Re: Fedora rawhide FTBFS status 2010-06-01 x86_64
Am Donnerstag, den 03.06.2010, 09:42 -0500 schrieb Matt Domsch: julian: diveintopython,sap I have already fixed sap, but I'm not able to do anything about diveintopython. The build process simply is too old to work with newer versions of the required libraries, and already worked only with patches before. I have already orphaned diveintopython about two weeks ago, any help/takeover would be greatly appreciated. Regards, Julian signature.asc Description: This is a digitally signed message part -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Fedora rawhide FTBFS status 2010-06-01 x86_64
On Thu, Jun 3, 2010 at 10:42 AM, Matt Domsch wrote: Fedora Fails To Build From Source Results for x86_64 using rawhide from 2010-06-01 [cut] Question 1: Suppose package A fails to build from source due to a bug in package B that is listed in package A's BuildRequires. Now that package B gets fixed, and it is possible to build package A from source without any modification. Do we need to bump A's release and do the rebuild in this case? Question 2: What is the likelihood of a mass rebuild in this cycle? Orcan -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Fedora rawhide FTBFS status 2010-06-01 x86_64
On Sun, Jun 20, 2010 at 10:24:27PM -0400, Orcan Ogetbil wrote: On Thu, Jun 3, 2010 at 10:42 AM, Matt Domsch wrote: Fedora Fails To Build From Source Results for x86_64 using rawhide from 2010-06-01 [cut] Question 1: Suppose package A fails to build from source due to a bug in package B that is listed in package A's BuildRequires. Now that package B gets fixed, and it is possible to build package A from source without any modification. Do we need to bump A's release and do the rebuild in this case? No. List in bug A that it Depends on the bug number for B. Close the bug against A once package B is fixed and its own bug closed. Question 2: What is the likelihood of a mass rebuild in this cycle? I've heard rumors that it's likely, but I don't know for certain what the trigger would be this time. Likely a glibc or gcc change. -- Matt Domsch Technology Strategist Dell | Office of the CTO -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Fedora rawhide FTBFS status 2010-06-01 x86_64
On Thu, Jun 3, 2010 at 4:42 PM, Matt Domsch matt_dom...@dell.com wrote: It also doesn't report any failing packages that have subsequently been built and published in koji's rawhide since 06-01. That should cut down on the but I just fixed that! responses from now on. One question. Is committing the fix in CVS enough or do I need to also push an update? In my case the package missed a BR that was only used to do tests in the %check section, I removed it now in CVS but the resulting binary package is basically the same as before. -- Gianluca Sforna http://morefedora.blogspot.com http://www.linkedin.com/in/gianlucasforna -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Fedora rawhide FTBFS status 2010-06-01 x86_64
Hedayat Vatankhah wrote, at 06/04/2010 03:07 PM +9:00: Hi, My packages (rcsslogplayer and rcssmonitor) are being failed because of the lack of Qt dependencies (QtNetwork requires ssl and crypto libraries from openssl-devel package, and gobject-2.0 from glib2-devel). As these packages are required by Qt, I think they should be added to qt-devel dependencies. Should I fill a bug against Qt? Thanks, Hedayat For now I just checked rcssmonitor and the cause for build failure is that - While rcssmonitor explicitly uses Libs.private (from m4/qt.m4 in your source tarball) like --- $ pkg-config --static --libs-only-l QtNetwork -lQtNetwork -lssl -lcrypto -lQtCore -lpthread -lz -lm -ldl -lgthread-2.0 -lrt -lglib-2.0 --- qt(4)-devel does pull openssl-devel or so in. However this is Libs.private dependency and man pkg-config says: --- Libs.private: This line should list any private libraries in use. Private libraries are libraries which are not exposed through your library, but are needed in the case of static linking. This differs from Requires.private in that it refer‐ ences libraries that do not have package files installed. --- So as we don't do static linking, this should not be needed. I tried to add --- sed -i.nostatic \ -e 's|--static||g' m4/qt.m4 --- and this seems to make build succeed (I just also tried the same fix for rcsslogplayer and it succeeds) http://koji.fedoraproject.org/koji/taskinfo?taskID=2229111 http://koji.fedoraproject.org/koji/taskinfo?taskID=2229126 Regards, Mamoru -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Fedora rawhide FTBFS status 2010-06-01 x86_64
/*Mamoru Tasaka mtas...@ioa.s.u-tokyo.ac.jp*/ wrote on 06/04/2010 12:00:50 PM +0450: Hedayat Vatankhah wrote, at 06/04/2010 03:07 PM +9:00: Hi, My packages (rcsslogplayer and rcssmonitor) are being failed because of the lack of Qt dependencies (QtNetwork requires ssl and crypto libraries from openssl-devel package, and gobject-2.0 from glib2-devel). As these packages are required by Qt, I think they should be added to qt-devel dependencies. Should I fill a bug against Qt? Thanks, Hedayat For now I just checked rcssmonitor and the cause for build failure is that - While rcssmonitor explicitly uses Libs.private (from m4/qt.m4 in your source tarball) like --- $ pkg-config --static --libs-only-l QtNetwork -lQtNetwork -lssl -lcrypto -lQtCore -lpthread -lz -lm -ldl -lgthread-2.0 -lrt -lglib-2.0 --- qt(4)-devel does pull openssl-devel or so in. However this is Libs.private dependency and man pkg-config says: --- Libs.private: This line should list any private libraries in use. Private libraries are libraries which are not exposed through your library, but are needed in the case of static linking. This differs from Requires.private in that it refer‐ ences libraries that do not have package files installed. --- So as we don't do static linking, this should not be needed. I tried to add --- sed -i.nostatic \ -e 's|--static||g' m4/qt.m4 --- and this seems to make build succeed (I just also tried the same fix for rcsslogplayer and it succeeds) Thanks, I'll try it Good luck, Hedayat http://koji.fedoraproject.org/koji/taskinfo?taskID=2229111 http://koji.fedoraproject.org/koji/taskinfo?taskID=2229126 Regards, Mamoru -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Fedora rawhide FTBFS status 2010-06-01 x86_64
On Fri, Jun 04, 2010 at 08:57:06AM +0200, Gianluca Sforna wrote: On Thu, Jun 3, 2010 at 4:42 PM, Matt Domsch matt_dom...@dell.com wrote: It also doesn't report any failing packages that have subsequently been built and published in koji's rawhide since 06-01. ??That should cut down on the but I just fixed that! ??responses from now on. One question. Is committing the fix in CVS enough or do I need to also push an update? In my case the package missed a BR that was only used to do tests in the %check section, I removed it now in CVS but the resulting binary package is basically the same as before. Given that we'll likely have a mass rebuild during the F14 dev cycle, you can just fix it in CVS and wait for the mass rebuild to generate the new pacakge. But please leave the bug open, change to MODIFIED status, until the package is rebuilt. Otherwise, I'll wind up filing a duplicate bug. I'm not inclined to try to teach my scripts to pull from CVS directly. It's been really nice only dealing with published SRPMS, even though there is some latency involved (packages fixed in CVS this morning won't appear until tomorrow). Thanks, Matt -- Matt Domsch Technology Strategist Dell | Office of the CTO -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Fedora rawhide FTBFS status 2010-06-01 x86_64
Mamoru Tasaka wrote: I tried to add --- sed -i.nostatic \ -e 's|--static||g' m4/qt.m4 --- and this seems to make build succeed (I just also tried the same fix for rcsslogplayer and it succeeds) Yes, --static is clearly wrong for dynamic linking. Kevin Kofler -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Fedora rawhide FTBFS status 2010-06-01 x86_64
On 3 June 2010 20:12, Matt Domsch wrote: Fedora Fails To Build From Source Results for x86_64 using rawhide from 2010-06-01 Small enhancement request for your scripts: I have two packages in list (one maintainer and another co-maintainer) and both are mentioned at the bottom, but I get mail twice. Can you update it to send mail once in case it is easy todo and you consider it worth. Thanks (specially for keeping good work going :) , -- Rakesh Pandit https://fedoraproject.org/wiki/User:Rakesh freedom, friends, features, first -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Fedora rawhide FTBFS status 2010-06-01 x86_64
On 3 June 2010 21:29, Rakesh Pandit wrote: On 3 June 2010 20:12, Matt Domsch wrote: Fedora Fails To Build From Source Results for x86_64 using rawhide from 2010-06-01 Small enhancement request for your scripts: I have two packages in list (one maintainer and another co-maintainer) and both are mentioned at the bottom, but I get mail twice. Can you update it to send mail once in case it is easy todo and you consider it worth. Thanks (specially for keeping good work going :) , Failed to notice, these are for two separate archs, so it is ok. thanks. -- Rakesh Pandit https://fedoraproject.org/wiki/User:Rakesh freedom, friends, features, first -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
Re: Fedora rawhide FTBFS status 2010-06-01 x86_64
On Thu, Jun 03, 2010 at 09:42:04AM -0500, Matt Domsch wrote: Fedora Fails To Build From Source Results for x86_64 using rawhide from 2010-06-01 This run continues from the previous run, rebuilding those packages that failed during the earlier run, or that changed between 2010-05-27 and 06-01. I filed a ton of bugzillas, basically this list. I apologize if there are some duplicates to already-existing FTBFS bugs opened for earlier releases - that wasn't intentional, but please take this opportunity to fix the problem and close both bugs then. Thanks, Matt -- Matt Domsch Technology Strategist Dell | Office of the CTO -- devel mailing list devel@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/devel
