View previous topic :: View next topic |
Author |
Message |
cynric Guru
Joined: 08 Oct 2004 Posts: 439
|
Posted: Sun Apr 02, 2006 6:08 am Post subject: Growing number of emerge fails |
|
|
Is it just me or is there an increasing number of emerge failures? I've been running ~amd64 for about 1.5 years I suppose. During that time, I've had very few problems and nothing major. Mostly these were architecture problems, config errors, or other general stuff. But, I have been getting my share of bad emerges for about the last 2 maybe 3 months. It seems to be getting progressively worse. I'm not asking about a flame war or a lecture on running ~arch. I'm merely questioning if something has been changed lately that makes portage or gcc-4.0 more sensitive to problems and if others are seeing an increase in problems as well. _________________ "This Snow Crash thing -- is it a virus, a drug, or a religion?" "What's the difference?"
-- Neal Stephenson - Snow Crash |
|
Back to top |
|
|
kill[h]er n00b
Joined: 02 Sep 2003 Posts: 30
|
Posted: Sun Apr 02, 2006 6:42 am Post subject: |
|
|
seeing it here too in ~x86... more than ever before... many of the things i've encountered have bug reports filed and some have patches in the reports, but they aren't in portage yet. |
|
Back to top |
|
|
nixnut Bodhisattva
Joined: 09 Apr 2004 Posts: 10974 Location: the dutch mountains
|
Posted: Sun Apr 02, 2006 6:44 am Post subject: |
|
|
As far as I can see gcc-4.0.x is not in ~arch for any arch except itanium based systems and gcc-4.1 is hard masked. So if you are using gcc-4.x you are using an unsupported compiler and that is totally unrelated to ~arch. _________________ Please add [solved] to the initial post's subject line if you feel your problem is resolved. Help answer the unanswered
talk is cheap. supply exceeds demand |
|
Back to top |
|
|
cynric Guru
Joined: 08 Oct 2004 Posts: 439
|
Posted: Sun Apr 02, 2006 6:48 am Post subject: |
|
|
Most of the stuff does seem to be mentioned in b.g.o prior to or shortly after my hitting them. I assume that part of this is in regards to a lot of revving up in QA -- something I've heard mentioned and see partially in the increased number of e[warn|info] messages. It could simply be that there aren't enough people to review all the user input. While this is debateable, there still exists the problem with broken stuff being pushed out to the servers. But, it is semi-encouraging that I'm not the only one. It's disappointing because it does mean I'm not the only one. _________________ "This Snow Crash thing -- is it a virus, a drug, or a religion?" "What's the difference?"
-- Neal Stephenson - Snow Crash |
|
Back to top |
|
|
cynric Guru
Joined: 08 Oct 2004 Posts: 439
|
Posted: Sun Apr 02, 2006 6:50 am Post subject: |
|
|
nixnut wrote: | 8O As far as I can see gcc-4.0.x is not in ~arch for any arch except itanium based systems and gcc-4.1 is hard masked. So if you are using gcc-4.x you are using an unsupported compiler and that is totally unrelated to ~arch. |
That is correct that it isn't in ~arch, but that does not dismiss my observation about an increase in broken builds. I also never said it was related to ~arch. It was only mentioned because it /wasn't/ related, but something I felt noteworthy nonetheless. _________________ "This Snow Crash thing -- is it a virus, a drug, or a religion?" "What's the difference?"
-- Neal Stephenson - Snow Crash |
|
Back to top |
|
|
nixnut Bodhisattva
Joined: 09 Apr 2004 Posts: 10974 Location: the dutch mountains
|
Posted: Sun Apr 02, 2006 6:56 am Post subject: |
|
|
cynric wrote: | That is correct that it isn't in ~arch, but that does not dismiss my observation about an increase in broken builds. I also never said it was related to ~arch. It was only mentioned because it /wasn't/ related, but something I felt noteworthy nonetheless. | Nor did you I say said so (and no you didn't say I said you said so etc etc). Anyway, my guess is that gcc-4.0 is causing a lot of the breakage you see. gcc-4.1 is much better in that regard, though a number of packages still need to be made 4.1 ready. I think gcc-4.1 might show up in ~arch in the near future since it's looking pretty good already. _________________ Please add [solved] to the initial post's subject line if you feel your problem is resolved. Help answer the unanswered
talk is cheap. supply exceeds demand |
|
Back to top |
|
|
cynric Guru
Joined: 08 Oct 2004 Posts: 439
|
Posted: Sun Apr 02, 2006 7:03 am Post subject: |
|
|
nixnut wrote: | Nor did you I say said so (and no you didn't say I said you said so etc etc). Anyway, my guess is that gcc-4.0 is causing a lot of the breakage you see. gcc-4.1 is much better in that regard, though a number of packages still need to be made 4.1 ready. I think gcc-4.1 might show up in ~arch in the near future since it's looking pretty good already. |
Thanks for the reply. Like you, I didn't intend to sound like I was on the offensive. GCC is the only thing I'm running out of p.mask so if stability is better in the 4.1 branch then it might be wise to move to that. I still maintain a 3.4.5 copy in reserve. The number of supported packages isn't a huge concern with me as I run a pretty bare desktop. Although gcc-4.0.x could be the source of much anguish, I'm not totally sold on that. Unfortunately, I have been a bit busy of late and haven't really followed up with any observed errors -- just been waiting the couple of days until fixes are rolled out. Upgrading GCC is definitely something I'll look into though. Thanks again. _________________ "This Snow Crash thing -- is it a virus, a drug, or a religion?" "What's the difference?"
-- Neal Stephenson - Snow Crash |
|
Back to top |
|
|
aidanjt Veteran
Joined: 20 Feb 2005 Posts: 1118 Location: Rep. of Ireland
|
Posted: Sun Apr 02, 2006 8:12 am Post subject: |
|
|
I'm getting occasional intermitent build failures on x86 much less ~x86. |
|
Back to top |
|
|
ercxy n00b
Joined: 26 Apr 2004 Posts: 55 Location: MA
|
Posted: Wed Apr 05, 2006 10:03 pm Post subject: |
|
|
AidanJT wrote: | I'm getting occasional intermitent build failures on x86 much less ~x86. |
I have x86 and I have some emerge failures too. And some packages compile but don't function.
(anjuta; stable(x86) compiles, crashes every single time at the start. ~x86 runs but importing project crashes the program).
I don't have aggresive *FLAGS by the way.
The only shaky stuff is I have 64 bit capable celeron, and set -march= nocona.
I did not want to go for amd64(stability issue), but if there is as much as random failures in am64 compared to x86, I may switch soon .
Portage 2.0.54 (default-linux/x86/2006.0, gcc-3.4.5, glibc-2.3.5-r3, 2.6.15-gentoo-r1 i686)
=================================================================
System uname: 2.6.15-gentoo-r1 i686 Intel(R) Celeron(R) CPU 3.06GHz
Gentoo Base System version 1.6.14
ccache version 2.3 [enabled]
dev-lang/python: 2.4.2
sys-apps/sandbox: 1.2.12
sys-devel/autoconf: 2.13, 2.59-r7
sys-devel/automake: 1.4_p6, 1.5, 1.6.3, 1.7.9-r1, 1.8.5-r3, 1.9.6-r1
sys-devel/binutils: 2.16.1
sys-devel/libtool: 1.5.22
virtual/os-headers: 2.6.11-r2
ACCEPT_KEYWORDS="x86"
AUTOCLEAN="yes"
CBUILD="i686-pc-linux-gnu"
CFLAGS="-march=nocona -O2 -pipe"
CHOST="i686-pc-linux-gnu"
CONFIG_PROTECT="/etc /usr/kde/2/share/config /usr/kde/3/share/config /usr/lib/X11/xkb /usr/share/config /usr/share/texmf/dvipdfm/config/ /usr/share/texmf/dvips/config/ /usr/share/texmf/tex/generic/config/ /usr/share/texmf/tex/platex/config/ /usr/share/texmf/xdvi/ /var/qmail/control"
CONFIG_PROTECT_MASK="/etc/eselect/compiler /etc/gconf /etc/terminfo /etc/env.d"
CXXFLAGS="-march=nocona -O2 -pipe"
DISTDIR="/usr/portage/distfiles"
FEATURES="autoconfig ccache distlocks sandbox sfperms strict"
MAKEOPTS="-j2"
PKGDIR="/usr/portage/packages"
PORTAGE_TMPDIR="/var/tmp"
PORTDIR="/usr/portage"
USE="x86 X acpi alsa apache2 asf audiofile avi bash-completion berkdb bitmap-fonts bzip2 cdr cli crypt ctype cups curl dba dbus dri dvd dvi eds emboss encode esd exif expat fam fastbuild fftw firefox foomaticdb force-cgi-redirect fortran ftp gd gdbm gif glut gnome gpm gstreamer gtk gtk2 gtkhtml hal howl imlib ipv6 isdnlog jpeg lcms libg++ libwww mad memlimit mikmod mmx mng motif mp3 mpeg mpi nautilus ncurses nls nptl ogg opengl oss pam pcre pdf pdflib perl plotutils png posix pppd python quicktime readline sdl session simplexml soap sockets spell spl sse sse2 ssl svg svga tcltk tcpd tetex tiff tokenizer truetype truetype-fonts trutype-fonts type1-fonts udev usb vorbis win32codecs wmf xml xml2 xsl xv zlib userland_GNU kernel_linux elibc_glibc"
Unset: ASFLAGS, CTARGET, INSTALL_MASK, LANG, LC_ALL, LDFLAGS, LINGUAS, PORTDIR_OVERLAY _________________ PC (Personal Cluster)
2 Athlon X2 3800,1 Celeron D 3GHZ, 1 Athlon xp 2400, 3 Athlon xp 1900, 6 Athlon xp 1800
OS:Rocks Cluster |
|
Back to top |
|
|
cynric Guru
Joined: 08 Oct 2004 Posts: 439
|
Posted: Wed Apr 05, 2006 10:22 pm Post subject: |
|
|
I think the problems are coming from the builds/patches themselves and not really related to unstable systems as such. I wen ahead and upgraded to gcc-4.1 though. Haven't used it enough to tell any differences, but after I worked out the initial build problems the new `emerge -ea world` didn't have a bit of problem. _________________ "This Snow Crash thing -- is it a virus, a drug, or a religion?" "What's the difference?"
-- Neal Stephenson - Snow Crash |
|
Back to top |
|
|
curmudgeon Veteran
Joined: 08 Aug 2003 Posts: 1740
|
Posted: Mon May 15, 2006 2:07 pm Post subject: |
|
|
Certainly an increase in failed packages on x86.
To cite three, just in the last week or so:
1. portage endlessly looping with the -e option (marked stable then not masked)
2. some package with a bad digest stopping an emerge -e world
3. some package with a non-existent file listed in the manifest (a ".orig" file), which also killed emerge -e world |
|
Back to top |
|
|
broken_chaos Guru
Joined: 18 Jan 2006 Posts: 370 Location: Ontario, Canada
|
Posted: Mon May 15, 2006 7:18 pm Post subject: |
|
|
curmudgeon wrote: | 1. portage endlessly looping with the -e option (marked stable then not masked)
2. some package with a bad digest stopping an emerge -e world
3. some package with a non-existent file listed in the manifest (a ".orig" file), which also killed emerge -e world |
1. Fixed with the latest stable upgrade to portage. Could have unmasked this version if you needed to to an emerge -e before it was stable.
2. Bad sync? Re-sync and emerge --resume should do the trick, I think.
3. Re-digest it until the problem is fixed in the portage tree.
Personally I've run into few, if any, major errors using the stable portage tree on x86. Relatively (but not extremely) conservative CFLAGs, liberal use of package.keywords where needed, and a basic set of 'world' USE flags, the rest set on a per-package basis (what I need and no more). |
|
Back to top |
|
|
curmudgeon Veteran
Joined: 08 Aug 2003 Posts: 1740
|
Posted: Thu May 18, 2006 3:41 pm Post subject: |
|
|
broken_chaos wrote: |
1. Fixed with the latest stable upgrade to portage. Could have unmasked this version if you needed to to an emerge -e before it was stable.
2. Bad sync? Re-sync and emerge --resume should do the trick, I think.
3. Re-digest it until the problem is fixed in the portage tree. |
I personally think the developers should have pulled 1 (forcing a downgrade if necessary). I just don't expect to run into problems like 3. That represents bad
QA. I almost always want to avoid syncing in the middle of emerge -e. It could
introduce unwanted major changes, and it makes resuming impossible.
I forgot 4. slang wouldn't build because of a sandbox problem. Each of these
took time to research and overcome, and combined, they made a usually
routine emerge -e a complete hell. It just seems to me that the number of
problems like these has increased significantly recently. |
|
Back to top |
|
|
Matteo Azzali Retired Dev
Joined: 23 Sep 2004 Posts: 1133
|
Posted: Fri May 19, 2006 9:19 am Post subject: |
|
|
I think the cause is lack of developers/testers.
In portage there are still elder packages unmantained (upstream) by years, and their cohabitation with
newer packages is becoming more and more difficult as the time goes and innovations come out.
This is not a critic to developers but to users, wanting good software but not being disposed to became mantainers/helpers,
here is a list of "by far unmaintained packages" :
xmltv (in portage from 5.34 to 5.39, most or all of them are useless, actually upstream is at 5.43 by far)
avidemux (in portage latest unmasked unstable is 2.0.42-r1 , upstream marked stable 2.1 and is previewing 2.2)
kover (unmantained by years, no latest stable upstream but a whole new app called koverartist)
gtkglextmm (1.0.1 latest unmasked in portage, actually upstream is at 1.2)
fox (1.4.x latest unmasked, upstream is at 1.6.x)
and there's more, these are just some examples I found...... _________________ Every day a new distro comes to birth. Every day a distro "eats" another.
If you're born distro, no matter what, start to run.
---- http://www.linuxprinting.org/ ---- http://tuxmobil.org/ |
|
Back to top |
|
|
pigeon768 l33t
Joined: 02 Jan 2006 Posts: 683
|
Posted: Sun May 21, 2006 1:04 am Post subject: |
|
|
curmudgeon wrote: | 2. some package with a bad digest stopping an emerge -e world
3. some package with a non-existent file listed in the manifest (a ".orig" file), which also killed emerge -e world | Use emwrap.sh to avoid one package not compiling killing an emerge -e world. Do a search for it. |
|
Back to top |
|
|
curmudgeon Veteran
Joined: 08 Aug 2003 Posts: 1740
|
Posted: Fri May 26, 2006 10:58 am Post subject: |
|
|
Matteo Azzali wrote: | I think the cause is lack of developers/testers.
In portage there are still elder packages unmantained (upstream) by years, and their cohabitation with
newer packages is becoming more and more difficult as the time goes and innovations come out.
This is not a critic to developers but to users, wanting good software but not being disposed to became mantainers/helpers,
here is a list of "by far unmaintained packages" : |
I don't know if that affected any of the packages that I ran into.
But you still make a good point. I would like to help (other than filing bug reports :) ), but I don't have enough programming skills. I can write some simple scripts, but the complexities and subtleties of portage baffle me. |
|
Back to top |
|
|
curmudgeon Veteran
Joined: 08 Aug 2003 Posts: 1740
|
Posted: Fri Jun 30, 2006 7:49 am Post subject: |
|
|
This seems like a good time to add:
5. marking binutils-2.16.1-r3 stable with a required file (binutils-2.16.1-patches-1.10.tar.bz2) not on the mirrors |
|
Back to top |
|
|
Genjix Apprentice
Joined: 23 Jun 2005 Posts: 163
|
Posted: Fri Jun 30, 2006 7:19 pm Post subject: |
|
|
Yup, Have to agree with you there- I've had to file quite a few bugs myself lately. Specifically gtkglextmm is the latest. |
|
Back to top |
|
|
|