#1 probably not #2 didn't used to, but ...
#5 not always - depends on the packaging of the libxml stuff, which as it's sometimes part of gnome and sometimes not isn't very reliable. Rather than changing the #include, I'd suggest altering the tests to support the 2nd 'common' location. -----Burton > -----Original Message----- > From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] > Behalf Of Daniel Macks > Sent: Thursday, August 19, 2004 6:38 PM > To: [EMAIL PROTECTED] > Subject: [Ntop-dev] ntop-3.0 compile bugs > > > 1. In configure.in starting at line 1315: > > CFLAGS="$CFLAGS -g" > AC_MSG_CHECKING([if compiler accepts -g (debugging information)]) > oCFLAGS=$CFLAGS > CFLAGS="-g $CFLAGS" > AC_TRY_COMPILE( , > return 0 , > AC_MSG_RESULT(yes), > CFLAGS="${oCFLAGS}" > AC_MSG_RESULT(no) > ) > > Should that first line really be there? It appears to *always* add > -g. That means the test may add a second -g, but if -g is not > supported, there will still be the first one passed. > > 2. ./configure does several checks for getopt, but then always builds > its own getopt.o regardless of whether getopt_long was found. > > 3. myrrd/Makefile.in gets CFLAGS but not CPPFLAGS, so I have to > duplicate some parts of the CPPFLAGS variable. > > 4. An empty directory "plugins" is created in libdir but all the > plugin dynamic libs go "loose" in libdir. > > 5. The configure check for xmlversion.h looks for that file "loose" in > the usual include places. But libxml appears to put all its .h in a > libxml/ subdirectory, and are loaded as '#include <libxml/foo.h>'. > > dan > > -- > Daniel Macks > [EMAIL PROTECTED] > http://www.netspace.org/~dmacks > > _______________________________________________ > Ntop-dev mailing list > [EMAIL PROTECTED] > http://listgateway.unipi.it/mailman/listinfo/ntop-dev _______________________________________________ Ntop-dev mailing list [EMAIL PROTECTED] http://listgateway.unipi.it/mailman/listinfo/ntop-dev
