[omniORB] Cannot get omniorb 3.0.2 to compile on Solaris 2.6

[Alwyn Schoeman]

 Back to the new error:

When checking for the _pure_error symbol I see that it is both
in libC and libCrun....
Omniorb seems to use Crun when compiling with 5.0.
Only thing I can see that might be a problem is that in 
libCrun the symbol is run::pure_error().
Do you think the namespace is the problem or the fact that it
is a function and not _pure_error...

-----Original Message-----
From: Duncan Grisby
To: Alwyn Schoeman
Cc: omniorb-list@uk.research.att.com
Sent: 10/13/2000 3:13 PM
Subject: Re: [omniORB] Cannot get omniorb 3.0.2 to compile on Solaris 2.6 

On Friday 13 October, Alwyn Schoeman wrote:

>  Thank you that helped.  I however now get the following...

> omniidl: (The error was `ld.so.1: python: fatal: relocation error:
file
> /usr/local/stuff/omni/lib/sun4_sosV_5.6/_omniidlmodule.so: symbol
> _pure_error_: referenced symbol not found')

I'm stumped now. I think it must be a compiler bug. Do you have all
the available patches applied to the compiler?  omniidl definitely
worked for us with Sun CC 5.0 on Solaris 2.7. We no longer have CC
5.0, but we have 5.1 (also known as Workshop 6...), which works fine.

The Sun compilers are basically a disaster area. If at all possible, I
recommend that you use gcc 2.95 instead.

Sorry I can't be more help,

Duncan.

-- 
 -- Duncan Grisby  \  Research Engineer  --
  -- AT&T Laboratories Cambridge          --
   -- http://www.uk.research.att.com/~dpg1 --