Re: [gentoo-user] two glibcs with different version
Kraus Philipp writes: > Hello, > > I must test a software with a older version of the glibc. I run the > 2.11.1 now but for one tool I need a previous version (2.6.1). > How can I compile the glibc without changing my system glibc. I would > like to set the previous glibc with the LD_PATH. > Can I run two different versions or is a better solution to downgrade > the system glib? I think that the only way you can do this is to create a chroot jail, in which you build everything using the old version of glibc (in a very similar way to building a new Gentoo system) and run your application in that.
Re: [gentoo-user] two glibcs with different version
On 01.05.2010 um 09:48 wrote Graham Murray: I think that the only way you can do this is to create a chroot jail, in which you build everything using the old version of glibc (in a very similar way to building a new Gentoo system) and run your application in that. That's a very goog idea, because the programm is a network service. Do you know a tutorial for creating root jails in this way? Thanks
Re: [gentoo-user] two glibcs with different version
On Samstag 01 Mai 2010, Graham Murray wrote: > Kraus Philipp writes: > > Hello, > > > > I must test a software with a older version of the glibc. I run the > > 2.11.1 now but for one tool I need a previous version (2.6.1). > > How can I compile the glibc without changing my system glibc. I would > > like to set the previous glibc with the LD_PATH. > > Can I run two different versions or is a better solution to downgrade > > the system glib? > > I think that the only way you can do this is to create a chroot jail, > in which you build everything using the old version of glibc (in a very > similar way to building a new Gentoo system) and run your application in > that. no, you can install glibc in /usr/local and then tell apps to either use the libs in /usr/local or /usr. It is just not easy because it easily breaks stuff in horrrible to fix ways.
[gentoo-user] Re: Compiling 32 bit library on x86_64
On 04/30/2010 07:29 PM, David W Noon wrote: On Fri, 30 Apr 2010 15:20:02 +0200, Nikos Chantziaras wrote about [gentoo-user] Re: Compiling 32 bit library on x86_64: On 04/30/2010 03:09 PM, David W Noon wrote: On Fri, 30 Apr 2010 12:10:02 +0200, Roger Mason wrote about [gentoo-user] Compiling 32 bit library on x86_64: Hello, I need to compile a 32 bit version of libtermcap on an x86_64 (multilib) system. Can someone tell me how to set up CFLAGS? This is what I have at the moment: CFLAGS="-O2 -m32 -march=native -msse3 -pipe" CXXFLAGS="-O2 -m32 -march=native -msse3 -pipe" The -march=native will shoot you in the foot. Pick a 32-bit architecture and use that instead; e.g. -march=i686 Then, -msse3 could also be problematic, unless the target is a very late model Pentium 4. I would ditch that too. None of those options are problematic. -march=native has nothing to do with 32/64 bit. Every 64-bit CPU is 32-bit compatible and has zero consequence. I think you fell into the logical trap that 32-bit CPUs are not 64-bit compatible but it's OK vice versa :) Meaning you can't use "-m64 -march=i686". But you *can* and *should* use "-m32 -march=core2". No, I stand by what I wrote. You are wrong. The -march=native option tells the compiler to issue the CPUID instruction to determine the architecture. This means that on an amd64 box it will return data for either an AMD K8 or an Intel Pentium D architecture. This, in turn, allows the compiler to generate K8 instructions that are not valid on IA32 processors. It even allows the compiler to use 64-bit registers, including the additional registers that were not in an IA32 processor. The OP was not interested whether the code runs on all IA32 processors. Just on his. Yes, if you compile with "-m32 -march=native" the code will not run everywhere else. But I fail to see what this has to do with 32-bit vs 64-bit; it also happens if you omit "-m32". The -m32 option instructs the compiler to generate code with 32-bit pointers and relocation dictionary. It does not constrain the compiler to generate code that will definitely run on an IA32 processor, but it does ensure that the code can be linked with 32-bit libraries. So, if one is compiling on, say, a Core2 Duo and one uses -march=native and -m32, the compiler can use all kinds of instructions valid on the Core2 Duo, but limits addressing to 32-bit. Yes, which is what we want. Again, the OP was not interested whether the code runs on other machines. If that was the case, he would not have used -march=native in the first place. And I have proof too. Compile whatever you want with -m32 -march=native. It will work just fine.
Re: [gentoo-user] two glibcs with different version
On 01.05.2010 um 10:32 wrote Volker Armin Hemmann: On Samstag 01 Mai 2010, Graham Murray wrote: Kraus Philipp writes: Hello, I must test a software with a older version of the glibc. I run the 2.11.1 now but for one tool I need a previous version (2.6.1). How can I compile the glibc without changing my system glibc. I would like to set the previous glibc with the LD_PATH. Can I run two different versions or is a better solution to downgrade the system glib? I think that the only way you can do this is to create a chroot jail, in which you build everything using the old version of glibc (in a very similar way to building a new Gentoo system) and run your application in that. no, you can install glibc in /usr/local and then tell apps to either use the libs in /usr/local or /usr. It is just not easy because it easily breaks stuff in horrrible to fix ways. Okay, can I downgrade my glibc? My Gentoo isn't a big system, it's a server installation, so I can recompile the whole system. I had forgotten to mask the glibc on the last update. I have add a line to the portage.mask but emerge says that it can't compile the older version, because will damage the system.
Re: [gentoo-user] Compiling 32 bit library on x86_64
Mike Edenfield writes: > On 4/30/2010 12:40 PM, Roger Mason wrote: >> Mike Edenfield writes: >> >>> Have you tried using sys-devel/crossdev? >> >> Not in the present context. >> >>> It will set up the entire 32-bit cross-compiler environment for you; >>> then it's just a matter of setting a couple of environment variables to >>> switch compilers. >> >> Some time ago I tried setting up cross-compilation so that I could use a >> rather heterogeneous collection of amd64, ppc and x86 machines in >> icecream. Unfortunately I could not get cross-compilation to work. I >> asked about it in this forum but did not get any replies. > > I have it set up on my laptop. I admit it's been a while since I used > it, but I know it worked at one point. > > Though I was using it on a standard PC, the best source of information I > found on the process was the Gentoo Embedded Handbook: > > > http://www.gentoo.org/proj/en/base/embedded/handbook/?part=1 > > The whole first section is on setting up a cross-compiler, just > substitute i686-pc-linux-gnu for your target architecture. Ah, thank you. I'll have another go at it this summer. Roger
Re: [gentoo-user] two glibcs with different version
On Sat, 2010-05-01 at 10:57 +0200, Kraus Philipp wrote: > > On 01.05.2010 um 10:32 wrote Volker Armin Hemmann: > > > On Samstag 01 Mai 2010, Graham Murray wrote: > > > Kraus Philipp writes: > > > > Hello, > > > > > > > > I must test a software with a older version of the glibc. I run > > > > the > > > > 2.11.1 now but for one tool I need a previous version (2.6.1). > > > > How can I compile the glibc without changing my system glibc. I > > > > would > > > > like to set the previous glibc with the LD_PATH. > > > > Can I run two different versions or is a better solution to > > > > downgrade > > > > the system glib? > > > > > > I think that the only way you can do this is to create a chroot > > > jail, > > > in which you build everything using the old version of glibc (in a > > > very > > > similar way to building a new Gentoo system) and run your > > > application in > > > that. > > > > no, you can install glibc in /usr/local and then tell apps to either > > use the > > libs in /usr/local or /usr. > > > > It is just not easy because it easily breaks stuff in horrrible to > > fix ways. > > > > > Okay, can I downgrade my glibc? My Gentoo isn't a big system, it's a > server > installation, so I can recompile the whole system. I had forgotten to > mask the > glibc on the last update. I have add a line to the portage.mask but > emerge says > that it can't compile the older version, because will damage the > system. > > Would LD_PRELOAD solve your problem? - worked for me when needing to run a legacy redhat app in the past on a more up-to-date gentoo system. There is also a LD_LIBRARY_PATH variable. Get a binary copy of the libs you need and put em somewhere convenient and let the rest of the system stay as is. google for LD_PRELOAD. BillK
[gentoo-user] convert google newsgroup digest to mbox
Hello. Does anybody know if there is an application to convert to the mbox format the email messages sent by groups.google.com in a digest format to the group subscribers (if the subscriber choosed this format)? Romildo
Re: [gentoo-user] two glibcs with different version
Am 01.05.2010 um 11:48 schrieb William Kenworthy: On Sat, 2010-05-01 at 10:57 +0200, Kraus Philipp wrote: On 01.05.2010 um 10:32 wrote Volker Armin Hemmann: On Samstag 01 Mai 2010, Graham Murray wrote: Kraus Philipp writes: Hello, I must test a software with a older version of the glibc. I run the 2.11.1 now but for one tool I need a previous version (2.6.1). How can I compile the glibc without changing my system glibc. I would like to set the previous glibc with the LD_PATH. Can I run two different versions or is a better solution to downgrade the system glib? I think that the only way you can do this is to create a chroot jail, in which you build everything using the old version of glibc (in a very similar way to building a new Gentoo system) and run your application in that. no, you can install glibc in /usr/local and then tell apps to either use the libs in /usr/local or /usr. It is just not easy because it easily breaks stuff in horrrible to fix ways. Okay, can I downgrade my glibc? My Gentoo isn't a big system, it's a server installation, so I can recompile the whole system. I had forgotten to mask the glibc on the last update. I have add a line to the portage.mask but emerge says that it can't compile the older version, because will damage the system. Would LD_PRELOAD solve your problem? - worked for me when needing to run a legacy redhat app in the past on a more up-to-date gentoo system. I think that can solve my problem, because it's only this one lib all other libs work very well. There is also a LD_LIBRARY_PATH variable. Get a binary copy of the libs you need and put em somewhere convenient and let the rest of the system stay as is. I don't have the glibc binary. I can't emerge it and if I try to compile from the sources. The configure script says: These critical programs are missing or too old: as ld How I can compile the from the sources (http://ftp.gnu.org/gnu/glibc/glibc-2.10.1.tar.gz ) ? Thanks Phil
Re: [gentoo-user] two glibcs with different version
On Sat, 2010-05-01 at 12:25 +0200, Kraus Philipp wrote: > > Am 01.05.2010 um 11:48 schrieb William Kenworthy: > > > On Sat, 2010-05-01 at 10:57 +0200, Kraus Philipp wrote: > > > > > > On 01.05.2010 um 10:32 wrote Volker Armin Hemmann: > > > > > > > On Samstag 01 Mai 2010, Graham Murray wrote: > > > > > Kraus Philipp writes: > > > > > > Hello, > > > > > > > > > > > > I must test a software with a older version of the glibc. I > > > > > > run > > > > > > the > > > > > > 2.11.1 now but for one tool I need a previous version > > > > > > (2.6.1). > > > > > > How can I compile the glibc without changing my system > > > > > > glibc. I > > > > > > would > > > > > > like to set the previous glibc with the LD_PATH. > > > > > > Can I run two different versions or is a better solution to > > > > > > downgrade > > > > > > the system glib? > > > > > > > > > > I think that the only way you can do this is to create a > > > > > chroot > > > > > jail, > > > > > in which you build everything using the old version of glibc > > > > > (in a > > > > > very > > > > > similar way to building a new Gentoo system) and run your > > > > > application in > > > > > that. > > > > > > > > no, you can install glibc in /usr/local and then tell apps to > > > > either > > > > use the > > > > libs in /usr/local or /usr. > > > > > > > > It is just not easy because it easily breaks stuff in horrrible > > > > to > > > > fix ways. > > > > > > > > > > > > > Okay, can I downgrade my glibc? My Gentoo isn't a big system, it's > > > a > > > server > > > installation, so I can recompile the whole system. I had forgotten > > > to > > > mask the > > > glibc on the last update. I have add a line to the portage.mask > > > but > > > emerge says > > > that it can't compile the older version, because will damage the > > > system. > > > > > > > > > > Would LD_PRELOAD solve your problem? - worked for me when needing to > > run > > a legacy redhat app in the past on a more up-to-date gentoo system. > > > > > I think that can solve my problem, because it's only this one lib all > other libs > work very well. > > > There is also a LD_LIBRARY_PATH variable. Get a binary copy of the > > libs > > you need and put em somewhere convenient and let the rest of the > > system > > stay as is. > > > I don't have the glibc binary. I can't emerge it and if I try to > compile from the sources. > The configure script says: These critical programs are missing or too > old: as ld > > > How I can compile the from the sources > (http://ftp.gnu.org/gnu/glibc/glibc-2.10.1.tar.gz) ? > > > Thanks > > > Phil You can use ebuild to build a package, but not install it - then just un-compress it and grab the wanted libs. Or if you tell us what arch, someone (me if its compatible) can create a binary pkg for you using quickpkg. It might also be on the livecd/install medium as well - dont have one here to check. BillK -- William Kenworthy Home in Perth!
Re: [gentoo-user] two glibcs with different version
On 2010-04-30 15:24, Kraus Philipp wrote: > I must test a software with a older version of the glibc. I run the > 2.11.1 now but for one tool I need a previous version (2.6.1). > How can I compile the glibc without changing my system glibc. I would > like to set the previous glibc with the LD_PATH. > Can I run two different versions or is a better solution to downgrade > the system glib? I would suggest dual-booting... Best regards Peter K
Re: [gentoo-user] two glibcs with different version
Am 01.05.2010 um 12:40 schrieb William Kenworthy: On Sat, 2010-05-01 at 12:25 +0200, Kraus Philipp wrote: Am 01.05.2010 um 11:48 schrieb William Kenworthy: On Sat, 2010-05-01 at 10:57 +0200, Kraus Philipp wrote: On 01.05.2010 um 10:32 wrote Volker Armin Hemmann: On Samstag 01 Mai 2010, Graham Murray wrote: Kraus Philipp writes: Hello, I must test a software with a older version of the glibc. I run the 2.11.1 now but for one tool I need a previous version (2.6.1). How can I compile the glibc without changing my system glibc. I would like to set the previous glibc with the LD_PATH. Can I run two different versions or is a better solution to downgrade the system glib? I think that the only way you can do this is to create a chroot jail, in which you build everything using the old version of glibc (in a very similar way to building a new Gentoo system) and run your application in that. no, you can install glibc in /usr/local and then tell apps to either use the libs in /usr/local or /usr. It is just not easy because it easily breaks stuff in horrrible to fix ways. Okay, can I downgrade my glibc? My Gentoo isn't a big system, it's a server installation, so I can recompile the whole system. I had forgotten to mask the glibc on the last update. I have add a line to the portage.mask but emerge says that it can't compile the older version, because will damage the system. Would LD_PRELOAD solve your problem? - worked for me when needing to run a legacy redhat app in the past on a more up-to-date gentoo system. I think that can solve my problem, because it's only this one lib all other libs work very well. There is also a LD_LIBRARY_PATH variable. Get a binary copy of the libs you need and put em somewhere convenient and let the rest of the system stay as is. I don't have the glibc binary. I can't emerge it and if I try to compile from the sources. The configure script says: These critical programs are missing or too old: as ld How I can compile the from the sources (http://ftp.gnu.org/gnu/glibc/glibc-2.10.1.tar.gz) ? You can use ebuild to build a package, but not install it - then just un-compress it and grab the wanted libs. I will try it on my system first, because I can lean something new. Or if you tell us what arch, someone (me if its compatible) can create a binary pkg for you using quickpkg. It might also be on the livecd/install medium as well - dont have one here to check. I run my Gentoo on amd64 (intel xeon), thank's for help.
Re: [gentoo-user] two glibcs with different version
On Saturday 01 May 2010 10:57:31 Kraus Philipp wrote: > On 01.05.2010 um 10:32 wrote Volker Armin Hemmann: > > On Samstag 01 Mai 2010, Graham Murray wrote: > >> Kraus Philipp writes: > >>> Hello, > >>> > >>> I must test a software with a older version of the glibc. I run the > >>> 2.11.1 now but for one tool I need a previous version (2.6.1). > >>> How can I compile the glibc without changing my system glibc. I > >>> would > >>> like to set the previous glibc with the LD_PATH. > >>> Can I run two different versions or is a better solution to > >>> downgrade > >>> the system glib? > >> > >> I think that the only way you can do this is to create a chroot jail, > >> in which you build everything using the old version of glibc (in a > >> very > >> similar way to building a new Gentoo system) and run your > >> application in > >> that. > > > > no, you can install glibc in /usr/local and then tell apps to either > > use the > > libs in /usr/local or /usr. > > > > It is just not easy because it easily breaks stuff in horrrible to > > fix ways. > > Okay, can I downgrade my glibc? My Gentoo isn't a big system, it's a > server > installation, so I can recompile the whole system. I had forgotten to > mask the > glibc on the last update. I have add a line to the portage.mask but > emerge says > that it can't compile the older version, because will damage the system. There is a way to downgrade, but it's far from bullet proof. You might end up with mutually incompatible code versions that makes the recompile world freak out. First, quickpkg your existing glibc Then read the glibc ebuild. You can bypass the version checks, IIRC its an envvar something like "I_AM_VERY_STUPID_OR_VERY_BRAVE". Failing that, just comment out the version checks, redigest the ebuild and run it. The rebuild world Obviously, this is not for the faint of heart. -- alan dot mckinnon at gmail dot com
Re: [gentoo-user] two glibcs with different version
Am 01.05.2010 um 12:53 schrieb pk: On 2010-04-30 15:24, Kraus Philipp wrote: I must test a software with a older version of the glibc. I run the 2.11.1 now but for one tool I need a previous version (2.6.1). How can I compile the glibc without changing my system glibc. I would like to set the previous glibc with the LD_PATH. Can I run two different versions or is a better solution to downgrade the system glib? I would suggest dual-booting... Sorry, that's no solution, because it is only one network service, that need's a older glibc. I can't reboot a server at any time or create a new physical server Phil
Re: [gentoo-user] Can't resolve package blocks
On Fri, 30 Apr 2010, Paul Hartman wrote: Use the --tree option in your emerge commandline so it will show which packages are trying to pull in these. I am using the -t flag - that's what I pasted earlier. It looks like you're using stable, I'm on ~amd64 and don't have these blocks so maybe it's something worked out by newer versions. This is a ppc machine... On the other hand, if you have unmasked any unstable packages maybe they're trying to pull in the conflictiong versions? I have practically nothing unmasked (just gnomeprint so unrelated). I have ffmpeg, libraw and libdc1394 unmerged, so its not clear what's pulling them in. You can see the "N" showing that they're new installs. --
Re: [gentoo-user] Can't resolve package blocks
On 4/30/10, Ajai Khattri wrote:
>
> I have unmerged ffmpeg, libraw1394 and libdc1394 and I still can't resolve
> this block:
>
> [nomerge ] media-video/ffmpeg-0.5_p20373 USE="X alsa amr encode
> hardcoded-tables ieee1394 ipv6 network theora threads vorbis x264 zlib
> (-3dnow) (-3dnowext) -altivec -bindist -cpudetection -custom-cflags -debug
> -dirac -doc -faac -faad -gsm -jack -jpeg2k (-mmx) (-mmxext) -mp3 -oss -pic
> -schroedinger -sdl -speex (-ssse3) -test -v4l -v4l2 -vdpau -xvid"
> VIDEO_CARDS="(-nvidia)" [0]
> [ebuild N] media-libs/opencore-amr-0.1.2 849 kB [0]
> [ebuild N] sys-libs/libraw1394-2.0.4 368 kB [0]
> [nomerge ] x11-libs/gtk+-2.18.7 [2.16.6] USE="cups jpeg tiff (-aqua)
> -debug -doc -jpeg2k -test -vim-syntax -xinerama (-X%*)" [0]
> [ebuild N] media-libs/tiff-3.9.2-r1 USE="cxx jpeg zlib -jbig" 1,387
> kB [0]
> [ebuild N] media-libs/jpeg-8a 951 kB [0]
> [blocks B ] is blocking sys-libs/libraw1394-2.0.4)
>
> Total: 79 packages (68 upgrades, 10 new, 1 in new slot, 6 uninstalls),
> Size of downloads: 174,892 kB
> Conflict: 13 blocks (1 unsatisfied)
> Portage tree and overlays:
> [0] /usr/portage
> [?] indicates that the source repository could not be determined
>
> * Error: The above package list contains packages which cannot be
> * installed at the same time on the same system.
>
>('ebuild', '/', 'media-libs/libdc1394-1.2.1', 'merge') pulled in by
> media-libs/libdc1394 required by ('ebuild', '/',
> 'media-video/ffmpeg-0.5_p20373', 'merge')
>
>('ebuild', '/', 'sys-libs/libraw1394-2.0.4', 'merge') pulled in by
> sys-libs/libraw1394 required by ('ebuild', '/',
> 'media-video/ffmpeg-0.5_p20373', 'merge')
> >=sys-libs/libraw1394-0.9.0 required by ('ebuild', '/',
> 'media-libs/libdc1394-1.2.1', 'merge')
This might sound slightly familiar:
http://bugs.gentoo.org/show_bug.cgi?id=305071#c3
I'd unmask the testing graded media-libs/libdc1394-2.1.2 in
package.keywords for the time being (or mask libraw1394-2.0.x), while
waiting for the stabilisation of newer libdc1394 versions. Which
unfortunately appears to be stuck:
http://bugs.gentoo.org/show_bug.cgi?id=315655
--
Arttu V.
Re: [gentoo-user] No /dev/sd? devices. Udev problem?
I wrote:
> I just migrated a friend's machine to ~am64. Everything was updated,
> but after a reboot some partitons are not found. No wonder, there are
> no /dev/sd? devices, only /dev/sg?. I suspect the problem is udev,
> because that was updated. Root is encrypted, so this machine makes use
> of an initramfs. At that point, all devices are found, so the system
> comes up, but later mounting of data partitions fails due to the
> missing devices.
>
> Any idea what the cause is? I will try to downgrade udev and see what
> happens. I have to wait for my friend to arrive here though, because I
> do not know the LUKS password. So I thought I'd ask here first, maybe
> someone knows this problem.
A downgrade from 151-r2 to 150-r1 did not change anything. I can try to go
to lower versions, but I wonder what the problem is. My own machine is
~x86 instead of ~am64, but has a similar setup, and all is working. Well,
not all, but I spare you my KDE4 rants for the moment.
The missing devices appear in the /sys/block/ hierarchy, so I can create
the device nodes by udevadm test. I created a new init script with
basically these commands, that activates all the devices:
for disk in /sys/block/sd*
do
udevadm test /block/${disk#/sys/block/}
done
for part in /sys/block/sd*/sd*
do
udevadm test /block/${part#/sys/block/}
done
Another thing I am missing is the /dev/vg/lvm entries, but I can also
access the LVM volumes as /dev/mapper/vg-lvm, so this is no problem. But I
wonder what else is missing that I do not know of yet. And I would prefer
a real solution over this hack, so if anyone has any ideas, I'd be happy
to hear them.
Wonko
[gentoo-user] weird portage warning
For the past few weeks, I have noticed the following bizarre msg : root:503 ~> emerge -Dup world These are the packages that would be merged, in order: Calculating dependencies | * Please fix your ebuild to not inherit the deprecated qt3.eclass * Please fix your ebuild to not inherit the deprecated qt3.eclass * Please fix your ebuild to not inherit the deprecated qt3.eclass ... done! Yes, I have the KDE 3 overlay installed, but even so how can a mere user be expected to "fix an ebuild", whatever that may mean anyway. Can anyone enlighten me as to the cause &/or purpose of these repeated msgs ? -- ,, SUPPORT ___//___, Philip Webb ELECTRIC /] [] [] [] [] []| Cities Centre, University of Toronto TRANSIT`-O--O---' purslowatchassdotutorontodotca
Re: [gentoo-user] weird portage warning
On Sonntag 02 Mai 2010, Philip Webb wrote: > For the past few weeks, I have noticed the following bizarre msg : > > root:503 ~> emerge -Dup world > These are the packages that would be merged, in order: > Calculating dependencies | > * Please fix your ebuild to not inherit the deprecated qt3.eclass > * Please fix your ebuild to not inherit the deprecated qt3.eclass > * Please fix your ebuild to not inherit the deprecated qt3.eclass > ... done! > > Yes, I have the KDE 3 overlay installed, > but even so how can a mere user be expected to "fix an ebuild", > whatever that may mean anyway. Can anyone enlighten me > as to the cause &/or purpose of these repeated msgs ? the cause: the qt3.eclass is deprecated (because qt3 and everything depending on it is 'dead'). The ebuilds use this eclass, portage warns about it. There is nothing you can do about it. But hopefully you have the eclass into your overlay too...
