Re: Shared library annoyance with gcc-3_0-branch

[Linus Torvalds <torvalds at transmeta dot com>]

In article <or7l2qe3cf.fsf@guarana.lsd.ic.unicamp.br> you write:
>On Feb 16, 2001, Brad Lucier <lucier@math.purdue.edu> wrote:
>
>> Is the following suggestion of Theodore Papadopoulo reasonable?
>
>It has been suggested before.  I myself tried it once.  At that time,
>it was rejected.  The argument was that hard-coding search paths in
>executables would cause them to fail should the hard-coded directory
>become unavailable (as in an NFS server down).  Not even setting
>LD_LIBRARY_PATH would get you around the problem.

Is there any real reason why libgcc can't just be statically linked?

Yes, it's big for C++ programs.  Tough.  It sounds like all the problems
people have with dynamic libgcc are rather much worse, though.  And most
C++ programs tend to be big anyway - anobody who wants a lean and mean
binary tends to be using C in the first place anyway. 

And maybe it will give people incentive to try to keep the size of
libgcc down, instead of just constantly making it bigger and bigger.

		Linus