[omniORB] Re: Newbie: Compiling problems omniORB-4.0.5 using Cygwin, WinXP, MS VC7.0

[Roland Goecke]

To follow-up my own posting, I managed to compile omniORB using 
omnipython, which I guess is something, but if anyone could give me a 
hint what to do to get it to work with a normal full installation of 
Python, I'd appreciate it.

I am not sure if omniORB works properly now. I followed the instructions 
in README.win32.txt:

- Set the PATH environment variable
- Set OMNINAMES_LOGDIR
- Started OmniNames with: omniNames -start
- Edited the registry key HKEY_LOCAL_MACHINE\SOFTWARE\omniORB\InitRef to 
name "1" and value "NameService=corbaname::myPC.anu.edu.au"

However, when I now try to check if things are OK using

 > nameclt list

I get the following error message:

Unexpected error when trying to narrow the NamingContext.

As suggested in README.win32.txt, I tried to set InitRef directly under 
the omniORB key instead of a string with name 1 under, but I get the 
same error message. I suppose that means the registry value isn't set 
correctly, but how can I fix it? What value or name do I need to set?

Any suggestions welcome!

Cheers,
Roland

> Hi,
> 
> I am trying to compile omniORB-4.0.5 on a Windows XP machine with 
> Cygwin, Python2.4 and MS Visual Studio 7.0. I followed the instructions 
> in README.win32.txt, selected the configuration
> 
> platform = x86_nt_4.0_vs_7
> 
> in config/config.mk and set the path to my Python2.4 installation in 
> \mk\platforms\x86_nt_4.0_vs_7.mk.
> 
> omniORB seems to compile fine for a while until I get the following 
> error message:
> 
> c:\omniORB-4.0.5\bin\x86_win32\omkdepend.exe: warning: idlpython.cc, 
> line 172: unknown directive == "#warning "Please use an up-to-date 
> version of gcc! (>2.91 recommended)""
> c:\omniORB4.0.5\bin\x86_win32\omkdepend.exe: warning: (from 
> idlpython.cc) c:\progra~1\python2.4\include\Python.h: 21: #error 
> "Something's broken. UCHAR_MAX should be defined in limits.h"
> c:\omniORB4.0.5\bin\x86_win32\omkdepend.exe: warning: (from 
> idlpython.cc) c:\progra~1\python2.4\include\Python.h: 25: #error 
> "Python's source code assumes C's unsigned char is an 8-bit type."
> c:\omniORB4.0.5\bin\x86_win32\omkdepend.exe: warning: (from 
> idlpython.cc) c:\progra~1\python2.4\include\Python.h: 34: #error 
> "Python.h requires that stdio.h define NULL."
> c:\omniORB4.0.5\bin\x86_win32\omkdepend.exe: warning: (from 
> idlpython.cc) c:\progra~1\python2.4\include\Python.h: 367: #error 
> "omniidl requires Python 1.5.2 or higher"
> + rm -f static/omnithread.lib
> + ../../../bin/x86_win32/libwrapper -gnuwin32 static/omnithread.lib 
> static/nt.o
> + install -c static/omnithread.lib ../../../lib/x86_win32
> + rm -f debug/omnithreadd.lib
> + ../../../bin/x86_win32/libwrapper -gnuwin32 debug/omnithreadd.lib 
> debug/nt.o
> + install -c debug/omnithread.lib ../../../lib/x86_win32
> 'import site' failed; use -v for traceback
> Traceback (most recent call last):
>   File "<string>", line 1, in ?
>   File "c:\Program Files\Python2.4\lib\compileall.py", line 15, in ?
>     import os
>   File "C:\Program Files\Python2.4\lib\os.py", line 651, in ?
>     import copy_reg as _copy_reg
>   File "C:\Program Files\Python2.4\lib\copy_reg.py", line 7, in ?
>     from types import ClassType as _ClassType
>   File "c:\omniORB-4.0.5\lib\python\omniidl_be\cxx\types.py", line 29, in ?
>     import string
>   File "C:\Program Files\Python2.4\lib\string.py", line 83, in ?
>     import re as _re
>   File "C:\Program Files\Python2.4\lib\re.py", line 5, in ?
>     from sre import *
>   File "C:\Program Files\Python2.4\lib\sre.py", line 270, in ?
>     copy_reg.pickle(_pattern_type, _pickle, _compile)
> AttributeError: 'module' object has no attribute 'pickle'
> make[4]: *** [export] Error 1
> ...
> 
> 
> I don't understand the first warning, because isn't the compilation 
> supposed to happen with VS7 instead of GCC given the above mentioned 
> configuration choices? Why does it crash with some Python problem?
> 
> My apologies if these are dumb questions. I have been searching through 
> the mail archive but couldn't find a solution. Is the compiler somehow 
> picking up the wrong paths?
> 
> Cheers,
> Roland

-- 
--------------------------------------------------
Dr Roland Goecke, PhD
Autonomous Systems and Sensing Technologies
NICTA Canberra Laboratory
Mail : Locked Bag 8001, Canberra ACT 2601, Australia
Phone: +61(0)2-6125 1259
Fax  : +61(0)2-6125 8651
Email: roland.goecke at ieee.org