On Friday 01 January 2010 16:29:55 Albert Astals Cid wrote: > A Dijous, 31 de desembre de 2009, Raphael Kubo da Costa va escriure: > > On Thursday 31 December 2009 10:32:39 Alexander Neundorf wrote: > > > Hi Jos, Raphael, > > > > > > since this patch: > > > http://websvn.kde.org/trunk/kdesupport/strigi/src/streamanalyzer/saxeve > > > nt > > > > > >an alyzer.cpp?r1=1067133&r2=1067134 > > > > > > strigi does not build anymore on FreeBSD: > > > http://my.cdash.org/viewBuildError.php?buildid=43229 > > > > > > > > > What can be done about this ? > > > > > > One thing which should be done, is to make this into a configure-check. > > > I think check_symbol_exists() ( > > > http://www.cmake.org/cmake/help/cmake2.6docs.html#module:CheckSymbolExi > > > st > > > > > >s ) should work for that. > > > > > > Then it would not be a build error anymore, but a configure error, > > > which is at least slightly better. > > > > > > What needs to be done to make it build again ? > > > > > > Alex > > > > Libxml2 defines LIBXML_THREAD_ENABLED if _REENTRANT, __MT__ or > > > > _POSIX_C_SOURCE are defined. glibc always defined _POSIX_C_SOURCE, > > whereas FreeBSD's libc seems to define it conditionally (either it must > > be defined to some value by the user and then it's correctly set, or > > _XOPEN_SOURCE must be set -- see > > http://fxr.watson.org/fxr/source/sys/cdefs.h?v=FREEBSD8). > > Till already defined _REENTRANT yesterday because he had the same problem > in the MAC AFAIR.
He's defined _REENTRANT inside an IF (APPLE) block. Should we move it outside so that it's defined for all platforms? _______________________________________________ Kde-buildsystem mailing list [email protected] https://mail.kde.org/mailman/listinfo/kde-buildsystem
