Re: [ClusterLabs] [Announce] libqb 10.rc3 release
On 26/02/16 08:58 +0100, Jan Pokorný wrote: > On 26/02/16 14:27 +0900, Keisuke MORI wrote: >> As of libqb-1.0rc3, Pacemaker fails to build upon it with the gcc >> warnings as below. >> There was no such a problem until 1.0rc2, and it seems that the >> changes in the pull request #175 is related. >> >> https://github.com/ClusterLabs/libqb/pull/175 >> >> {{{ >> [root@build-centos71 pacemaker ((Pacemaker-1.1.14))]# rpm -qa | grep libqb >> libqb-1.0rc3-1.el7.x86_64 >> libqb-devel-1.0rc3-1.el7.x86_64 >> >> [root@build-centos71 pacemaker ((Pacemaker-1.1.14))]# git checkout >> Pacemaker-1.1.14 >> HEAD is now at 70404b0... Merge pull request #892 from kgaillot/1.1 >> [root@build-centos71 pacemaker ((Pacemaker-1.1.14))]# make release >> (snip) >> In file included from ../../include/crm_internal.h:33:0, >> from ipc.c:19: >> ipc.c: In function 'crm_ipcs_flush_events': >> ../../include/crm/common/logging.h:140:23: error: format '%d' expects >> argument of type 'int', but argument 10 has type 'ssize_t' >> [-Werror=format=] >> static struct qb_log_callsite *trace_cs = NULL; \ >>^ >> ../../include/crm/common/logging.h:254:37: note: in expansion of macro >> 'do_crm_log_unlikely' >> # define crm_trace(fmt, args...) do_crm_log_unlikely(LOG_TRACE, >> fmt , ##args) >> ^ >> [...] > > Looks like that PR of mine turned out to be generally a good thing as it > effectively discovers wrong printf-format usage compared to the actual > arguments, exactly according to what you report :-) > > Will take care of the problematic code in Pacemaker. https://github.com/ClusterLabs/pacemaker/pull/939 -- Jan (Poki) pgpaNFhZ0yryR.pgp Description: PGP signature ___ Users mailing list: Users@clusterlabs.org http://clusterlabs.org/mailman/listinfo/users Project Home: http://www.clusterlabs.org Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf Bugs: http://bugs.clusterlabs.org
Re: [ClusterLabs] [Announce] libqb 10.rc3 release
On 25/02/16 13:24 +, Christine Caulfield wrote: > I am pleased to announce the third 1.0 release candidate release of > libqb. Huge thanks to all those who have contributed to this release. Once again, there are EPEL/Fedora builds for convenient testing: https://copr.fedorainfracloud.org/coprs/jpokorny/libqb/build/163316/ Note that upsteam spec file was used (make libqb.spec), only with a slight adjustment of the Version & Release tags + pointing Source0 to https://github.com/ClusterLabs/libqb/releases/download/v1.0rc3/libqb-1.0rc3.tar.xz This time around, all users of libqb are strongly advised to try building with this version of libqb to discover potential misuse of printf-like format string vs. actual arguments serving for substitution within it. This is due to the newly introduced annotation to give the compiler a hint about the printf-like semantics of the arguments to particular functions: https://github.com/ClusterLabs/libqb/pull/175 Warm thanks to Keisuke-san for pointing the possible issues arising from this coding defensive measure, as he was first to report failure to build Pacemaker as a direct consequence (I am fixing that right now): http://oss.clusterlabs.org/pipermail/users/2016-February/002375.html -- Jan (Poki) pgpQS2GXKuRde.pgp Description: PGP signature ___ Users mailing list: Users@clusterlabs.org http://clusterlabs.org/mailman/listinfo/users Project Home: http://www.clusterlabs.org Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf Bugs: http://bugs.clusterlabs.org
Re: [ClusterLabs] [Announce] libqb 10.rc3 release
On 26/02/16 14:27 +0900, Keisuke MORI wrote: > As of libqb-1.0rc3, Pacemaker fails to build upon it with the gcc > warnings as below. > There was no such a problem until 1.0rc2, and it seems that the > changes in the pull request #175 is related. > > https://github.com/ClusterLabs/libqb/pull/175 > > {{{ > [root@build-centos71 pacemaker ((Pacemaker-1.1.14))]# rpm -qa | grep libqb > libqb-1.0rc3-1.el7.x86_64 > libqb-devel-1.0rc3-1.el7.x86_64 > > [root@build-centos71 pacemaker ((Pacemaker-1.1.14))]# git checkout > Pacemaker-1.1.14 > HEAD is now at 70404b0... Merge pull request #892 from kgaillot/1.1 > [root@build-centos71 pacemaker ((Pacemaker-1.1.14))]# make release > (snip) > In file included from ../../include/crm_internal.h:33:0, > from ipc.c:19: > ipc.c: In function 'crm_ipcs_flush_events': > ../../include/crm/common/logging.h:140:23: error: format '%d' expects > argument of type 'int', but argument 10 has type 'ssize_t' > [-Werror=format=] > static struct qb_log_callsite *trace_cs = NULL; \ >^ > ../../include/crm/common/logging.h:254:37: note: in expansion of macro > 'do_crm_log_unlikely' > # define crm_trace(fmt, args...) do_crm_log_unlikely(LOG_TRACE, > fmt , ##args) > ^ > [...] Looks like that PR of mine turned out to be generally a good thing as it effectively discovers wrong printf-format usage compared to the actual arguments, exactly according to what you report :-) Will take care of the problematic code in Pacemaker. -- Jan (Poki) pgp2pRrOiPRpq.pgp Description: PGP signature ___ Users mailing list: Users@clusterlabs.org http://clusterlabs.org/mailman/listinfo/users Project Home: http://www.clusterlabs.org Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf Bugs: http://bugs.clusterlabs.org