> On Friday 13 April 2001 2:22 am, Kevin Ryde wrote:
> >
> > Without wanting to repeat myself, I posted libtool-patches in January
> > about SCO 5 and -lc on shared libraries, but the problem still exists
> > in 1.3d and the current cvs.
>
> > Something like below (re-diffed for current libtool.m4) will be much
> > wanted, since it avoids seg faults if a shared library uses ctype.h
> > but the mainline program doesn't. I don't have an actual explanation
> > for what goes wrong, but the advice from SCO is apparently "don't do
> > that".
"Gary V. Vaughan" <[EMAIL PROTECTED]> writes:
>
> Oh yuck!
>
> This is definitely the wrong approach. Could you try to find a way to fix
> the test that discovers the correct value for lt_cv_archive_cmds_need_lc
> (aound line 2158 in libtool.m4).
cc and gcc on sco know not to add -lc, so the test there looks like it
would probably be right, except that archive_cmds is setup to use $LD,
so it doesn't come into play.
Maybe archive_cmds should switch to using $CC, though I wouldn't know
if that might upset anything else.
> If this is just a bug in SCO's linker, maybe we just need to move
> your patch into this section of libtool.m4?
I don't know if it counts as a bug, though rumour has it there's a
patch for unixware to address it. Certainly something strange is
going on, it doesn't look like a basic ld.so/ELF setup, libc.so is an
ar archive for a start. Cc to a sco expert for further remarks.
_______________________________________________
Libtool mailing list
[EMAIL PROTECTED]
http://mail.gnu.org/mailman/listinfo/libtool
