Re: mico compile fails, lots of memory

[Raja R Harinath <harinath cs umn edu>]

Michael Harnois <mharnois@sbt.net> writes:
> I have 56meg of RAM and 64MB of swap space, so I don't think I'm
> experiencing the same problem others have reported, but mico 2.0.5
> from cvs is not building for me. Everything is fine until it tries to
> use the newly compiled idl in building in auxdir; then I get
> 
> /src/mico/./idl/idl --name dynany ../include/mico/dynany.idl
> make[1]: *** [dynany.cc] Segmentation fault (core dumped)
> 
> It's idl that's segfaulting. Ideas?

Are you using a `egcs' snapshot or something.  I had a similar problem,
and the fix that worked was:

  Index: ir/ir_impl.cc
  ===================================================================
  RCS file: /debian/home/gnomecvs/mico/ir/ir_impl.cc,v
  retrieving revision 1.2
  diff -u -p -r1.2 mico/ir/ir_impl.cc
  --- mico/ir/ir_impl.cc  23:01:01	1.2
  +++ mico/ir/ir_impl.cc  17:26:19
  @@ -532,7 +532,7 @@ CORBA::Contained_ptr Container_impl::loo
	 if( CORBA::is_nil( con ) )
	  return CORBA::Contained::_nil();
	 strip_first_scope( name );
  -      if( name == "" )
  +      if( name.length() == 0 )
	  return CORBA::Contained::_duplicate( con );

	 // con should be a container

This looks like a bug in the `egcs' compiler, but I haven't had the
time to come up with a testcase to report to them.  Also, I haven't
compiled it with newer snapshots: things might have gotten better.

HTH
- Hari
-- 
Raja R Harinath ------------------------------ harinath@cs.umn.edu
"When all else fails, read the instructions."      -- Cahn's Axiom
"Our policy is, when in doubt, do the right thing."   -- Roy L Ash