gcc question/bug

[David Spindler]

The pre-processed source code is at http://hf.utexas.edu/gccbug.tar.gz
The original source code is from mysql-3.23 and causes a crash on freebsd
machines.

There are three files in the tar:
sql22.cc is the source file after pre-processing
sql22.s is the result of compiling sql22.cc with the -S option but w/o the -g 
sql22g.s is the result of compiling sql22.cc with the -S and -g option

I commented the differences between them by using the #ds and #de to designate
where differences started and ended.

They were compiled with:
c++ -S (-g) O3  -fno-implicit-templates -c sql22.cc



David Spindler
Network Analyst
University of Texas at Austin
512-475-9299
512-415-1434

On 2 Aug 2000, Alexandre Oliva wrote:

> On Aug  2, 2000, David Spindler <spindler@mail.utexas.edu> wrote:
> 
> > Isn't g++ supposed to produce the same assembly instructions
> > regardless of the debug mode?
> 
> Yep.  Do you have any sample code for which GCC generates different
> assembly code when given `-g'?
> 
> -- 
> Alexandre Oliva   Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
> Red Hat GCC Developer                  aoliva@{cygnus.com, redhat.com}
> CS PhD student at IC-Unicamp        oliva@{lsd.ic.unicamp.br, gnu.org}
> Free Software Evangelist    *Please* write to mailing lists, not to me
>