I've created a script[1] to find the latest failed build that hasn't yet been fixed during our mass rebuild. The output is an html page broken down by packager. Each package listed under the packager is a link to the latest failed build attempt. Shortly I'll setup a cron job to keep this list updated every 10 minutes or so.
http://jkeating.fedorapeople.org/failed-f11-rebuilds.html
If you have a package on this list, please fix the build error and do a normal build (make build). Your package will go into dist-f11 and this script will see that it has been fixed and no longer report it.
Since builds are still ongoing, more failures may be added to the list. Please check back often.
[1]: http://git.fedorahosted.org/git/?p=releng;a=blob;f=scripts/find-failures.py;...
Jesse Keating wrote:
I've created a script[1] to find the latest failed build that hasn't yet been fixed during our mass rebuild. The output is an html page broken down by packager. Each package listed under the packager is a link to the latest failed build attempt. Shortly I'll setup a cron job to keep this list updated every 10 minutes or so.
http://jkeating.fedorapeople.org/failed-f11-rebuilds.html
If you have a package on this list, please fix the build error and do a normal build (make build). Your package will go into dist-f11 and this script will see that it has been fixed and no longer report it.
Since builds are still ongoing, more failures may be added to the list. Please check back often.
If anyone wants to help fix any of my broken packages, I'd appreciate it. It's going to be a few days until I can get mock builds for rawhide working again due to the recent rpm changes.
--Wart
On Thu 26-Feb-2009 at 14:21 -0800, Michael Thomas wrote:
If anyone wants to help fix any of my broken packages, I'd appreciate it.
Same here, I'm not going to get a chance to fix hugin for at least a week:
http://koji.fedoraproject.org/koji/taskinfo?taskID=1166227
Jesse Keating wrote:
I've created a script[1] to find the latest failed build that hasn't yet been fixed during our mass rebuild.
http://jkeating.fedorapeople.org/failed-f11-rebuilds.html
If you have a package on this list, please fix the build error and do a normal build (make build). Your package will go into dist-f11 and this script will see that it has been fixed and no longer report it.
jima (1):
> graphviz http://koji.fedoraproject.org/koji/taskinfo?taskID=1164446 I suspect Patrick "Jima" Laughton must be out on a sabbatical, or something?
I'm one of the upstream authors of graphviz. I've posted to Bugzilla some patches that I think should fix the mass rebuild failure. https://bugzilla.redhat.com/show_bug.cgi?id=486045
Could someone else apply them and kick off a rebuild?
On 2009-02-27 at 7:32:17 -0500, John Ellson john.ellson@comcast.net wrote:
Jesse Keating wrote:
I've created a script[1] to find the latest failed build that hasn't yet been fixed during our mass rebuild. http://jkeating.fedorapeople.org/failed-f11-rebuilds.html
If you have a package on this list, please fix the build error and do a normal build (make build). Your package will go into dist-f11 and this script will see that it has been fixed and no longer report it.
jima (1):
> graphviz
http://koji.fedoraproject.org/koji/taskinfo?taskID=1164446 I suspect Patrick "Jima" Laughton must be out on a sabbatical, or something?
I'm one of the upstream authors of graphviz. I've posted to Bugzilla some patches that I think should fix the mass rebuild failure. https://bugzilla.redhat.com/show_bug.cgi?id=486045
Could someone else apply them and kick off a rebuild?
Yep, I've got this.
~spot
Tom "spot" Callaway wrote:
On 2009-02-27 at 7:32:17 -0500, John Ellson john.ellson@comcast.net wrote:
Jesse Keating wrote:
I've created a script[1] to find the latest failed build that hasn't yet been fixed during our mass rebuild. http://jkeating.fedorapeople.org/failed-f11-rebuilds.html
If you have a package on this list, please fix the build error and do a normal build (make build). Your package will go into dist-f11 and this script will see that it has been fixed and no longer report it.
jima (1):
> graphviz
http://koji.fedoraproject.org/koji/taskinfo?taskID=1164446 I suspect Patrick "Jima" Laughton must be out on a sabbatical, or something?
I'm one of the upstream authors of graphviz. I've posted to Bugzilla some patches that I think should fix the mass rebuild failure. https://bugzilla.redhat.com/show_bug.cgi?id=486045
Could someone else apply them and kick off a rebuild?
Yep, I've got this.
~spot
Thanks Spot,
I see there is now a problem with BR for java-devel. Are you Cc'd on Bug #486045 ? Can I just correspond with you through there to get the graphviz rebuild fixed?
On 2009-02-27 at 13:19:41 -0500, John Ellson john.ellson@comcast.net wrote:
I see there is now a problem with BR for java-devel. Are you Cc'd on Bug #486045 ? Can I just correspond with you through there to get the graphviz rebuild fixed?
Sure, that works.
~spot
On Thu, 26 Feb 2009 12:20:54 -0800, "JK" == Jesse Keating jkeating@redhat.com wrote:
JK> I've created a script[1] to find the latest failed build that hasn't yet JK> been fixed during our mass rebuild. The output is an html page broken JK> down by packager. Each package listed under the packager is a link to JK> the latest failed build attempt. Shortly I'll setup a cron job to keep JK> this list updated every 10 minutes or so.
JK> http://jkeating.fedorapeople.org/failed-f11-rebuilds.html
for kakasi,
fail only happened on ppc64, which looks strange to me:
http://koji.fedoraproject.org/koji/getfile?taskID=1169050&name=build.log /bin/sh ../libtool --mode=link gcc -O2 -g -pipe -Wall -Wp,-D_FORTIFY_SOURCE=2 -fexceptions -fstack-protector --param=ssp-buffer-size=4 -m64 -mminimal-toc -Wall -Wunused -Wuninitialized -Wmissing-prototypes -Wmissing-declarations -pedantic -o libkakasi.la -rpath /usr/lib64 -version-info 3:0:1 -export-dynamic libdict.lo libkakasi.lo libkanjiio.lo liba2.lo libg2.lo libj2.lo libk2.lo libee2.lo libhh2.lo libjj2.lo libkk2.lo libitaiji.lo lib78_83.lo mkdir .libs rm -fr .libs/libkakasi.la .libs/libkakasi.* .libs/libkakasi.* (cd . && ln -s libdict.lo libdict.o) (cd . && ln -s libkakasi.lo libkakasi.o) (cd . && ln -s libkanjiio.lo libkanjiio.o) (cd . && ln -s liba2.lo liba2.o) (cd . && ln -s libg2.lo libg2.o) (cd . && ln -s libj2.lo libj2.o) (cd . && ln -s libk2.lo libk2.o) (cd . && ln -s libee2.lo libee2.o) (cd . && ln -s libhh2.lo libhh2.o) (cd . && ln -s libjj2.lo libjj2.o) (cd . && ln -s libkk2.lo libkk2.o) (cd . && ln -s libitaiji.lo libitaiji.o) (cd . && ln -s lib78_83.lo lib78_83.o) gcc -shared libdict.lo libkakasi.lo libkanjiio.lo liba2.lo libg2.lo libj2.lo libk2.lo libee2.lo libhh2.lo libjj2.lo libkk2.lo libitaiji.lo lib78_83.lo -Wl,-soname -Wl, -o .libs/ /usr/bin/ld: cannot open output file .libs/: Is a directory collect2: ld returned 1 exit status make[2]: *** [libkakasi.la] Error 1
This application is quite old and just suspected any files generated by autotools might affects. then tried to regenerate with libtoolize, aclocal, automake and autoconf.
However still no lucks. the result is worst:
https://koji.fedoraproject.org/koji/getfile?taskID=1204738&name=build.lo... checking build system type... Invalid configuration `ppc64-redhat-linux-gnu': machine `ppc64-redhat' not recognized configure: error: /bin/sh ./config.sub ppc64-redhat-linux-gnu failed
So what am I missing? either of the above didn't happen on other arches though.
Any idea?
-- Akira TAGOH
Akira TAGOH wrote, at 02/27/2009 10:22 PM +9:00:
On Thu, 26 Feb 2009 12:20:54 -0800, "JK" == Jesse Keating jkeating@redhat.com wrote:
JK> I've created a script[1] to find the latest failed build that hasn't yet JK> been fixed during our mass rebuild. The output is an html page broken JK> down by packager. Each package listed under the packager is a link to JK> the latest failed build attempt. Shortly I'll setup a cron job to keep JK> this list updated every 10 minutes or so.
JK> http://jkeating.fedorapeople.org/failed-f11-rebuilds.html
for kakasi,
fail only happened on ppc64, which looks strange to me: http://koji.fedoraproject.org/koji/getfile?taskID=1169050&name=build.log
This application is quite old and just suspected any files generated by autotools might affects. then tried to regenerate with libtoolize, aclocal, automake and autoconf. However still no lucks. the result is worst:
So what am I missing? either of the above didn't happen on other arches though.
Any idea?
Overwriting config.{guess,sub} like: -------------------------------------------------------- %prep %setup -q %patch1 -p1 %patch2 -p1 %patch3 -p1 %patch4 -p1 cp -p /usr/lib/rpm/config.{guess,sub} .
%build %configure --disable-static make %{?_smp_mflags} -------------------------------------------------------- seems good: http://koji.fedoraproject.org/koji/taskinfo?taskID=1204932
Mamoru
On Fri, 27 Feb 2009 23:03:06 +0900, "MT" == Mamoru Tasaka mtasaka@ioa.s.u-tokyo.ac.jp wrote:
MT> Overwriting config.{guess,sub} like: MT> -------------------------------------------------------- MT> %prep MT> %setup -q MT> %patch1 -p1 MT> %patch2 -p1 MT> %patch3 -p1 MT> %patch4 -p1 MT> cp -p /usr/lib/rpm/config.{guess,sub} .
MT> %build MT> %configure --disable-static MT> make %{?_smp_mflags} MT> -------------------------------------------------------- MT> seems good: MT> http://koji.fedoraproject.org/koji/taskinfo?taskID=1204932
Great! Thanks!
-- Akira TAGOH
MT> Mamoru
Akira TAGOH wrote:
On Fri, 27 Feb 2009 23:03:06 +0900, "MT" == Mamoru Tasaka mtasaka@ioa.s.u-tokyo.ac.jp wrote:
MT> Overwriting config.{guess,sub} like: MT> -------------------------------------------------------- MT> %prep MT> %setup -q MT> %patch1 -p1 MT> %patch2 -p1 MT> %patch3 -p1 MT> %patch4 -p1 MT> cp -p /usr/lib/rpm/config.{guess,sub} .
<nitpicking> I wouldn't want to do this, but would add a patch containing appropriate upstream versions to your package, instead of making your package vulnerable to using a random (and not unlikely to become) outdated version from rpm, which is not unlikely to go away at any time. </nitpicking>
[Upstream for config.guess|config.sub is the "config"-project: cf. http://savannah.gnu.org]
Ralf
Ralf Corsepius rc040203@freenet.de writes:
Akira TAGOH wrote:
MT> cp -p /usr/lib/rpm/config.{guess,sub} .
<nitpicking> I wouldn't want to do this,
I agree with that, but IMHO the approved way is like this: rm -f config.guess config.sub automake --add-missing
regards, tom lane