Hi Roman, On Mon, 2006-01-23 at 18:11 +0100, Roman Kennke wrote: > > If investigating this > > takes more than a day could you revert the patch temporarily so we don't > > have a broken tree for too long? > > This specific problem can be fixed by adding autoconf checks like these: > > AC_CHECK_CONSTANT_INCLUDES(F_SETLK, fcntl.h,AC_DEFINE(HAVE_F_SETLK, > 1,[define if the system has F_SETLK constant is in fcntl.h.])) > AC_CHECK_CONSTANT_INCLUDES(F_SETLKW,fcntl.h,AC_DEFINE(HAVE_F_SETLKW,1,[define > if the system has F_SETLKW constant is in fcntl.h.])) > AC_CHECK_CONSTANT_INCLUDES(F_RDLCK, fcntl.h,AC_DEFINE(HAVE_F_RDLCK, > 1,[define if the system has F_RDLCK constant is in fcntl.h.])) > AC_CHECK_CONSTANT_INCLUDES(F_WRLCK, fcntl.h,AC_DEFINE(HAVE_F_WRLCK, > 1,[define if the system has F_WRLCK constant is in fcntl.h.])) > AC_CHECK_CONSTANT_INCLUDES(F_UNLCK, fcntl.h,AC_DEFINE(HAVE_F_UNLCK, > 1,[define if the system has F_UNLCK constant is in fcntl.h.])) > > These are copy+pasted from Jamaica, I'll adjust this for Classpath asap > and check this in. Or maybe you'd look after this in the meantime? I > won't have time for this this evening ...
I looked, but I don't know where AC_CHECK_CONSTANT_INCLUDES comes from. I assume it is some AC_EGREP_CPP magic that sees if there is a define with that name. But I haven't tried to recreate the magic yet. Cheers, Mark
signature.asc
Description: This is a digitally signed message part
_______________________________________________ Classpath-patches mailing list [email protected] http://lists.gnu.org/mailman/listinfo/classpath-patches
