remove gnumeric and more
# apt-get upgrade Reading package lists... Done Building dependency tree... Done The following packages have been kept back: libgnomeui-common 0 upgraded, 0 newly installed, 0 to remove and 1 not upgraded. recycle:~# apt-get install libgnomeui-common Reading package lists... Building dependency tree... The following packages will be REMOVED: gnumeric gnumeric-common libgnomedb2-4 libgnomedb2-common libgnomeui-0 libgoffice-1 libgucharmap4 The following packages will be upgraded: libgnomeui-common 1 upgraded, 0 newly installed, 7 to remove and 0 not upgraded. Need to get 606kB of archives. After unpacking 27.1MB disk space will be freed. Do you want to continue [Y/n]? Surely this is not intended behavior. - Larry signature.asc Description: Digital signature
libwxgtk2.4-python 2.4.2.6 install error
# apt-get upgrade Reading Package Lists... Building Dependency Tree... The following packages will be upgraded: libwxgtk2.4-python 1 upgraded, 0 newly installed, 0 to remove and 0 not upgraded. Need to get 0B/2592kB of archives. After unpacking 0B of additional disk space will be used. Do you want to continue? [Y/n] [chop] Reading changelogs... (Reading database ... 68563 files and directories currently installed.) Preparing to replace libwxgtk2.4-python 2.4.2.6 (using .../libwxgtk2.4-python_2.4.2.6.1_amd64.deb) ... Unpacking replacement libwxgtk2.4-python ... dpkg: error processing /var/cache/apt/archives/libwxgtk2.4-python_2.4.2.6.1_amd64.deb (--unpack): trying to overwrite `/usr/bin/helpviewer', which is also in package wxpython2.5.3 dpkg-deb: subprocess paste killed by signal (Broken pipe) Errors were encountered while processing: /var/cache/apt/archives/libwxgtk2.4-python_2.4.2.6.1_amd64.deb E: Sub-process /usr/bin/dpkg returned an error code (1) # echo $? 100 # signature.asc Description: Digital signature
Re: fujitsu-siemens primergy rx300 s2
Len Sorensen wrote - > Do [Broadcom] provide a driver that works without a firmware file? Do they > include sources to any firmware files required? If not they can call it > GPL all they want, but it still won't be GPL. They wouldn't be the only > company to make that mistake (For example Sangoma's wanpipe drivers > claim to be GPL but have vinary only modules and firmware files in > them). AFAICT, binary-only firmware is OK, if that blob is redistributable. That's why hooks to allow user-space firmware loading were added, it keeps those blobs out of the GPL-covered binary. Not everything on a Debian CD has to be GPL. The firmware file is "merely aggregated". This is the situation with my Prism54 wireless card. Supplying the firmware from the main CPU saves the manufacturer a few bucks and square cm of board space for the PROM, makes firmware upgrades less tricky. The only technical downside is the few milliseconds it takes to schlep the bits onto the card. The legal issues arise if the manufacturer fails to license the binary for redistribution, then Debian etc. can't put it on the general-issue CD, and the end-user has to cajole the file from the manufacturer before they can use the hardware. Binary-only modules for the main CPU are, of course, deadly. - Larry signature.asc Description: Digital signature
Re: Still confused about pure64 package changelogs
Thanks, everyone, for your comments! On Tue, Feb 01, 2005 at 06:56:36PM +0100, Kurt Roeckx wrote: > On Mon, Jan 31, 2005 at 04:04:14PM -0800, Larry Doolittle wrote: > > > > Today, for instance, I see netbase is ready to upgrade, from > > version 4.19 to version 4.20. The trouble is, looking at > > packages.debian.org, the changelog only goes up to 4.19. > > It seems to be there now? Maybe they're just slow in updating > it? I guess. Yes, I see it there now, too. My point is that there was a window where the update was available but the documentation was not. My first and best reason to run Linux (since 1992) is reliability. If there isn't a perfect match between the update and packages.debian.org, that isn't the primary source of information about updates, and I want to know what is. > > Another particular example is kernel-image-2.6.10-9-amd64-k8, > > which doesn't even show up on packages.debian.org. > > This is one of the excpetion of packages that are first uploaded > to the amd64 archive and only later to the debian archive. It's > stuck in NEW for some time now. Can anyone elaborate on this comment? Who controls and/or where is the master list of such packages, and how can we mere users find out what's going on with them? To the several people who pointed me to apt-listchanges: yes, that sure looks like what I want. I don't mind downloading new .deb files first, as long as I can find out what's in them before installing. It looks (from the man page) like apt-listchanges is designed to slide into apt-get somehow. Google found a few pages suggesting that connection is supposed to happen automagically in the install, but it didn't for me. Well, maybe it tried: right after it ran apt-listbugs (also new on my system [*], I ran into suggestions to run it as well as apt-listchanges), I do see the line Reading changelogs... Done but nothing came out. You can probably tell I haven't used Debian for very long: only about six months. I hope it's the last distribution I have to learn! - Larry [*] It sure would be nice if all these system tools were written in the same language. Pulling in apt-listbugs had the side effect of pulling in ruby and friends. Do we really need a copy of every scripting language on the planet just to administer a debian system? signature.asc Description: Digital signature
Still confused about pure64 package changelogs
Let me assure everyone before I start that I'm really happy with debain pure64. Fast and solid. Most of the time, when I see a package ready to download with apt-get upgrade, I can go to packages.debian.org, and find out what changed and why. Every now and then, that doesn't work. Today, for instance, I see netbase is ready to upgrade, from version 4.19 to version 4.20. The trouble is, looking at packages.debian.org, the changelog only goes up to 4.19. Now, netbase is kind of an important package, and I'd like a way to read about the changes before I load it up. Another particular example is kernel-image-2.6.10-9-amd64-k8, which doesn't even show up on packages.debian.org. In that case, I downloaded the source package, and buried in there I did indeed find a changelog that appeared up-to-date (although with a typo in it). Do these changelogs appear on the net anywhere, in a way such that I can avoid downloading the source every time? This isn't Gentoo! - Larry signature.asc Description: Digital signature
kernel package documentation
Guys - Can someone give me a pointer to documentation on the AMD64 kernel package builds on Alioth? For generic debian packages, I can get Changelogs from packages.debian.org. Two specific questions, that I'd like to learn how to answer myself: when I "apt-get upgrade"d today, including kernel-image-2.6.8-10-amd64-k8 and kernel-image-2.6-amd64-k8, the automagically generated /boot/grub/menu.lst changed my default kernel from 2.6.9-9 to 2.6.8-10. I don't know if this is intended or desired. Today's upgrade also gave me a new 2.6.9 kernel patch (2.6.9-4), but no corresponding kernel image. Is one on the way? Lastly, what are the plans regarding 2.6.10? I'd really like to start up SMART on my SATA drive. - Larry
contention for /usr/lib64 in libc6 and base-files
If I try to install (update) libc6, I get dpkg: error processing /var/cache/apt/archives/libc6_2.3.2.ds1-19.0.0.1.pure64_amd64.deb (--unpack): trying to overwrite `/usr/lib64', which is also in package base-files If I try to install (update) base-files, I get The following packages have unmet dependencies: base-files: PreDepends: libc6 (>= 2.3.2.ds1-19.0.0.1.pure64) but 2.3.2.ds1-19 is to be installed libc6-dev: Depends: libc6 (= 2.3.2.ds1-19.0.0.1.pure64) but 2.3.2.ds1-19 is to be installed E: Unmet dependencies. Try 'apt-get -f install' with no packages (or specify a solution). And of course, apt-get -f install tries to install libc6, so I'm back to square one. How do I recover from this? I understand from reading e.g., http://lists.debian.org/debian-amd64/2004/12/msg00318.html that some of these links are critical for the system's operation, so I don't want to charge ahead without some idea where I'm going. - Larry
Re: VFAT: Very stupid question
Hi - Using debian x86_64 2.6.8-9-amd64-k8, I think I see the same problem that greg reported in http://lists.debian.org/debian-amd64/2004/10/msg00315.html It gets stranger, though: 1. The same memory stick reads and writes fine on my x86 sarge laptop. 2. When I check to see if VFAT works, using the method described by Harald Dunkel in that thread, it does work. 3. fdisk /dev/sdb works fine, I see a 128Meg partition as sdb1 4. I get one more message out of the failed mount command when I run it on the console: FAT: invalid media value (0xb9) 5. After upgrading to 2.6.9-9-amd64-k8, I can no longer access the partition table on the memory stick. /var/log/messages cut-and paste appended. Test conditions: Kingston DataTraveler 2.0 Rev: 4.10 MSI K8T Neo2, MS-6702E AMD Athlon64 3500+ This motherboard has four USB controllers, each one shows up as VIA Technologies, Inc. VT82x UHCI USB 1.1 Controller This uses the memory stick with its out-of-the-box partition table. Don't tell me to put a new partition table on it, that approach would leave me in the dark when someone gives me their memory stick with useful files on it. I haven't recompiled a custom kernel for this box yet, but I will if someone suggests a useful experiment. - Larry P.S. I really like debian-amd64 on this computer. For an unstable distribution, it's really very clean. This is one of only a handful of warts. And the performance is phenomenal! ---start snip from /var/log/messages--- Nov 1 11:44:29 localhost kernel: usb 1-1: new high speed USB device using addre ss 5 Nov 1 11:44:29 localhost usb.agent[3533]: ub: already loaded Nov 1 11:44:29 localhost kernel: uba: device 5 capacity nsec 50 bsize 512 Nov 1 11:44:31 localhost kernel: uba: made changed Nov 1 11:44:33 localhost kernel: uba: device 5 capacity nsec 50 bsize 512 Nov 1 11:44:35 localhost kernel: uba: device 5 capacity nsec 50 bsize 512 Nov 1 11:44:35 localhost kernel: /dev/ub/a:end_request: I/O error, dev uba, se ctor 0 Nov 1 11:44:35 localhost kernel: end_request: I/O error, dev uba, sector 2 Nov 1 11:44:35 localhost kernel: end_request: I/O error, dev uba, sector 4 Nov 1 11:44:35 localhost kernel: end_request: I/O error, dev uba, sector 6 Nov 1 11:44:35 localhost kernel: end_request: I/O error, dev uba, sector 6 Nov 1 11:44:35 localhost kernel: end_request: I/O error, dev uba, sector 4 Nov 1 11:44:35 localhost kernel: end_request: I/O error, dev uba, sector 2 Nov 1 11:44:35 localhost kernel: end_request: I/O error, dev uba, sector 0 Nov 1 11:44:35 localhost kernel: unable to read partition table Nov 1 11:44:35 localhost kernel: /dev/ub/a:end_request: I/O error, dev uba, se ctor 2 Nov 1 11:44:35 localhost kernel: end_request: I/O error, dev uba, sector 4 Nov 1 11:44:35 localhost kernel: end_request: I/O error, dev uba, sector 6 Nov 1 11:44:35 localhost kernel: end_request: I/O error, dev uba, sector 0 Nov 1 11:44:35 localhost kernel: unable to read partition table ---end snip from /var/log/messages---