On Sat, 2011-11-26 at 14:21 +1300, Olly Betts wrote: > On Mon, Oct 24, 2011 at 01:13:05PM +0200, Mònica Ramírez Arceda wrote: > > During a rebuild of all packages in sid, your package failed to build on > > amd64. > > > > Relevant part: > > > x86_64-linux-gnu-g++ -DHAVE_CONFIG_H -I. -I.. > > > -DDATA_DIR=\"/usr/share/pgadmin3/\" -I../pgadmin/include > > > -I/usr/include/postgresql/9.1/server -I/usr/include/postgresql > > > -I/usr/include/postgresql/9.1/server -I/usr/include/postgresql -DSSL > > > -DHAVE_CONNINFO_PARSE -I/usr/lib/wx/include/gtk2-unicode-release-2.8 > > > -I/usr/include/wx-2.8 -D_FILE_OFFSET_BITS=64 -D_LARGE_FILES -D__WXGTK__ > > > -DEMBED_XRC -I/usr/include/libxml2 -I/usr/include/libxml2 -Wall -O2 -c > > > -o ctlCheckTreeView.o `test -f './ctl/ctlCheckTreeView.cpp' || echo > > > './'`./ctl/ctlCheckTreeView.cpp > > > In file included from ../pgadmin/include/utils/sysSettings.h:15:0, > > > from ../pgadmin/include/pgAdmin3.h:37, > > > from ./agent/pgaStep.cpp:16: > > > ../pgadmin/include/utils/sysLogger.h:62:1: error: expected initializer > > > before 'ATTRIBUTE_PRINTF_1' > > This is due to this patch which I recently applied to wxwidgets2.8: > > http://patch-tracker.debian.org/patch/series/view/wxwidgets2.8/2.8.12.1-3/fix-macro-namespace-pollution.patch > > The bug motivating this change is here: > > http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=521924 > > My take on this is that the ATTRIBUTE_PRINTF* macros weren't intended to > be used outside of the wx headers - there aren't publicly documented > anywhere, and wx upstream has renamed them in 2.9 without mentioning the > fact in docs/changes.txt (which is intended to document all wx API > changes). > > So I would say it's a bug in pgadmin that it makes use of these macros - > it should define them itself if it needs them (and to avoid clashes with > everyone else who carelessly defines them without thinking through the > implications, it ought to define something like PG_ATTRIBUTE_PRINTF* > instead of ATTRIBUTE_PRINTF*). > > These macros expand to GCC attributes which allow printf format > checking, which allow the compiler to perform more checks but don't > matter for correct compilation. So a crude workaround to address this > in the short-term would be to just define them to be empty if not > already defined, for example: > > #ifndef ATTRIBUTE_PRINTF_1 > # define ATTRIBUTE_PRINTF_1 > #endif >
Here is the patch applied upstream: http://git.postgresql.org/gitweb/?p=pgadmin3.git;a=commitdiff;h=41ed1ae4f177e60e770e326b75ae3160a4041efc Regards. -- Guillaume http://blog.guillaume.lelarge.info http://www.dalibo.com -- To UNSUBSCRIBE, email to debian-bugs-rc-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org