Hello, On Mon, Jan 16, 2006 at 11:21:48PM +0100, spiral voice wrote: > Hi, > > testing MLDonkey 2.7.1 on Solaris 9/Sparc did not show the tiger tree error. > > This error was fixed in 2.7.1: > 2005/12/01 > 4657: Fix for C files for AMD64, mipsel platforms > improve endianness checking in tiger.c, md5_c.c and CryptoPP.h, > print endianness in buildinfo, removed sha1_old code > > Therefore I am a bit surprised that this error still exists. > Could you please compile Ocaml and MLDonkey yourself? > > Remove all Ocaml and MLDonkey stuff from Debian/Sparc > > wget http://caml.inria.fr/pub/distrib/ocaml-3.09/ocaml-3.09.1.tar.bz2 > tar -xjvf ocaml-3.09.1.tar.bz2 > cd ocaml-3.09.1 > ./configure > make world.opt > make install > > wget > http://download.savannah.gnu.org/releases/mldonkey/mldonkey-2.7.2.tar.bz2 > tar -xjvf mldonkey-2.7.2.tar.bz2 > cd mldonkey-2.7.2 > ./configure > make > ./mlnet -stdout > > No tiger tree error should appear now. If it still does, please check > /usr/local/lib/ocaml/caml/config.h. A sparc CPU is big-endian > (this should also be reported by MLDonkey buildinfo command) > and ARCH_BIG_ENDIAN should be defined in config.h. > MLDonkey uses this header file since the patch mentioned above. > So if 2.7.1 still reports failed tiger tree check then something could > be wrong when detecting endianess. >
I just can't use a custom ocaml binary (it will be really complicated). However, i think ocaml 3.09.1 doesn't suffer the tiger tree problem (the problem arise on ocaml 3.09.0). I think it should be fixed in the next upload (mldonkey 2.7.2 + ocaml 3.09.1). But for now, i am waiting for a fixed lablgtk2 version (one that will compile on all debian arches). Kind regard Sylvain Le Gall -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]