(Illustration by Gaich Muramatsu)
In some email I received from Jan Harkes <jaharkes_at_cs.cmu.edu> on Fri, 4 Apr 2003 09:42:23 -0500, wrote: > On Fri, Apr 04, 2003 at 11:47:25AM +0100, lou wrote: > > hate to tell you guys but it`s broken on FreeBSD -current(however it > > works fine on my -stable box) too > > > > <snip net/coda5_server> > > /u1/portsbuild/u1/ports/net/coda5_server/work/coda-5.3.20/include/auth2.h:173:8: > > warning: extra tokens at end of #endif directive fixed. > Ok, can you first of all use a more recent rpc2 version than v1.11. > This was fixed May 22 16:31:46 2001. i.e. almost 2 years ago. We're > currently at version 1.15. > > > > comm.cc: In member function `connent* conn_iterator::operator()()': > > comm.cc:606: warning: invalid offsetof from non-POD type `class connent'; use > > pointer to member instead > > Non-fatal. gone too. > > comm.cc: In function `void MultiProbe(int, RPC2_Handle*)': > > comm.cc:924: invalid conversion from `long int (*)(int, RPC2_Handle*, long int, long > > int, ...)' to `long int (*)()' > > > And this one was fixed May 23 16:50:16 2001. same. > > comm.cc: In member function `srvent* srv_iterator::operator()()': > > comm.cc:1454: warning: invalid offsetof from non-POD type `class srvent'; use pointer > > to member instead > > Again non-fatal. A fix for these warnings is in CVS, but some versions > of gcc-3.2 seem to generate broken code when C structs contain C++ > objects as members. But the latest gcc-3.2.3 seems to do the right thing. right, thanks Jan, it works and compiles fine now. wey! good pointer, it looks like i was using the old rvm/lwp/rpc2, and somehow i overlooked it. anyways you can add freebsd -current to your list, if you havent done it already! thanks again. cheersReceived on 2003-04-04 10:06:45