Re: Why is libc-2.1.*.so not stripped?

[Steve Dunham <dunham@cse.msu.edu>]

Steve Dunham <dunham@cse.msu.edu> writes:

> Nils Rennebarth <nils@ipe.uni-stuttgart.de> writes:
> 
> > [1  <text/plain; us-ascii (7bit)>]
> > Subject says it all:
> >    Why is the libc DLL not stripped on potato systems? It enlarges *every*
> > executable (linked dynamically!) by about 100k
> 
> > Long Story:
> > I compiled a short gtk program that I wrote as an example to learn gtk
> > programming on a potato system. The executable (linked dynamically,
> > stripped) was > 100k in size.
> 
> > Then I transferred the program to a slink system, compiled again, using the
> > same Makefile and presto, the size of the executable shrank to 7k stripped,
> > a much more reasonable size.
> 
> > After much searching around I compiled the simplest hello world program on
> > the potato machine. source code like this:
> > -------------------------------------------------------------
> > #include <stdio.h>
> > 
> > int main(int argc, char **argv)
> > {
> >         printf("Hi folks\n");
> > }
> > -------------------------------------------------------------
> > again > 100k in size.
> 
> This has nothing to do with whether libc is stripped or not.
> 
> Also:
> 
>  # gcc foo.c
>  # strip a.out
>  # ls -l a.out
>  -rwx------   1 dunham   gnats        3064 Jul 26 16:30 a.out*

> This is on an up-to-date potato system, using your code.

> I'm guessing that your problem is that you are compiling this as a C++
> program, with exceptions enabled.  (Exceptions are enabled in g++ by
> default because if you compile any code with exceptions disabled, you
> can't use it in a program that needs exceptions.)

I take this back, g++ makes the program 4028 bytes stripped.


Steve
dunham@cse.msu.edu