[omniORB] omniORB_280 and Sun C++ 5.0 - and omniEvents

[Sai-Lai Lo]

As Robert said, its no point try to use Sun C++ 5 until one has got all
their patches applied. Even so, the last time I give Sun C++ 5 a try was
back in January when they release the last patch 107311-09. The result was
disappointing.  When compiled with -O2 and without -g, eg2_impl SEGV after
it replies to eg2_clt. It simply got the exception handling list all mess
up with -O2. Could be a problem with raising an exception from a shared
library. I don't know and don't have time to keep testing that compiler.

The generated code seems alright with -O2 -g but I think you should know
its limitation before you rely too much on that compiler.

For me, I've been using gcc-2.95 on that platform. At least I don't have to
worry about the compiler.

To fellow omniORBers, the question about getting omniORB to compile with
Sun C++ 5 has come up every month or so. Please search through the archive
for the discussion (http://www.uk.research.att.com/search.html).

Sai-Lai


>>>>> Robert Jonsson writes:

> Hi everybody,
> I spent eons trying to compile omniORB and omniEvents on Solaris 7 and Sun C++
> 5.0, I think we have everything working now.
> I'm sorry to say though that it has been a while since I got it working and I
> don't remember all the strange things we where faced with.

> Firstly, we are NOT using compatibility mode which makes a lot of difference.
> Also we are using Rogue Wave which also makes some difference.

> Looking at our current makefile I think there are two changes that are
> necessary.

> The AR line should be declared as:
> AR = CC -xar -o        # I think it has to do with new object file format,
> perhaps it isn't needed

> Also add -library=iostream to the propriate FLAG, I seem to have put it under
> CXXDEBUGFLAGS seems strange but works.

> As far as omniEvents are concerned we have yet to test it consistently but it
> seems to work. The changes where quite a few, I won't try to remember all the
> changes if there is any interest I'll try to find out. As I remember it most had
> to do with namespaces.

> But then again I've tried so many different compile changes it isn't at all
> certain the above achieves anything!


-- 
Sai-Lai Lo                                   S.Lo@uk.research.att.com
AT&T Laboratories Cambridge           WWW:   http://www.uk.research.att.com 
24a Trumpington Street                Tel:   +44 1223 343000
Cambridge CB2 1QA                     Fax:   +44 1223 313542
ENGLAND