On Sun, Feb 08, 2004 at 03:08:32PM -0800, Guy Harris wrote:
> I installed autoconf 2.52, automake 1.6, and libtool 1.5 on my Solaris
> partition (automake 1.5 blew up with some weird problems, but 1.6
> didn't; 1.6 requires a minimum of autoconf 2.52).
AFAIK, automake had some strange porblems - IIRc the KDE people refused
to use it for a long time, but I don't remember why.
Btw, I'd be very happy if we could require automake 1.6, autoconf 2.52
and libtool 1.5 :)
> /bin/sh ./libtool --mode=link gcc -D_U_="__attribute__((unused))"
> -Wall -W -g -O2 -I/usr/local/include
> -I/usr/local/lib/glib/include -I/usr/local/include -o
> libethereal.la -rpath /usr/local/lib -version-info 0:1:0
> atalk-utils.lo bitswap.lo circuit.lo conversation.lo
> column-utils.lo epan.lo except.lo filesystem.lo frame_data.lo
> int-64bit.lo ipv4.lo osi-utils.lo packet.lo plugins.lo proto.lo
> resolv.lo sna-utils.lo strutil.lo to_str.lo tvbuff.lo
> value_string.lo inet_aton.o inet_pton.o inet_ntop.o
> ftypes/libftypes.a -L/usr/local/lib -R/usr/local/lib -lgmodule
> -lglib -ldl
>
> *** Warning: Linking the shared library libethereal.la against the
> *** static library ftypes/libftypes.a is not portable!
Ah, OK, I'll try to fix this later on Monday or Tuesday.
> Speaking of position-independent code, note that making libethereal PIC
> will slow down performance somewhat, and making the dissectors PIC will
> do so even more. I don't know how much - as I remember from my SunOS
> 4.x days, the claim was that it was about a 10% performance hit.
Hmm, I'm confused: This shouldn't cost performance on a SPARC processor
with its many registers. The platform that once caused trouble was the
x86 processor, but that was "cured" by the move from a.out to elf. OK,
its late and maybe I'm dreaming up things
Good night
Jörg
--
Joerg Mayer <jmayer@xxxxxxxxx>
We are stuck with technology when what we really want is just stuff that
works. Some say that should read Microsoft instead of technology.
