Re: [orbitcpp-list] Re: [orbitcpp-commits] orbitcpp inducer

[Andreas K. <ak...@ix...>]

> I prefer to stick to glib types, because we already require glib and we
> know it's portable (it even implements some functions if they dont exist
> on systems). I don't think I can be sure to say the same for unsigned
> int and assert. You may have more knowledge on this than me, but I take
> the working assumption that if glib doesn't port properly then we can
> fix it.

Those unsigned int's are gone already, now they're CORBA::ULong's, which
I think is best. concerning the assert, I dunno. I think we should
minimize or at least encapsulate glib dependencies because <illusion>
maybe someone else also has a c orb he needs wrapped to be a c++ orb.
</illusion> 

> 
> > * added FAQ (wanna know what i was doing lately? look here. :((()
> > * fixed some tests to remedy problem from FAQ question 1
> > * added exceptionlist mechanism
> > * removed check for glib from configure.in:
> > is necessarily included with ORBit
> > * fixed bad bug in smartptr.hh _retn methods
> > * fixed bug with empty exceptions
> > * reimplemented PortableServer::ObjectId in terms of sequences
> > * removed code duplication from sequence template impl, implemented
> > missing copy and assignment, made implementation out-of-memory-aware,
> > implemented the specified allocbuf/freebuf functions, made the
> > sequence template roll its own CPP struct...
> >
> > ah well, call it a reimplementation of orbitcpp_sequence. :)
> 
> Cool! Can't wait to see what you've done!

not too much new code, actually. hardest were those darn segfaults i've
been hunting. :(((

cya
andy