On Sat, 10 Apr 1999, John Polstra wrote:
> Brian Feldman wrote:
>
> > YES! Make the mdXXX symbols all weak, so libmd can be added manually!
>
> Peter already mentioned why that isn't necessary. I just want to add
> a caution to folks out there. Don't go wild with weak symbols. Think
> VERY ca
Mark Murray wrote:
> John Polstra wrote:
>> I ran into a similar problem for static PAM. After some discussion
>> with Bruce, I held my nose and put this into bsd.libnames.mk:
>
>:-) I can see why you don't like that solution; I can also see why
> it is unavoidable.
>
> Any objections to me doin
John Polstra wrote:
> I ran into a similar problem for static PAM. After some discussion
> with Bruce, I held my nose and put this into bsd.libnames.mk:
:-) I can see why you don't like that solution; I can also see why
it is unavoidable.
Any objections to me doing something similar with libcryp
Peter Wemm wrote:
> What I would like is a minimal /lib, containing a few key libraries like
> libc.so, libutil.so, etc and have everything dynamic.
I support that idea too. It's time. We can't support all-static
systems forever. It's becoming a ball and chain as more and more
applications use
Brian Feldman wrote:
> YES! Make the mdXXX symbols all weak, so libmd can be added manually!
Peter already mentioned why that isn't necessary. I just want to add
a caution to folks out there. Don't go wild with weak symbols. Think
VERY carefully (and get a review) before using them. They shou
Mark Murray wrote:
>
> The "undefined references" are all in libmd which is listed as above
> in libcrypt's Makefile.
>
> I would hate to have to hunt down all usages of -lcrypt to add -lmd.
As Peter already mentioned, the problem is that library dependencies
aren't recorded in archive libraries
Brian Feldman wrote:
> On Sat, 10 Apr 1999, Peter Wemm wrote:
>
> > Mark Murray wrote:
> > > Hi
> > >
> > > I've been helping somebody along with getting the new libcrypt going,
> > > and something has broken recently (post-egcs?), and I'm having a wee
> > > problem debugging it.
> > >
> > > Lib
On Sat, 10 Apr 1999, Peter Wemm wrote:
> Mark Murray wrote:
> > Hi
> >
> > I've been helping somebody along with getting the new libcrypt going,
> > and something has broken recently (post-egcs?), and I'm having a wee
> > problem debugging it.
> >
> > Libcrypt uses routines out of libmd (MD[45]*
Peter Wemm wrote:
> > cc -O -pipe -I/usr/src/gnu/usr.bin/perl/miniperl/../../../../contrib/perl5
-I
> /usr/obj/usr/src/gnu/usr.bin/perl/miniperl -I/usr/obj/usr/src/tmp/usr/i
nc
> lude -static -o miniperl miniperlmain.o -lperl -lm -lcrypt
>^^^
> > /usr/obj/usr/src/tmp/us
Mark Murray wrote:
> Hi
>
> I've been helping somebody along with getting the new libcrypt going,
> and something has broken recently (post-egcs?), and I'm having a wee
> problem debugging it.
>
> Libcrypt uses routines out of libmd (MD[45]* and SHA*), and a while
> back (pre-egcs), it was possib
Hi
I've been helping somebody along with getting the new libcrypt going,
and something has broken recently (post-egcs?), and I'm having a wee
problem debugging it.
Libcrypt uses routines out of libmd (MD[45]* and SHA*), and a while
back (pre-egcs), it was possible to have an "LDADD+=-lmd" in libc
11 matches
Mail list logo
