Re: [gentoo-user] glibc 2.13 warning
On Wed, 09 Feb 2011 00:03:36 -0500, Allan Gottlieb wrote: The postfix issue is separate and needs a glibc downgrade to fix. I am not sure I understand. I did the glibc upgrade and have not downgraded. I run postfix and my mail is coming and going. I had thought/hoped that postfix does work with the new glibc if you don't do prelinking (which is my configuration: 2-13 glibc, postfix, no prelink). If glibc-2.13 kills postfix w/o prelinking then perhaps, I am just lucky and my overall configuration is such that it works for me. I think you're lucky, as is Volker. A number of people, me included, have had Postfix fail with postfix/local[4452]: fatal: unable to determine open file limit postfix/master[4001]: warning: process /usr/lib/postfix/local pid 4452 exit status 1 postfix/master[4001]: warning: /usr/lib/postfix/local: bad command startup -- throttling Prelinking is not involved, so it must be down to how you have configured Postfix, probably the local delivery settings. signature.asc Description: PGP signature
Re: [gentoo-user] copy a bunch of files...
On Tue, Feb 8, 2011 at 8:27 PM, Mark Knecht markkne...@gmail.com wrote: Hi, Looking for a simple way to do a big copy at the command line. I have a bunch of files (maybe 100 right now, but it will grow) that I can find with locate and grep: c2stable ~ # locate Correlation | grep Builder | grep csv /home/mark/Builder/TF/TF.D-17M-2009_06-2010_11/Correlation/TF.D-17M-2009_06-2010_11-V1.csv /home/mark/Builder/TF/TF.D-17M-2009_06-2010_11/Correlation/TF.D-17M-2009_06-2010_11-V2.csv /home/mark/Builder/TF/TF.D-17M-2009_06-2010_11/Correlation/TF.D-17M-2009_06-2010_11-V3.csv SNIP /home/mark/Builder/TF/TF.D-31M-2009_06-2010_11/Correlation/TF.D-31M-2009_06-2010_11-V4.csv /home/mark/Builder/TF/TF.D-31M-2009_06-2010_11/Correlation/TF.D-31M-2009_06-2010_11-V5.csv c2stable ~ # I need to copy these files to a new directory (~mark/CorrelationTests) where I will modify what's in them before running correlation tests on the contents. How do I feed the output of the command above to cp at the command line to get this done? I've been playing with things like while read but I can't get it right. Thanks, Mark Another way to do it is with find: find /home/mark/Builder -type f -iname '*csv' -exec cp {} ~mark/CorrelationTests \; If catching the *csv is not enough, you can use -ipath instead of -iname and do something like this: -ipath '*Builder*csv' this will match all the path. Regards, Kfir
[gentoo-user] cups hplib - how to divorce
Hi, I think cups-1.4.6 and my HP printer via USB can't live in harmony. Often, it gets stuck. Then it tries to use feed 1 (manual feed) although I have configured cups to use feed 2. Can I just remove hlip from my system, delete and add my printer again in cups. Has anybody made similar experience? Many thanks for a hint, Helmut.
Re: [gentoo-user] copy a bunch of files...
On Wed, 9 Feb 2011 13:38:37 +0200, Kfir Lavi wrote: Another way to do it is with find: find /home/mark/Builder -type f -iname '*csv' -exec cp {} ~mark/CorrelationTests \; Replace \; with + for a faster process, as Mark said there are hundreds of these files. Or, if you use zsh instead of bash, it can be as simple as cp Builder/**/*.csv CorrelationTests -- Neil Bothwick Learn from your parents' mistakes - use birth control! signature.asc Description: PGP signature
Re: [gentoo-user] cups hplib - how to divorce
2011/2/9 Helmut Jarausch jarau...@igpm.rwth-aachen.de: Hi, I think cups-1.4.6 and my HP printer via USB can't live in harmony. Often, it gets stuck. Then it tries to use feed 1 (manual feed) although I have configured cups to use feed 2. Can I just remove hlip from my system, delete and add my printer again in cups. Has anybody made similar experience? Does it work with stable cups? Do you use in kernel USB printer support or libusb? -- Daniel Pielmeier
[gentoo-user] FIXED:Re: [drm] loading RV710 Microcode fails
Nikos Chantziaras realnc at arcor.de writes: I use ati-drivers-10.11 All the instructions you received assume that you were using xf86-video-ati, *not* ati-drivers (also known as AMD Catalyst for Linux, AMD's proprietary binary driver for Radeon cards). Do *not* install any firmware files and make sure you disable the radeon driver completely in the kernel. Not only KMS, but no driver *at all*. Lesson well learned. ati-drivers working fine now. snip # Generic Driver Options # CONFIG_UEVENT_HELPER_PATH=/sbin/hotplug # CONFIG_DEVTMPFS is not set # CONFIG_STANDALONE is not set # CONFIG_PREVENT_FIRMWARE_BUILD is not set CONFIG_FW_LOADER=y # CONFIG_FIRMWARE_IN_KERNEL is not set CONFIG_EXTRA_FIRMWARE= # CONFIG_DEBUG_DRIVER is not set # CONFIG_DEBUG_DEVRES is not set # CONFIG_SYS_HYPERVISOR is not set # CONFIG_CONNECTOR is not set # CONFIG_MTD is not set # CONFIG_PARPORT is not set CONFIG_PNP=y CONFIG_PNP_DEBUG_MESSAGES=y Look correct? thx, James
Re: [gentoo-user] cups hplib - how to divorce
On 02/09/2011 12:58:07 PM, Daniel Pielmeier wrote: 2011/2/9 Helmut Jarausch jarau...@igpm.rwth-aachen.de: Hi, I think cups-1.4.6 and my HP printer via USB can't live in harmony. Often, it gets stuck. Then it tries to use feed 1 (manual feed) although I have configured cups to use feed 2. Can I just remove hlip from my system, delete and add my printer again in cups. Has anybody made similar experience? Does it work with stable cups? Do you use in kernel USB printer support or libusb? AFAIR it worked just fine with CUPS-1.3.x I do have CONFIG_USB_PRINTER=y in my kernel configuration. Thanks, Daniel, Helmut.
Re: [gentoo-user] copy a bunch of files...
On 02/09/2011 12:56:09 PM, Neil Bothwick wrote: On Wed, 9 Feb 2011 13:38:37 +0200, Kfir Lavi wrote: Another way to do it is with find: find /home/mark/Builder -type f -iname '*csv' -exec cp {} ~mark/CorrelationTests \; Replace \; with + for a faster process, as Mark said there are hundreds of these files. Or, if you use zsh instead of bash, it can be as simple as cp Builder/**/*.csv CorrelationTests There is a problem with this approach, though. It can easily give command line too long. Helmut.
[gentoo-user] Re: FIXED:Re: [drm] loading RV710 Microcode fails
James wireless at tampabay.rr.com writes: Do *not* install any firmware files and make sure you disable the radeon driver completely in the kernel. Not only KMS, but no driver *at all*. oops, I forgot to post the graphics section, for review: # Graphics support # CONFIG_AGP=y CONFIG_AGP_AMD64=y CONFIG_AGP_INTEL=y # CONFIG_AGP_SIS is not set # CONFIG_AGP_VIA is not set CONFIG_VGA_ARB=y CONFIG_VGA_ARB_MAX_GPUS=16 # CONFIG_VGA_SWITCHEROO is not set # CONFIG_DRM is not set # CONFIG_VGASTATE is not set # CONFIG_VIDEO_OUTPUT_CONTROL is not set # CONFIG_FB is not set # CONFIG_BACKLIGHT_LCD_SUPPORT is not set # # Display device support # # CONFIG_DISPLAY_SUPPORT is not set # Console display driver support # CONFIG_VGA_CONSOLE=y CONFIG_VGACON_SOFT_SCROLLBACK=y CONFIG_VGACON_SOFT_SCROLLBACK_SIZE=256 CONFIG_DUMMY_CONSOLE=y Any other suggestions? Ati-drivers is working again. James
Re: [gentoo-user] glibc 2.13 warning
On Wed, Feb 09 2011, Neil Bothwick wrote: On Wed, 09 Feb 2011 00:03:36 -0500, Allan Gottlieb wrote: The postfix issue is separate and needs a glibc downgrade to fix. I am not sure I understand. I did the glibc upgrade and have not downgraded. I run postfix and my mail is coming and going. I had thought/hoped that postfix does work with the new glibc if you don't do prelinking (which is my configuration: 2-13 glibc, postfix, no prelink). If glibc-2.13 kills postfix w/o prelinking then perhaps, I am just lucky and my overall configuration is such that it works for me. I think you're lucky, as is Volker. A number of people, me included, have had Postfix fail with postfix/local[4452]: fatal: unable to determine open file limit postfix/master[4001]: warning: process /usr/lib/postfix/local pid 4452 exit status 1 postfix/master[4001]: warning: /usr/lib/postfix/local: bad command startup -- throttling Prelinking is not involved, so it must be down to how you have configured Postfix, probably the local delivery settings. Thanks neil for the explanation and thank you my lucky stars for sparing me this mail failure. allan
Re: [gentoo-user] cups hplib - how to divorce
2011/2/9 Helmut Jarausch jarau...@igpm.rwth-aachen.de: AFAIR it worked just fine with CUPS-1.3.x Then why don't you use stable cups? I do have CONFIG_USB_PRINTER=y in my kernel configuration. New cups 1.4 can make use of the libusb package via the usb use flag. Without it uses the kernel USB support CONFIG_USB_PRINTER. If you want to use cups-1.4, I recommend to compile kernel USB support as module CONFIG_USB_PRINTER=m instead of built-in. This way you can test both methods without conflicts. Compile cups-1.4 with usb use flag and blacklist the usblp module to test if it works with libusb. To test if it works with the kernel module remove the blacklist entry and compile cups without usb use flag. Btw: cups-1.4 is not stable because of issues with usb support ([1] see also the depending bugs), so if both methods do not work for you I recommend using cups-1.3 for the time being. [1] http://bugs.gentoo.org/show_bug.cgi?id=333781 PS: You can only get rid of hplip if your printer is directly supported by cups and cups only supports a few HP printer. So in general you are better of with hplip. -- Daniel Pielmeier
[gentoo-user] Install unstable package in stage4 catalyst?
Hi, I'm trying to create a stage4 spec file. Running the spec, sys-apps/net-tools-1.60-r13 doesn't compile. There is a bug about this. So I want to specify for catalyst to install a later version of this package. How can I do that? Regards, Kfir
[gentoo-user] Disk Labels in Handbook
Hello, So looking at the handbook, I was wondering why it does not describe how to use Disk Labels during the installation process. Dunno. So I poised this question on gentoo-doc and got this encouraging response from *JOSH* snip Some discussion on modifying the Gentoo Handbook to describe how to use Disk Labels is warranted? Many have switched to disk labels in fstab. Maybe the Gentoo handbook should include a section on disk label usage to prepare for the future? and Josh replied: These aren't needed to get a system up and running. Yeah, Ubuntu uses them for various ID purposes, but nothing really critical. Unless there's a clear need for them, for example if some package in the @system set will use them in a way that our users will see, I doubt the handbook needs to cover them. Adding labels using udev just seems very finicky and time-consuming, and doesn't seem like it would be of much utility when users are already swamped with everything else the handbook asks them to do. If someone's willing to provide the text, then sure, I'll consider adding it to the handbooks, but if not, then I won't spend any energy trying to write something up. Things like GPT for hard disks are far more important, now that the large-size disks are shipping with it rather than the old MSDOS partition arrangement.. So, I opened up a bug on disk labels and the handbook, hoping that knowledgable users I will post to this bug # 354229. James
Re: [gentoo-user] Disk Labels in Handbook
James wrote: Hello, So looking at the handbook, I was wondering why it does not describe how to use Disk Labels during the installation process. Dunno. So I poised this question on gentoo-doc and got this encouraging response from *JOSH* snip James Given that some folks on here have ran into USB drives changing the order of partitions, I think this is a good idea. If needed, they could at least introduce the subject then have it link to another page. Even if it is the simplest label of using boot, root and such labels and maybe a mention that there are other ways to accomplish the same thing. I ran into this issue a while back when I added a hard drive and it was not easy to work with. When I boot a CD/DVD, it sees them as hd* instead of sd* so that didn't help since the OS kernel sees them as sd*. It may be uphill to get this included or at least linked to something else explaining it but I think it is a good idea. I also added myself to the bug as well. I saw the post on -doc. Dale :-) :-)
Re: [gentoo-user] Disk Labels in Handbook
On Wed, Feb 9, 2011 at 6:16 AM, Dale rdalek1...@gmail.com wrote: James wrote: Hello, So looking at the handbook, I was wondering why it does not describe how to use Disk Labels during the installation process. Dunno. So I poised this question on gentoo-doc and got this encouraging response from *JOSH* snip James Given that some folks on here have ran into USB drives changing the order of partitions, I think this is a good idea. If needed, they could at least introduce the subject then have it link to another page. Even if it is the simplest label of using boot, root and such labels and maybe a mention that there are other ways to accomplish the same thing. I ran into this issue a while back when I added a hard drive and it was not easy to work with. When I boot a CD/DVD, it sees them as hd* instead of sd* so that didn't help since the OS kernel sees them as sd*. It may be uphill to get this included or at least linked to something else explaining it but I think it is a good idea. I also added myself to the bug as well. I saw the post on -doc. Dale Following Walt's recent thread about his experiences using grub2 I think getting folks used to disk labels at installation time, be they names or even better UUID's, might fit in very well with installation instructions that cover using grub2 instead of grub as a boot loader. - Mark
Re: [gentoo-user] copy a bunch of files...
On Wed, Feb 9, 2011 at 4:09 AM, Helmut Jarausch jarau...@igpm.rwth-aachen.de wrote: On 02/09/2011 12:56:09 PM, Neil Bothwick wrote: On Wed, 9 Feb 2011 13:38:37 +0200, Kfir Lavi wrote: Another way to do it is with find: find /home/mark/Builder -type f -iname '*csv' -exec cp {} ~mark/CorrelationTests \; Replace \; with + for a faster process, as Mark said there are hundreds of these files. Or, if you use zsh instead of bash, it can be as simple as cp Builder/**/*.csv CorrelationTests There is a problem with this approach, though. It can easily give command line too long. Helmut. Lots of interesting ideas. I use apps so often I've never become very strong at the command line and yet people built this whole Linux empire using it. It's very powerful. One thing I didn't make clear in my original post - it didn't seem important to confuse my real question which was the copy itself and not locating the files - but which likely changes how well some of these commands would work in my specific case was that the Builder directory actually has _many_ CSV files ut specifically I needed only the ones in the Correlation directories. Additionally, being that this is stock futures trading data, generally at a given time I need the CSV files for a specific symbol, for instance in the original post: c2stable ~ # locate Correlation | grep Builder | grep csv /home/mark/Builder/TF/TF.D-17M-2009_06-2010_11/Correlation/TF.D-17M-2009_06-2010_11-V1.csv /home/mark/Builder/TF/TF.D-17M-2009_06-2010_11/Correlation/TF.D-17M-2009_06-2010_11-V2.csv /home/mark/Builder/TF/TF.D-17M-2009_06-2010_11/Correlation/TF.D-17M-2009_06-2010_11-V3.csv SNIP /home/mark/Builder/TF/TF.D-31M-2009_06-2010_11/Correlation/TF.D-31M-2009_06-2010_11-V4.csv /home/mark/Builder/TF/TF.D-31M-2009_06-2010_11/Correlation/TF.D-31M-2009_06-2010_11-V5.csv c2stable ~ # I knew I wanted the Correlation directory but it turned out I had other directories with that name on the system, so I added the grep Builder to get me into the right tree and CSV to find only the CSV files. However at that point I only had Russell futures data (TF.D) so I didn't have to go further. Now, however, as I bring in Dow futures (YM, YM.D) , SP 500 futures (ES, ES.D), and NASDAQ futures (NQ, ND.D) I just an an extra grep in and I'm there in terms of finding the files I need for a certain test. Additionally I have test results for other date ranges that will show up soon, (2001-2005, or 2003-2011, etc. Additional greps are an easy way to just winnow it down to a point where I'm finding what I need to find. This thread has given me a lot of new commands to go look at. Thanks, Mark
Re: [gentoo-user] Disk Labels in Handbook
On 9. 2. 2011 15:16, Dale wrote: So looking at the handbook, I was wondering why it does not describe how to use Disk Labels during the installation process. Dunno. Given that some folks on here have ran into USB drives changing the order of partitions, I think this is a good idea. The same happened to me when I attached one more sata-drive. After that those two which used to be /dev/sda and /dev/sdb suddenly became /dev/sdb and /dev/sdc (and the new one was now /dev/sda). No matter how I tried to switch cables or ports, they kept to be detected as 2nd and 3rd sata-drives. So I agree there should be at least some basic info about disk-labels in the handbook. Jarry -- ___ This mailbox accepts e-mails only from selected mailing-lists! Everything else is considered to be spam and therefore deleted.
[gentoo-user] Re: Bug#354229 Disk Labels in Handbook
Jarry mr.jarry at gmail.com writes: So I agree there should be at least some basic info about disk-labels in the handbook. OK guys let's all chime on on the bug with ideas and verbiage, or we just need one really smart admin, to post some text and we're off to the races. Me, I still struggle with Disk labels and UUIDs and such, so I'm looking for leadership here. You know one of those smart kids or elder (older than me) linux gentlemen types... Don't be shy! bug 354229 James
Re: [gentoo-user] Disk Labels in Handbook
On Wed, 9 Feb 2011 13:51:46 + (UTC), James wrote: These aren't needed to get a system up and running. Yeah, Ubuntu uses them for various ID purposes, but nothing really critical. Unless there's a clear need for them, for example if some package in the @system set will use them in a way that our users will see, I doubt the handbook needs to cover them. Adding labels using udev just seems very finicky and time-consuming, and doesn't seem like it would be of much utility when users are already swamped with everything else the handbook asks them to do. Josh seems to be referring to UUID type disk labels, not filesystem labels. Which are you talking about? signature.asc Description: PGP signature
Re: [gentoo-user] Re: Bug#354229 Disk Labels in Handbook
In linux.gentoo.user, you wrote: Jarry mr.jarry at gmail.com writes: So I agree there should be at least some basic info about disk-labels in the handbook. OK guys let's all chime on on the bug with ideas and verbiage, or we just need one really smart admin, to post some text and we're off to the races. Me, I still struggle with Disk labels and UUIDs and such, so I'm looking for leadership here. You know one of those smart kids or elder (older than me) linux gentlemen types... Don't be shy! bug 354229 James I'm a little confused about use of the term disk labels in this discussion. Isn't a disk label a fs level ID (I create those when I make my fs)? Using UUID in fstab for quite awhile here (due to multiple external drives), but those aren't disk labels... are they? It seems there are two subjects here, one to address in the section of preparing disks and one for setting up fstab. Unless of course I'm missing something, which is entirely possible... -- caveat utilitor ♫ ❤ ♫ ❤ ♫ ❤ ♫
Re: [gentoo-user] copy a bunch of files...
On Wed, 9 Feb 2011 06:46:34 -0800, Mark Knecht wrote: One thing I didn't make clear in my original post - it didn't seem important to confuse my real question which was the copy itself and not locating the files - but which likely changes how well some of these commands would work in my specific case was that the Builder directory actually has _many_ CSV files ut specifically I needed only the ones in the Correlation directories. Bear in mind that locate only returns files in its database, not any created since the last time its cron job was run. Find seems a more appropriate tool for this task. signature.asc Description: PGP signature
Re: [gentoo-user] i586 stage3 tarball
On 20:07 Tue 08 Feb , Mark Shields wrote: Interesting. Did you build this up from a stage1 or stage2 tarball? Hi Mark, actually, neither. ;-) I just took an existing stage3 tarball that I unpacked and chrooted into. I then changed the CHOST / CFLAGS and rebuild everything, finally tarring the whole thing back up again. Greetings, Nils -- Nils Holland * Ti Systems, Wunstorf-Luthe (Germany) Powered by GNU/Linux since 1998
Re: [gentoo-user] Disk Labels in Handbook
Apparently, though unproven, at 16:27 on Wednesday 09 February 2011, Mark Knecht did opine thusly: On Wed, Feb 9, 2011 at 6:16 AM, Dale rdalek1...@gmail.com wrote: James wrote: Hello, So looking at the handbook, I was wondering why it does not describe how to use Disk Labels during the installation process. Dunno. So I poised this question on gentoo-doc and got this encouraging response from *JOSH* snip James Given that some folks on here have ran into USB drives changing the order of partitions, I think this is a good idea. If needed, they could at least introduce the subject then have it link to another page. Even if it is the simplest label of using boot, root and such labels and maybe a mention that there are other ways to accomplish the same thing. I ran into this issue a while back when I added a hard drive and it was not easy to work with. When I boot a CD/DVD, it sees them as hd* instead of sd* so that didn't help since the OS kernel sees them as sd*. It may be uphill to get this included or at least linked to something else explaining it but I think it is a good idea. I also added myself to the bug as well. I saw the post on -doc. Dale Following Walt's recent thread about his experiences using grub2 I think getting folks used to disk labels at installation time, be they names or even better UUID's, might fit in very well with installation instructions that cover using grub2 instead of grub as a boot loader. From a practical perspective, fs labels are easier than GUIDs, so I would recommend labels. Users can invent their own descriptive labels at install time and enter that into fstab. LABEL=SERVER1-ROOT is not much more effort than /dev/sda3 GUIDs are another story. They get autogenerated, are invariably displayed on the screen along with a huge number of other GUIDs (Murphy) and one has to copy paste the damn things into vi. GUIDs are great for ubuntu where an install script does all the heavy lifting, but Gentoo, being a vastly superior operating system, has made the devastatingly astounding assumption that users are actually able to think and type. Whodathunkedit? If we expect users to type stuff, we should set it up so they type easy stuff :-) -- alan dot mckinnon at gmail dot com
Re: [gentoo-user] copy a bunch of files...
On Wed, Feb 9, 2011 at 12:10 PM, Neil Bothwick n...@digimed.co.uk wrote: On Wed, 9 Feb 2011 06:46:34 -0800, Mark Knecht wrote: One thing I didn't make clear in my original post - it didn't seem important to confuse my real question which was the copy itself and not locating the files - but which likely changes how well some of these commands would work in my specific case was that the Builder directory actually has _many_ CSV files ut specifically I needed only the ones in the Correlation directories. Bear in mind that locate only returns files in its database, not any created since the last time its cron job was run. Find seems a more appropriate tool for this task. Good point. As I bring these files down from a bunch of Windows VMs I typically run updatedb before doing this, but find would probably be safer. Thanks, Mark
Re: [gentoo-user] glibc 2.13 warning
On 13:34 Mon 07 Feb , Neil Bothwick wrote: Don't install glibc-2.13 if you either use prelinking or run postfix. After testing it on my netbook, which uses neither, I installed it on my desktop and home server and broke both. Thanks a lot, I've read your mail just in time. Actually, glibc 2.13 krept onto my first machine Monday night - I generally test new versions of such far reaching stuff as glibc on a single machine first before letting to onto all of my boxes. I didn't have any problems with the new glibc, and tonight I would have updated a few additional machines, one of which happens to run Postfix. I guess I'm going to delay that a bit now. ;-) Greetings, Nils -- Nils Holland * Ti Systems, Wunstorf-Luthe (Germany) Powered by GNU/Linux since 1998
Re: [gentoo-user] glibc 2.13 warning
On Wed, 9 Feb 2011 23:23:50 +0100, Nils Holland wrote: Thanks a lot, I've read your mail just in time. Actually, glibc 2.13 krept onto my first machine Monday night - I generally test new versions of such far reaching stuff as glibc on a single machine first before letting to onto all of my boxes. I didn't have any problems with the new glibc, and tonight I would have updated a few additional machines, one of which happens to run Postfix. That's what happened to me, I updated one box, rebooted, made sure things worked and then updated the Postfix server and the prelinked desktop. signature.asc Description: PGP signature
[gentoo-user] Re: Bug#354229 Disk Labels in Handbook
On 02/09/2011 12:05 PM, Elaine C. Sharpe wrote: I'm a little confused about use of the term disk labels in this discussion. Isn't a disk label a fs level ID (I create those when I make my fs)? Using UUID in fstab for quite awhile here (due to multiple external drives), but those aren't disk labels... are they? Correct, they are two different (but equivalent) ways of naming a filesystem (partition) for use in fstab. mkfs generates a UUID automatically when the fs is created, but it does *not* generate a label unless you give it one using the -L flag, or create one later using e2fslabel or some other utility. The syntax in fstab is either UUID=foo or LABEL=bar, but the idea is exactly the same. The whole point is to divorce the fs from the /dev/ it happens to reside on at boot time.
[gentoo-user] Re: glibc 2.13 warning
On 02/09/2011 02:31 PM, Neil Bothwick wrote: On Wed, 9 Feb 2011 23:23:50 +0100, Nils Holland wrote: Thanks a lot, I've read your mail just in time. Actually, glibc 2.13 krept onto my first machine Monday night - I generally test new versions of such far reaching stuff as glibc on a single machine first before letting to onto all of my boxes. I didn't have any problems with the new glibc, and tonight I would have updated a few additional machines, one of which happens to run Postfix. That's what happened to me, I updated one box, rebooted, made sure things worked and then updated the Postfix server and the prelinked desktop. Could you explain a bit about prelinking? Does it have anything to do with the sys-devel/prelink package or the gentoo 'prelink' useflag? Or maybe Alan, if he's done polishing his humility plugin :p
Re: [gentoo-user] possible portage or ebuild bug?
Apparently, though unproven, at 00:03 on Wednesday 09 February 2011, Alan McKinnon did opine thusly: Apparently, though unproven, at 17:41 on Monday 07 February 2011, cov...@ccs.covici.com did opine thusly: On trying my last world update with --deep and --newuse, etc. I get the following message: Calculating dependencies .. done! !!! All ebuilds that could satisfy =dev-libs/glib-2.27.5 have been masked. !!! One of the following masked packages is required to complete your request: - dev-libs/glib-::gnome (masked by: package.mask, missing keyword) /usr/portage/profiles/package.mask: # Gilles Dartiguelongue e...@gentoo.org (04 Feb 2011) # New glib/gtk+ mask, for testing purpose # Needs a new gvfs as well - dev-libs/glib-2.27.93::gnome (masked by: package.mask) - dev-libs/glib-2.27.93::gentoo (masked by: package.mask) (dependency required by x11-libs/gtk+-2.99.3 [ebuild]) (dependency required by media-gfx/graphviz-2.26.3-r3[gtk] [installed]) (dependency required by app-doc/doxygen-1.7.3[-nodot] [ebuild]) (dependency required by media-libs/id3lib-3.8.3-r8 [installed]) (dependency required by media-sound/id3v2-0.1.12 [installed]) (dependency required by @selected [set]) (dependency required by @world [argument]) For more information, see the MASKED PACKAGES section in the emerge man page or refer to the Gentoo Handbook. Now I tried the following in my package.mask with no difference: =x11-libs/gtk+-2.25.0 Do you have some gnome overlay installed? None of the high-versioned packages you need (glib, gtk+) are in the regular tree. Looks like my sync yesterday was behind. Today's sync gives me the same error you got. Latest unstable gtk+-2.24.0 DEPENDS on =dev-libs/glib-2.27.3 (hard masked). $PORTDIR/profiles/package.mask lists =dev-libs/glib-2.27 =x11-libs/gtk+-2.24 So the gtk+ line is wrong. It should be either = or version 2.24.0 Pacho fixed it http://bugs.gentoo.org/show_bug.cgi?id=354275 so a resync again should resolve your issue So what am I missing here, or is something broke? Thanks much for any help you can give. My guess is something in the overlay depends on a version of gtk+ that isn't in the overlay yet, except for the dev versions *99*. And those are hard and keyword masked as usual for cvs/svn versions. resync the overlay, try again. It's an overlay. It's certain to break once a day. Less than that is a bonus. -- alan dot mckinnon at gmail dot com
Re: [gentoo-user] glibc 2.13 warning
Apparently, though unproven, at 00:23 on Thursday 10 February 2011, Nils Holland did opine thusly: On 13:34 Mon 07 Feb , Neil Bothwick wrote: Don't install glibc-2.13 if you either use prelinking or run postfix. After testing it on my netbook, which uses neither, I installed it on my desktop and home server and broke both. Thanks a lot, I've read your mail just in time. Actually, glibc 2.13 krept onto my first machine Monday night - I generally test new versions of such far reaching stuff as glibc on a single machine first before letting to onto all of my boxes. I didn't have any problems with the new glibc, and tonight I would have updated a few additional machines, one of which happens to run Postfix. I guess I'm going to delay that a bit now. ;-) This raises an interesting point. glibc is a problematic package, it's tentacles run very deep in any GNU system, it has a less than stellar history in terms of breaking gentoo systems (mostly due to inadequate testing before releasing to ~arch) And it's very difficult to downgrade it due to that hidden barf check in the ebuild. I have yet to find a supported, documented way to back out of glibc screw-ups; my way is to keep binpkgs of @system and use those. Yes it's true that downgrading glibc is often a sure road to suicide, but the current method is also unworkable. Surely, surely, there's a better way? I'd even go so far as to support a portage feature-request: automatic binpkgs of a sub-set of @system that the user must opt-out of in make.conf: python, portage, glibc, gcc, maybe a few other highly critical packages. What say you all? -- alan dot mckinnon at gmail dot com
Re: [gentoo-user] glibc 2.13 warning
On Thu, 10 Feb 2011 01:48:18 +0200, Alan McKinnon wrote: And it's very difficult to downgrade it due to that hidden barf check in the ebuild. I have yet to find a supported, documented way to back out of glibc screw-ups; my way is to keep binpkgs of @system and use those. The trouble is that binpkgs keep a copy of the ebuild in them, so even if you remove the downgrade check fro the in-tree ebuild, it still fails. That one had me scratching my head for a few minutes. The thing is, a downgrade like that one is not a problem, especially if done soon after the upgrade. The problems arise when you build other packages against the later glibc and then downgrade. We need a more intelligent test and we need a way of circumventing the restriction that doesn't involve editing the ebuild, something like I_KNOW_ITS_DODGY_BUT_IM_DESPERATE=true emerge \sys-libs/glibc-2.13 -- Neil Bothwick Famed tautologist dies of suicide in distressing tragedy signature.asc Description: PGP signature
Re: [gentoo-user] Re: glibc 2.13 warning
On Wed, 09 Feb 2011 15:34:59 -0800, walt wrote: On 02/09/2011 02:31 PM, Neil Bothwick wrote: On Wed, 9 Feb 2011 23:23:50 +0100, Nils Holland wrote: Thanks a lot, I've read your mail just in time. Actually, glibc 2.13 krept onto my first machine Monday night - I generally test new versions of such far reaching stuff as glibc on a single machine first before letting to onto all of my boxes. I didn't have any problems with the new glibc, and tonight I would have updated a few additional machines, one of which happens to run Postfix. That's what happened to me, I updated one box, rebooted, made sure things worked and then updated the Postfix server and the prelinked desktop. Could you explain a bit about prelinking? Does it have anything to do with the sys-devel/prelink package or the gentoo 'prelink' useflag? http://www.gentoo.org/doc/en/prelink-howto.xml explains it, although it misses out the part about it speeding up the reduction of your system to a useless wreck with a broken glibc :( Apparently, prelinking is less of a benefit than it used to be. I can't say I've noticed any difference having removed it this wekk and may well not reinstall it when glibc is fixed. I suspect a lot will feel the same. Or maybe Alan, if he's done polishing his humility plugin :p What make you think he's got one... or even know what it is :) -- Neil Bothwick Sir! Romulan warbird decloaki»®õ÷üÁ NO CARRIER signature.asc Description: PGP signature
Re: [gentoo-user] glibc 2.13 warning
110210 Alan McKinnon wrote: glibc is problematic, it's tentacles run very deep in any GNU system, it has a less than stellar history in terms of breaking gentoo systems, mostly due to inadequate testing before releasing to ~arch. it's v difficult to downgrade it due to that hidden barf check in the ebuild. I'd support a portage feature-request: automatic binpkgs of a sub-set of @system that the user must opt-out of in make.conf: python, portage, glibc, gcc, maybe a few others What say you all? I avoid such problems by 2 simple precautions: (1) I never use testing versions of system pkgs like Glibc (2) I have FEATURES=buildsyspkg in make.conf . Beyond those, I'ld say Gentoo users are grown-ups who don't need coddling. -- ,, SUPPORT ___//___, Philip Webb ELECTRIC /] [] [] [] [] []| Cities Centre, University of Toronto TRANSIT`-O--O---' purslowatchassdotutorontodotca
Re: [gentoo-user] Re: glibc 2.13 warning
Apparently, though unproven, at 02:05 on Thursday 10 February 2011, Neil Bothwick did opine thusly: Or maybe Alan, if he's done polishing his humility plugin :p What make you think he's got one... or even know what it is :) Yeah, come to think of it, what is that thing anyway? Must be related to the certificate I got at last year's end-of-year function. After all the usual ones (employee of the quarter), and the unusual ones (most gratuitous use of the word fuck in a work context - [my manager]) there was a special presentation: PEDANTIC OVERLORD For ruling Unix shell and Tacacs access with an adamantium fist. No exceptions! = Presented to: you guessed it :-) -- alan dot mckinnon at gmail dot com
Re: [gentoo-user] glibc 2.13 warning
On Wed, 9 Feb 2011 19:09:14 -0500, Philip Webb wrote: (1) I never use testing versions of system pkgs like Glibc Someone has to or they'll never get tested. (2) I have FEATURES=buildsyspkg in make.conf . It didn't help here. Beyond those, I'ld say Gentoo users are grown-ups who don't need coddling. The problem is that we are being coddled with the we won't let you downgrade because we don't think it is safe for you ebuilds. We're asking for less coddling, to be able to make our own decisions and be able to keep the pieces if they turn out to be wrong. A stern warning should be sufficient. -- Neil Bothwick Documentation: (n.) a novel sold with software, designed to entertain the operator during episodes of bugs or glitches. signature.asc Description: PGP signature
Re: [gentoo-user] Re: glibc 2.13 warning
Alan McKinnon wrote: Apparently, though unproven, at 02:05 on Thursday 10 February 2011, Neil Bothwick did opine thusly: Or maybe Alan, if he's done polishing his humility plugin :p What make you think he's got one... or even know what it is :) Yeah, come to think of it, what is that thing anyway? Must be related to the certificate I got at last year's end-of-year function. After all the usual ones (employee of the quarter), and the unusual ones (most gratuitous use of the word fuck in a work context - [my manager]) there was a special presentation: PEDANTIC OVERLORD For ruling Unix shell and Tacacs access with an adamantium fist. No exceptions! = Presented to: you guessed it :-) This reminds me of when I left my puter job. Windows 3.1 came out. I worked with it for a few months then turned in my notice. The song they played at the going away party, you are my hero. I went to work in sales for a magazine distributor, no computer needed. They thought it was brave of me to change careers. Come to think of it, some may have wished they had changed to right now. The computer field around here is a bit . . . crazy. It's also a low paying job if you can find one. Now some of you know how much I hate windows. Of course, DOS wasn't much better, Dale :-) :-)
Re: [gentoo-user] Re: glibc 2.13 warning
=== On Wed, 02/09, Dale wrote: === Now some of you know how much I hate windows. Of course, DOS wasn't much better, === Yep. I've been using Linux on my desktop since version 1.2, and UnixWare before that. Some Mac in there too. I avoid Windows like the plague that it is. -- Keith Dart -- -- ~ Keith Dart ke...@dartworks.biz public key: ID: 19017044 http://www.dartworks.biz/ =
Re: [gentoo-user] possible portage or ebuild bug?
Alan McKinnon alan.mckin...@gmail.com wrote: Apparently, though unproven, at 00:03 on Wednesday 09 February 2011, Alan McKinnon did opine thusly: Apparently, though unproven, at 17:41 on Monday 07 February 2011, cov...@ccs.covici.com did opine thusly: On trying my last world update with --deep and --newuse, etc. I get the following message: Calculating dependencies .. done! !!! All ebuilds that could satisfy =dev-libs/glib-2.27.5 have been masked. !!! One of the following masked packages is required to complete your request: - dev-libs/glib-::gnome (masked by: package.mask, missing keyword) /usr/portage/profiles/package.mask: # Gilles Dartiguelongue e...@gentoo.org (04 Feb 2011) # New glib/gtk+ mask, for testing purpose # Needs a new gvfs as well - dev-libs/glib-2.27.93::gnome (masked by: package.mask) - dev-libs/glib-2.27.93::gentoo (masked by: package.mask) (dependency required by x11-libs/gtk+-2.99.3 [ebuild]) (dependency required by media-gfx/graphviz-2.26.3-r3[gtk] [installed]) (dependency required by app-doc/doxygen-1.7.3[-nodot] [ebuild]) (dependency required by media-libs/id3lib-3.8.3-r8 [installed]) (dependency required by media-sound/id3v2-0.1.12 [installed]) (dependency required by @selected [set]) (dependency required by @world [argument]) For more information, see the MASKED PACKAGES section in the emerge man page or refer to the Gentoo Handbook. Now I tried the following in my package.mask with no difference: =x11-libs/gtk+-2.25.0 Do you have some gnome overlay installed? None of the high-versioned packages you need (glib, gtk+) are in the regular tree. Looks like my sync yesterday was behind. Today's sync gives me the same error you got. Latest unstable gtk+-2.24.0 DEPENDS on =dev-libs/glib-2.27.3 (hard masked). $PORTDIR/profiles/package.mask lists =dev-libs/glib-2.27 =x11-libs/gtk+-2.24 So the gtk+ line is wrong. It should be either = or version 2.24.0 Pacho fixed it http://bugs.gentoo.org/show_bug.cgi?id=354275 so a resync again should resolve your issue So what am I missing here, or is something broke? Thanks much for any help you can give. My guess is something in the overlay depends on a version of gtk+ that isn't in the overlay yet, except for the dev versions *99*. And those are hard and keyword masked as usual for cvs/svn versions. resync the overlay, try again. It's an overlay. It's certain to break once a day. Less than that is a bonus. Thanks much -- I had unmasked glib2.27.5, but I will get rid of that and resync again -- did not really want to use that. Thanks. -- Your life is like a penny. You're going to lose it. The question is: How do you spend it? John Covici cov...@ccs.covici.com
Re: [gentoo-user] glibc 2.13 warning
110210 Neil Bothwick wrote: On Wed, 9 Feb 2011 19:09:14 -0500, Philip Webb wrote: (1) I never use testing versions of system pkgs like Glibc Someone has to or they'll never get tested. Come on ! -- not on a production system ! (2) I have FEATURES=buildsyspkg in make.conf . It didn't help here. The OP was making a proposal to solve the more general problem, incl requiring users to adopt (2) by default. -- ,, SUPPORT ___//___, Philip Webb ELECTRIC /] [] [] [] [] []| Cities Centre, University of Toronto TRANSIT`-O--O---' purslowatchassdotutorontodotca