orbitcpp-list

Re: [orbitcpp-list] packages, etc.

[Brian C. <bc...@th...>]

On Sun, Feb 24, 2002 at 07:00:43PM +1100, Sam Couter wrote:
> It's safer to compile it yourself or use the version compiled by your
> distribution vendor (if there is one)

Requiring my users to compile support libraries isn't an option. Near as I can
tell there are no vendors that ship orbitcpp. Guess I'll use C wrappers. 

I suppose I could also distribute orbitcpp with the app and install the lib in
an app specific directory. Using wrappers is still probably easier.

Thanks,
b.c.

Re: [orbitcpp-list] packages, etc.

[Sam C. <sc...@bi...>]

Brian Craft <bc...@th...> wrote:
> Are there plans to provide binaries for the major linux distributions?=20

A binary package is available as part of the Debian GNU/Linux system,
but I have no plans to provide them for other distributions.

It's safer to compile it yourself or use the version compiled by your
distribution vendor (if there is one), and it's not a difficult package
to compile. You wouldn't really be able to trust whatever packages I
made available for download, as they'd have to be donated by others. I
don't have the resources to compile the package for multiple
distributions.

> Is orbitcpp stable and complete enough to use in a currently shipping app?

That depends on what you're trying to do with it. What features do you
need? It works for the commercial application that I've used it for, but
my application was relatively straightforward.
--=20
Sam "Eddie" Couter  |  mailto:sc...@bi...
Debian Developer    |  mailto:ed...@de...
                    |  jabber:sa...@ja...
OpenPGP fingerprint:  A46B 9BB5 3148 7BEA 1F05  5BD5 8530 03AE DE89 C75C

[orbitcpp-list] packages, etc.

[Brian C. <bc...@th...>]

Hi all -- I'm in the process of adding corba to a c++ application. It would be
nice to have native language bindings, rather than having to maintain c
wrappers. But it doesn't look like orbitcpp is together enough, yet. In
particular, I can't introduce a dependence on a library which is only
distributed as source.

Are there plans to provide binaries for the major linux distributions? 

Is orbitcpp stable and complete enough to use in a currently shipping app?

b.c.

Re: [orbitcpp-list] Re: cpp branch: CORBA_Object struct hidden?

[<bb...@ya...>]

> I can't think of any way of retaining this C <-> C++ compatibility
> if
> the CORBA_Object struct definition remains private and subject to
> change. But I'm certainly not a C++ expert by any stretch of the
> imagination.
> 
> So... thoughts, anyone?


More thinking aloud, if anyone still want to pursue this course...

If you did inherit CORBA::Object privately from CORBA_Object, then
the same piece of memory could be used for C and C++ objects.
You would have to have an overloaded operator new and delete for the
CORBA::Object which used the C allocator and deallocator for
CORBA_Object. 
It doesn't seem possible that a CORBA::Object could ever exist as a
concrete object on the stack since its size cannot be determined but
only pointers should ever exist to this type anyway. Allocation on
the heap through new should be sufficient and possible. 
The concrete types are X_var and X_ptr which hold pointers, so this
seems OK.

A narrow operation might operate like

CORBA::XXX_ptr 
CORBA::XXX_ptr::narrow( const CORBA::Object_ptr & rhs )
{
    // Note: rhs.internalPtr_ is a CORBA::Object *

    // a CORBA_Object * is the same as a CORBA::Object *
    CORBA_Object * rhsPtr 
        = reinterpret_cast<CORBA_Object *>( rhs.internalPtr_ );

    // do ORBit narrow on internal pointer
    CORBA_XXX * newObj = CORBA_XXX_narrow( rhsPtr, env );
     
    // check for exceptions and throw if a bad narrow

    // this should be OK since CORBA::XXX inherits from CORBA_XXX
    CORBA::XXX * internalPtr = static_cast<CORBA::XXX *>(newObj);

    // create new XXX_ptr which is now responsible for newObj
    return CORBA::XXX_ptr(internalPtr);
}

When the XXX_ptr deletes its internal pointer, the operator delete
for the XXX * will actually call the ORBit routine to deallocate
itself (because it is also a CORBA_XXX pointer).

(Disclaimer: I've come in half way through this discussion so I may
be missing some important point. Also I don't know the C binding.)

Or is this all just wrong?




http://movies.yahoo.com.au - Yahoo! Movies
- Vote for your nominees in our online Oscars pool.

Re: [orbitcpp-list] Re: cpp branch: CORBA_Object struct hidden?

[Michael M. <mi...@xi...>]

Hi Sam,

On Mon, 2002-02-18 at 06:25, Sam Couter wrote:
> There shouldn't be any reason to hide the CORBA_Object from the C++
> binding, as the C++ binding will be rebuilt any time the CORBA_Object
> structure is redefined anyway.

	Quite :-)

> Hiding it from third-party bindings is a different matter.

	Yes.

> Or am I missing something?

	No :-)

		Michael.

-- 
 mm...@gn...  <><, Pseudo Engineer, itinerant idiot

[orbitcpp-list] [OT] ORBit2 missing configure pieces

[<bb...@ya...>]

OK, so I checkout the cpp branch of ORBit2 but I have problems.

1. autconf creates a configure which calls ltconfig --no-verify but
doesn't specify a host type. This causes ltconfig to fail so it
doesn't generate libtool.

2. When I hack aclocal to provide the host-type configure stops half
way through with 

./configure: line 3648: syntax error near unexpected token
`PKG_CHECK_MODULES(ORBIT,'

What versions of configure tools are people using?
I have
autconf 2.52
libtool 1.4.2

Please excuse my lack of knowledge about autoconf et al.

I'm assuming that PKG_CHECK_MODULES is an autoconf macro that I don't
have for some reason.

Anyone know what I'm missing here?


http://my.yahoo.com.au - My Yahoo!
- It's My Yahoo! Get your own!

[orbitcpp-list] Manifest for working with ORBit2

[<bb...@ya...>]

Murray,

Can you describe a manifest of what is needed to get started using
orbitcpp with ORBit2.
I seem to remember this was covered some time ago but I couldn't find
the info in the archives.

Which versions are known to work together and where can I get them?
Eg. how do I pull down the cpp branch of ORBit2, linc etc.
Is there any particular branch of orbitcpp required or is this
included in the ORBit2 area as was suggested some time ago?

--
Rich



http://my.yahoo.com.au - My Yahoo!
- It's My Yahoo! Get your own!

Re: [orbitcpp-list] PATCH : make orbitcpp work with ISO C++ compiler

[<bb...@ya...>]

> I didn't mean to be insulting, I just meant to impress upon you
> that this isn't a sensible way of us all working on the ORBit2 
> branch.
 
OK. Point taken. But for me ORBit2 is a low priority.

If ORBit2 is happy to coexist with ORBit1 on my system in the same
way that gcc3 coexists with gcc2 then one day soon I will ensure that
the patch works with ORBit2.

My appologies for taking your comments in the wrong light.

--
RIch


http://my.yahoo.com.au - My Yahoo!
- It's My Yahoo! Get your own!

[orbitcpp-list] [ orbitcpp-Bugs-503269 ] generate skel code not legal C++

[<no...@so...>]

Bugs item #503269, was opened at 2002-01-14 03:01
You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=100646&aid=503269&group_id=646

Category: None
Group: None
>Status: Closed
>Resolution: Fixed
Priority: 5
Submitted By: Richard Andrews (bbmaj7)
>Assigned to: Sam Couter (eddiesam)
Summary: generate skel code not legal C++

Initial Comment:

Let me present an example of what currently occurs.

This is a skel generated from the everything test

test_VariableLengthStruct * 
_skel_opVariable( ::PortableServer_Servant _servant,
const test_VariableLengthStruct * inArg,
test_VariableLengthStruct *inoutArg,
test_VariableLengthStruct **outArg, 
::CORBA_Environment *_ev) {
	::test::VariableLengthStruct *_retval = NULL;
	bool _results_valid = true;

	try {
		POA_test::StructServer * _self = ((_orbitcpp_Servant
*)_servant)->m_cppimpl;
		_retval = _self->opVariable( 
reinterpret_cast< const VariableLengthStruct&>(*inArg),
reinterpret_cast< VariableLengthStruct&>(*inoutArg),
reinterpret_cast< VariableLengthStruct*&>(*outArg) );
	}

I have removed namespace qualifications for clarity.

Now what happens here is that _self->opVariable expects
a Data_out<VariableLengthStruct> & (non-const ref) as
argument 3. Under C++ implied construction rules this
is no problem because Data_out<...> has a constructor
from VariableLengthStruct*&.

BUT... and here's the big problem.

This constructor creates a temporary object which is
passed as a reference through parameter 3, however only
const-references to temporaries may be passed in
standard C++ (as is now enforced in gcc3).

The only legal way to get around this problem is to
construct a wrapper object before the call, then pass
this as the non-const reference.

Eg.

try 
{ 
    POA_test::StructServer * _self =
((_orbitcpp_Servant *)_servant)->m_cppimpl; 

    Data_out< VariableLengthStruct> 
    wrapper3( reinterpret_cast<
VariableLengthStruct*&>(*outArg) );
    _retval = _self->opVariable( 
reinterpret_cast< const VariableLengthStruct&>(*inArg),
reinterpret_cast< VariableLengthStruct&>(*inoutArg), 
wrapper3); 

} 


In this case because wrapper3 is not a temporary it can
be passed through a non-const reference as is required
for argument 3 of opVariable().


In general the skeleton code for invoking operations
needs to be modified so that for each argument to the
servant implementation that is passed by non-const
reference, a temporary is constructed and passed.

I've been having a look at the code. The main areas of
interest are 

IDLPassSkels::doOperationSkel()

IDL_X_X_X_::getCPPSkelParameterTerm()

--

I put in a kludge in orbitcpp_smrtptr.hh about a year
ago using const_cast<>() to get around this problem but
it is no longer sufficient. The problem needs to be
fixed at the source which is the generated skel code.

I'm willing to put some time into fixing this if
someone familiar with this code can help me out.




----------------------------------------------------------------------

>Comment By: Sam Couter (eddiesam)
Date: 2002-01-21 03:43

Message:
Logged In: YES 
user_id=14486

Richard's supplied patch has been applied.

----------------------------------------------------------------------

You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=100646&aid=503269&group_id=646

Re: [orbitcpp-list] PATCH : make orbitcpp work with ISO C++ compiler

[Sam C. <sc...@bi...>]

Richard Andrews <bb...@ya...> wrote:
> Attached are some patches to make the skel output of orbitcpp legal
> ISO C++. See bug report 503269 on the orbitcpp page at sourceforge.

Patches have been applied and committed to the CVS repository.

Your contribution is appreciated. :)
--=20
Sam "Eddie" Couter  |  mailto:sc...@bi...
Debian Developer    |  mailto:ed...@de...
                    |  jabber:edd...@ja...
OpenPGP fingerprint:  A46B 9BB5 3148 7BEA 1F05  5BD5 8530 03AE DE89 C75C

[orbitcpp-list] [ orbitcpp-Bugs-475580 ] Compilation fail

[<no...@so...>]

Bugs item #475580, was opened at 2001-10-27 09:24
You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=100646&aid=475580&group_id=646

Category: None
Group: None
>Status: Closed
>Resolution: Fixed
Priority: 5
Submitted By: Francesco Lamonica (franci)
Assigned to: Sam Couter (eddiesam)
Summary: Compilation fail

Initial Comment:
Compilation fails on the test suite (at least in 
slackware 7.x and 8). This is due to the static 
linking against the glib.a libraray... maybe the 
compiler (gcc 2.95.3) should be explicitly told that 
is a static lib (static linking has always been a 
little annoying). Btw adding the -lglib flags made 
the compilation go smoothlessly


----------------------------------------------------------------------

>Comment By: Sam Couter (eddiesam)
Date: 2002-01-21 03:03

Message:
Logged In: YES 
user_id=14486

Braden has changed the build and removed the -static
compiler option.

----------------------------------------------------------------------

Comment By: Sam Couter (eddiesam)
Date: 2001-12-09 15:10

Message:
Logged In: YES 
user_id=14486

I think this bug is just a symptom, and the real problem is
that the tests are linking statically.
If the tests are modified to link dynamically instead, this
bug should go away.

----------------------------------------------------------------------

Comment By: Francesco Lamonica (franci)
Date: 2001-12-08 04:44

Message:
Logged In: YES 
user_id=1349

in the test suite the command line tell compiler to just use the 
glib.a file that would cos the linker to got nut and do not work 
properly.... instead of using the glib.a file using the shared library 
causes no problem at all. So i think that the linker is always 
checking for the shared library and that should be taold explicitly 
to usa static linking.... maybe with the -Wl,static flag but i am not 
sure about the correct syntax....


----------------------------------------------------------------------

Comment By: Sam Couter (eddiesam)
Date: 2001-12-07 22:49

Message:
Logged In: YES 
user_id=14486

Francesco: Can you provide more information on this? Can you
explain how the static linking causes problems?
I don't have any problems, using gcc 2.95.4 20011006.

----------------------------------------------------------------------

You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=100646&aid=475580&group_id=646

[orbitcpp-list] [ orbitcpp-Bugs-437571 ] error while compiling

[<no...@so...>]

Bugs item #437571, was opened at 2001-06-30 07:50
You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=100646&aid=437571&group_id=646

Category: None
Group: None
>Status: Closed
>Resolution: Works For Me
Priority: 5
Submitted By: Nobody/Anonymous (nobody)
Assigned to: Sam Couter (eddiesam)
Summary: error while compiling 

Initial Comment:
Making all in services
make[1]: Entering directory
`/usr/local/orbitcpp-0.30/services'
Making all in name
make[2]: Entering directory
`/usr/local/orbitcpp-0.30/services/name'
/opt/gnome/bin/orbit-idl -l c++
--backenddir=../../compiler/.libs ./CosNaming.id

** WARNING **: Module load failed:
../../compiler/.libs/liborbit-idl-c++-backend

** CRITICAL **: file orbit-idl-driver.c: line 50
(orbit_idl_to_backend): asserti

** WARNING **: ./CosNaming.idl compilation failed
sed -e 's/#include \CosNaming.h\/#include
\ORBitservices/CosNaming.h\/' Co
sed: can't read CosNaming-cpp-common.hh: No such file
or directory
make[2]: *** [CosNaming.cc] Error 2
make[2]: Leaving directory
`/usr/local/orbitcpp-0.30/services/name'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory
`/usr/local/orbitcpp-0.30/services'
make: *** [all-recursive] Error 1r

----------------------------------------------------------------------

Comment By: Sam Couter (eddiesam)
Date: 2001-12-07 22:35

Message:
Logged In: YES 
user_id=14486

I need more information to work this one out. I can't
reproduce it, and since it was submitted anonymously, I will
close it if I do not receive more information.
I need to know if this still happens with 0.30.2, and what
versions of ORBit and glib are installed, as well as what
compilation options were used.

----------------------------------------------------------------------

Comment By: Francesco Lamonica (franci)
Date: 2001-10-25 06:54

Message:
Logged In: YES 
user_id=1349

have u stripped the Orbit libs? that was causing a similar 
error on my box (slack 8 with gcc 2.95.3)


----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2001-06-30 07:54

Message:
Logged In: NO 

btw: running Slackware 7.1 
Kernel 2.4.5
gcc 2.95.3
ORBit 0.5.8

(get same errs with other packages, so it's prob me!!) 


----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2001-06-30 07:53

Message:
Logged In: NO 

btw: running Slackware 7.1 
Kernel 2.4.5
gcc 2.95.3
ORBit 0.5.8

(get same errs with other packages, so it's prob me!!) 


----------------------------------------------------------------------

You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=100646&aid=437571&group_id=646

[orbitcpp-list] PATCH : make orbitcpp work with ISO C++ compiler

[<bb...@ya...>]

 

Folks,

Attached are some patches to make the skel output of orbitcpp legal
ISO C++. See bug report 503269 on the orbitcpp page at sourceforge.

I've tested the patches under gcc-3.0.2. All tests pass.

The main change affects the production of skel code (pass_skels.cc)
and minor alterations to orbitcpp_smartptr.hh to make copy
constructors sane for out types.

Other minor necessary changes are made to client.cc and server.cc
under test/everything so they compile.

All tests passed successfully.

Can someone please verify that these changes are OK, and check them
in.


Thanks,

Rich

--

More info about what changed

For out arguments, implied construction was used with non-const
reference passing to servant methods. Implied construction creates
temporary objects which can only be passed though const references.
The patch constructs local objects in the skel code to allow the
arguments to be passed through non-const reference.
The exception is arrays of fixed length type which are passed as
pointers so construction is neither necessary nor possible.

Copy constructors in some of the smart pointer types needed to be
changed for the same reason.

client and server changes were just to make cout and endl visible in
the local namespace.


 





http://my.yahoo.com.au - My Yahoo!
- It's My Yahoo! Get your own!

[orbitcpp-list] Re: [gtkmm] ORBit2 cpp: cstddef include error

[Joe Y. <jw...@di...>]

> While trying to get the ORBit2 cpp branch (see link below) working with
> some example code, I'm getting a
> cstddef: No such file or directory
> error. This file is part of gcc, so  I'm hoping that somebody can
> suggest why it's not being found.
> 
> make[3]: Entering directory `/home/murrayc/bonoboui2-hello-cpp/src'
> gcc -DHAVE_CONFIG_H -I. -I. -I.. -I..
> hello.c
> In file included from
> /gnome/head/INSTALL/include/orbit-2.0/orbit/orb-cpp/orbitcpp.hh:42,
>                  from corba/Bonobo-cpp-common.hh:11,
>                  from corba/Bonobo-cpp.hh:6,
>                  from hello.c:36:
> /gnome/head/INSTALL/include/orbit-2.0/orbit/orb-cpp/orbitcpp_sequence.hh:32:19: cstddef: No such file or directory
> 

The problem is that you're compiling a C source file, so the libstdc++ 
headers aren't being implicitly -I'd in.  You need to explicitly 
set the language to c++ by adding "-x c++" to CFLAGS.

cheers,
--
Joe Yandle
http://www.divisionbyzero.com/jwy/pubkey.asc

If video games really affected kids, then we'd all be running around in
dark rooms, munching on pills, and listening to electronic music.

[orbitcpp-list] Finding servant types (long)

[<bb...@ya...>]

I submitted bug 503269 earlier this week.
I'm now working on a fix for it.
I've gotten half way but need some help from folks that know the code
better than me.

Basically in some cases it is necessary to construct an object (eg.
_out) to pass to the skel implementation (see the bug report for
why).


I'll take an (abbreviated) example of a generated skel from
test/everything/generated/everthing-cpp-skel.cc (orbitcpp-0.30.2)

LongArray_slice *POA_BasicServer::_skel_opLongArray(
    ::PortableServer_Servant _servant, 
    const LongArray inArg, 
    LongArray inoutArg, 
    LongArray outArg,
    ::CORBA_Environment *_ev){
...

retval = _self->opLongArray(
    reinterpret_cast< const ::LongArray_slice *&>(inArg), 
    reinterpret_cast< ::LongArray_slice *&>(inoutArg), 
    reinterpret_cast< ::LongArray_slice *&>(outArg) );

...
}

_self points to an object of type POA_test::BasicServer - the servant
implementation type.
When generating this skel code (ie. in
IDLPassSkels::doOperationSkel() ) how can I get strings which define
the types expected for each argument of
POA_test::BasicServer::opLongArray() ?

The code I need to generate would look something like

{
    TypeOfArg1 paramWraper1( reinterpret_cast< const
::LongArray_slice *&>(inArg) );
    TypeOfArg2 paramWrapper2( reinterpret_cast< ::LongArray_slice
*&>(inoutArg) );
    TypeOfArg3 paramWrapper3( reinterpret_cast< ::LongArray_slice
*&>(outArg) ) );

...

    _self->opLongArray( paramWrapper1, paramWrapper2, paramWrapper3
);
}

I need to know what to substitute for TypeOfArgX above.

In this case I would need (see everything-cpp-skel.hh)
    const ::test:LongArray  paramWrapper1( . );
    ::test:LongArray        paramWrapper2( . );
    ::test:LongArray_out    paramWrapper3( . );

--
TIA

Rich



http://my.yahoo.com.au - My Yahoo!
- It's My Yahoo! Get your own!

[orbitcpp-list] [ orbitcpp-Bugs-503269 ] generate skel code not legal C++

[<no...@so...>]

Bugs item #503269, was opened at 2002-01-14 03:01
You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=100646&aid=503269&group_id=646

Category: None
Group: None
Status: Open
Resolution: None
Priority: 5
Submitted By: Richard Andrews (bbmaj7)
Assigned to: Nobody/Anonymous (nobody)
Summary: generate skel code not legal C++

Initial Comment:

Let me present an example of what currently occurs.

This is a skel generated from the everything test

test_VariableLengthStruct * 
_skel_opVariable( ::PortableServer_Servant _servant,
const test_VariableLengthStruct * inArg,
test_VariableLengthStruct *inoutArg,
test_VariableLengthStruct **outArg, 
::CORBA_Environment *_ev) {
	::test::VariableLengthStruct *_retval = NULL;
	bool _results_valid = true;

	try {
		POA_test::StructServer * _self = ((_orbitcpp_Servant
*)_servant)->m_cppimpl;
		_retval = _self->opVariable( 
reinterpret_cast< const VariableLengthStruct&>(*inArg),
reinterpret_cast< VariableLengthStruct&>(*inoutArg),
reinterpret_cast< VariableLengthStruct*&>(*outArg) );
	}

I have removed namespace qualifications for clarity.

Now what happens here is that _self->opVariable expects
a Data_out<VariableLengthStruct> & (non-const ref) as
argument 3. Under C++ implied construction rules this
is no problem because Data_out<...> has a constructor
from VariableLengthStruct*&.

BUT... and here's the big problem.

This constructor creates a temporary object which is
passed as a reference through parameter 3, however only
const-references to temporaries may be passed in
standard C++ (as is now enforced in gcc3).

The only legal way to get around this problem is to
construct a wrapper object before the call, then pass
this as the non-const reference.

Eg.

try 
{ 
    POA_test::StructServer * _self =
((_orbitcpp_Servant *)_servant)->m_cppimpl; 

    Data_out< VariableLengthStruct> 
    wrapper3( reinterpret_cast<
VariableLengthStruct*&>(*outArg) );
    _retval = _self->opVariable( 
reinterpret_cast< const VariableLengthStruct&>(*inArg),
reinterpret_cast< VariableLengthStruct&>(*inoutArg), 
wrapper3); 

} 


In this case because wrapper3 is not a temporary it can
be passed through a non-const reference as is required
for argument 3 of opVariable().


In general the skeleton code for invoking operations
needs to be modified so that for each argument to the
servant implementation that is passed by non-const
reference, a temporary is constructed and passed.

I've been having a look at the code. The main areas of
interest are 

IDLPassSkels::doOperationSkel()

IDL_X_X_X_::getCPPSkelParameterTerm()

--

I put in a kludge in orbitcpp_smrtptr.hh about a year
ago using const_cast<>() to get around this problem but
it is no longer sufficient. The problem needs to be
fixed at the source which is the generated skel code.

I'm willing to put some time into fixing this if
someone familiar with this code can help me out.




----------------------------------------------------------------------

You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=100646&aid=503269&group_id=646

[orbitcpp-list] Re: cpp branch

[<mu...@t-...>]

Sam, I have applied your patches, though I needed to swap a ] and )
around in your conditional compilation patch, and add some #include
<string.h> lines for your C++ ORB patch. I really think that you should
get cvs write access and commit stuff directly.

Unfortunately, this seems to have made things worse:

[root@localhost ORBit2]# cd test
[root@localhost test]# orbit-idl-2 -lc++ empty.idl
orbit-idl 2.3.101 compiling
 small mode, show preprocessor errors, passes: stubs skels common
headers skel_impl imodule


(process:19818): ** CRITICAL **: file orbit-idl-driver.c: line 50
(orbit_idl_to_backend): assertion `binfo && binfo->op_output' failed

(process:19818): ** WARNING **: empty.idl compilation failed
[root@localhost test]# 


By the way, can you tell me something about this 'C++ ORB'. Why do we
need a separate ORB? I thought that Orbit/C++ just provided C++
stubs/skels which in turn used the C stubs/skels. Please forgive my
ignorance.


On Sat, 2002-01-05 at 11:25, Sam Couter wrote:
> Sam Couter <sa...@to...> wrote:
> > 2) Actually getting the C++ ORB to compile at all. No guarantees about
> > the code actually being useful. :)
> 
> Okay, the attached patch will make the C++ ORB build, but it won't be a
> useful ORB. Some of the internals have changed a fair bit, so the few
> problems I couldn't solve quickly got #if 0'd out of existence. Getting
> it to build is the first step towards making it work.
> -- 
> Sam "Eddie" Couter  |  mailto:sa...@to...     |  I need a short and
> Internet Engineer   |  jabber:edd...@ja...  |  clever comment for
> tSA Consulting      |  http://www.topic.com.au/    |  my .signature file
> OpenPGP fingerprint:  A46B 9BB5 3148 7BEA 1F05  5BD5 8530 03AE DE89 C75C
> ----
> 

> ? orb_cpp.diff
> ? test/test-mem
> Index: include/orbit/orb-core/corba-orb.h
> ===================================================================
> RCS file: /cvs/gnome/ORBit2/include/orbit/orb-core/corba-orb.h,v
> retrieving revision 1.11
> diff -u -r1.11 corba-orb.h
> --- include/orbit/orb-core/corba-orb.h	2001/11/21 10:46:18	1.11
> +++ include/orbit/orb-core/corba-orb.h	2002/01/05 10:20:57
> @@ -21,6 +21,8 @@
>  
>  void      ORBit_ORB_start_servers (CORBA_ORB orb);
>  
> +CORBA_TypeCode CORBA_TypeCode_allocate(void);
> +
>  #endif /* ORBIT2_INTERNAL_API */
>  
>  #endif
> Index: include/orbit/orb-core/corba-typecode.h
> ===================================================================
> RCS file: /cvs/gnome/ORBit2/include/orbit/orb-core/corba-typecode.h,v
> retrieving revision 1.13
> diff -u -r1.13 corba-typecode.h
> --- include/orbit/orb-core/corba-typecode.h	2001/11/26 01:42:14	1.13
> +++ include/orbit/orb-core/corba-typecode.h	2002/01/05 10:20:57
> @@ -4,6 +4,8 @@
>  #include <orbit/orb-core/corba-typecode-type.h>
>  #include <orbit/orb-core/corba-any-type.h>
>  #include <orbit/orb-core/orbit-object.h>
> +#include <orbit/orb-core/orb-types.h>
> +#include <orbit/GIOP/giop-basics.h>
>  
>  #define CORBA_tk_recursive -1
>  #define CORBA_tk_last (CORBA_tk_abstract_interface + 1)
> Index: src/Makefile.am
> ===================================================================
> RCS file: /cvs/gnome/ORBit2/src/Makefile.am,v
> retrieving revision 1.4
> diff -u -r1.4 Makefile.am
> --- src/Makefile.am	2000/11/17 01:15:44	1.4
> +++ src/Makefile.am	2002/01/05 10:20:57
> @@ -1,2 +1 @@
> -SUBDIRS=idl-compiler idl orb services
> -
> +SUBDIRS=idl-compiler idl orb orb-cpp services
> Index: src/orb/orb-core/corba-orb.c
> ===================================================================
> RCS file: /cvs/gnome/ORBit2/src/orb/orb-core/corba-orb.c,v
> retrieving revision 1.47
> diff -u -r1.47 corba-orb.c
> --- src/orb/orb-core/corba-orb.c	2001/12/05 13:58:13	1.47
> +++ src/orb/orb-core/corba-orb.c	2002/01/05 10:21:00
> @@ -378,7 +378,7 @@
>  	return ORBit_RootObject_duplicate (objref);
>  }
>  
> -static CORBA_TypeCode
> +CORBA_TypeCode
>  CORBA_TypeCode_allocate (void)
>  {
>  	CORBA_TypeCode tc = g_new0 (struct CORBA_TypeCode_struct, 1);
> Index: src/orb-cpp/Makefile.am
> ===================================================================
> RCS file: /cvs/gnome/ORBit2/src/orb-cpp/Attic/Makefile.am,v
> retrieving revision 1.1.2.2
> diff -u -r1.1.2.2 Makefile.am
> --- src/orb-cpp/Makefile.am	2001/12/11 09:39:14	1.1.2.2
> +++ src/orb-cpp/Makefile.am	2002/01/05 10:21:00
> @@ -1,7 +1,12 @@
> -## Process this file with automake to produce Makefile.in
> -
>  lib_LTLIBRARIES = liborbitcpp.la
>  
> +INCLUDES = \
> +	-I. \
> +	-I$(top_srcdir)/include \
> +	-DORBIT2_INTERNAL_API \
> +	$(WARN_CFLAGS) \
> +	$(ORBIT_CFLAGS)
> +
>  liborbitcppincludedir = $(includedir)/orb
>  
>  liborbitcppinclude_HEADERS = \
> @@ -33,5 +38,5 @@
>  			orbitcpp_typecode.cc
>  			$(liborbitcppinclude_HEADERS)
>  
> -liborbitcpp_la_LDFLAGS = -version-info $(LT_CURRENT):$(LT_REVISION):$(LT_AGE)
> +liborbitcpp_la_LDFLAGS = -version-info $(LT_VERSION)
>  liborbitcpp_la_LIBADD = $(ORBIT_LIBS) -lstdc++
> Index: src/orb-cpp/orbitcpp.hh
> ===================================================================
> RCS file: /cvs/gnome/ORBit2/src/orb-cpp/Attic/orbitcpp.hh,v
> retrieving revision 1.1.2.1
> diff -u -r1.1.2.1 orbitcpp.hh
> --- src/orb-cpp/orbitcpp.hh	2001/12/11 07:59:42	1.1.2.1
> +++ src/orb-cpp/orbitcpp.hh	2002/01/05 10:21:00
> @@ -32,19 +32,19 @@
>  
>  
>  
> -#include <orb/orbitcpp_constants.hh>
> -#include <orb/orbitcpp_types.hh>
> -#include <orb/orbitcpp_smartptr.hh>
> -#include <orb/orbitcpp_tools.hh>
> -#include <orb/orbitcpp_exception.hh>
> -#include <orb/orbitcpp_object.hh>
> -#include <orb/orbitcpp_orb.hh>
> -#include <orb/orbitcpp_sequence.hh>
> -#include <orb/orbitcpp_poa.hh>
> -#include <orb/orbitcpp_any.hh>
> +#include "orbitcpp_constants.hh"
> +#include "orbitcpp_types.hh"
> +#include "orbitcpp_smartptr.hh"
> +#include "orbitcpp_tools.hh"
> +#include "orbitcpp_exception.hh"
> +#include "orbitcpp_object.hh"
> +#include "orbitcpp_orb.hh"
> +#include "orbitcpp_sequence.hh"
> +#include "orbitcpp_poa.hh"
> +#include "orbitcpp_any.hh"
>  
>  // hopefully this will be removed when the object_var stuff gets
>  // sorted out.  - PD
> -#include <orb/orbitcpp_var_smartptr.hh>
> +#include "orbitcpp_var_smartptr.hh"
>  
>  #endif
> Index: src/orb-cpp/orbitcpp_any.cc
> ===================================================================
> RCS file: /cvs/gnome/ORBit2/src/orb-cpp/Attic/orbitcpp_any.cc,v
> retrieving revision 1.1.2.1
> diff -u -r1.1.2.1 orbitcpp_any.cc
> --- src/orb-cpp/orbitcpp_any.cc	2001/12/11 07:59:42	1.1.2.1
> +++ src/orb-cpp/orbitcpp_any.cc	2002/01/05 10:21:00
> @@ -24,20 +24,24 @@
>   *  Description: CORBA Any implementation
>   */
>  #include <cwchar>
> -#include <orb/orbit.h>
> -#include <orb/orbitcpp.hh>
> +#include <orbit/orbit.h>
> +#include "orbitcpp.hh"
>  #include "orbitcpp_any.hh"
>  
>  
>  /* ORBit_demarshal_allocate_mem cut & paste - this should ideally be
>     global */
>  
> +/* FIXME: This is sure to be broken for ORBit2. */
>  static gpointer
>  ORBit_demarshal_allocate_mem(CORBA_TypeCode tc, gint nelements)
>  {
> -    size_t block_size;
>      gpointer retval = NULL;
>  
> +	/* And it seems to be quite broken. */
> +#if 0
> +    size_t block_size;
> +
>      if(!nelements) return retval;
>  
>      block_size = ORBit_gather_alloc_info(tc);
> @@ -50,6 +54,7 @@
>  
>  	*(CORBA_TypeCode *)((char *)retval-sizeof(ORBit_mem_info)-sizeof(CORBA_TypeCode)) = (CORBA_TypeCode)CORBA_Object_duplicate((CORBA_Object)tc, NULL);
>      }
> +#endif
>  
>      return retval;
>  }
> Index: src/orb-cpp/orbitcpp_any.hh
> ===================================================================
> RCS file: /cvs/gnome/ORBit2/src/orb-cpp/Attic/orbitcpp_any.hh,v
> retrieving revision 1.1.2.1
> diff -u -r1.1.2.1 orbitcpp_any.hh
> --- src/orb-cpp/orbitcpp_any.hh	2001/12/11 07:59:42	1.1.2.1
> +++ src/orb-cpp/orbitcpp_any.hh	2002/01/05 10:21:00
> @@ -27,12 +27,11 @@
>  #define __ORBITCPP_ANY_HH
>  
>  #include <cstdlib>
> -#include <orb/orbit.h>
> -#include <orb/orbitcpp_types.hh>
> -#include <orb/orbitcpp_object.hh>
> -#include <orb/orbitcpp_exception.hh>
> -//#include <orb/corba_typecode.h>
> -#include <orb/orbitcpp_typecode.hh>
> +#include <orbit/orbit.h>
> +#include "orbitcpp_types.hh"
> +#include "orbitcpp_object.hh"
> +#include "orbitcpp_exception.hh"
> +#include "orbitcpp_typecode.hh"
>  
>  namespace CORBA {
>  	class Any;
> Index: src/orb-cpp/orbitcpp_exception.cc
> ===================================================================
> RCS file: /cvs/gnome/ORBit2/src/orb-cpp/Attic/orbitcpp_exception.cc,v
> retrieving revision 1.1.2.1
> diff -u -r1.1.2.1 orbitcpp_exception.cc
> --- src/orb-cpp/orbitcpp_exception.cc	2001/12/11 07:59:42	1.1.2.1
> +++ src/orb-cpp/orbitcpp_exception.cc	2002/01/05 10:21:00
> @@ -27,7 +27,7 @@
>  
>  
>  
> -#include <orb/orbitcpp_exception.hh>
> +#include "orbitcpp_exception.hh"
>  
>  
>  
> @@ -69,7 +69,7 @@
>  	const char *repo_id = CORBA_exception_id(&m_env);	
>  
>      #define P(name,quotedrepoid) ORBITCPP_SYSEX_THROW_O_MATIC(name,quotedrepoid)
> -	#include <orb/orbitcpp_exceptionlist.hh>
> +	#include "orbitcpp_exceptionlist.hh"
>      #undef P
>  	
>  	CORBA_exception_free(&m_env);
> Index: src/orb-cpp/orbitcpp_exception.hh
> ===================================================================
> RCS file: /cvs/gnome/ORBit2/src/orb-cpp/Attic/orbitcpp_exception.hh,v
> retrieving revision 1.1.2.1
> diff -u -r1.1.2.1 orbitcpp_exception.hh
> --- src/orb-cpp/orbitcpp_exception.hh	2001/12/11 07:59:42	1.1.2.1
> +++ src/orb-cpp/orbitcpp_exception.hh	2002/01/05 10:21:01
> @@ -33,8 +33,8 @@
>  
>  
>  
> -#include <orb/orbitcpp_types.hh>
> -#include <orb/orbitcpp_tools.hh>
> +#include "orbitcpp_types.hh"
> +#include "orbitcpp_tools.hh"
>  
>  
>  
> @@ -160,7 +160,7 @@
>  	};
>  	
>  	#define P(name,quotedname) ORBITCPP_SYSEXCEPTION_FACTORY(name,quotedname)
> -	#include <orb/orbitcpp_exceptionlist.hh>
> +	#include "orbitcpp_exceptionlist.hh"
>  	#undef P
>  
>  	class UserException : public Exception {
> Index: src/orb-cpp/orbitcpp_object.cc
> ===================================================================
> RCS file: /cvs/gnome/ORBit2/src/orb-cpp/Attic/orbitcpp_object.cc,v
> retrieving revision 1.1.2.1
> diff -u -r1.1.2.1 orbitcpp_object.cc
> --- src/orb-cpp/orbitcpp_object.cc	2001/12/11 07:59:42	1.1.2.1
> +++ src/orb-cpp/orbitcpp_object.cc	2002/01/05 10:21:01
> @@ -27,8 +27,8 @@
>  
>  
>  
> -#include <orb/orbitcpp_object.hh>
> -#include <orb/orbitcpp_exception.hh>
> +#include "orbitcpp_object.hh"
> +#include "orbitcpp_exception.hh"
>  
>  
>  
> Index: src/orb-cpp/orbitcpp_object.hh
> ===================================================================
> RCS file: /cvs/gnome/ORBit2/src/orb-cpp/Attic/orbitcpp_object.hh,v
> retrieving revision 1.1.2.1
> diff -u -r1.1.2.1 orbitcpp_object.hh
> --- src/orb-cpp/orbitcpp_object.hh	2001/12/11 07:59:42	1.1.2.1
> +++ src/orb-cpp/orbitcpp_object.hh	2002/01/05 10:21:01
> @@ -29,9 +29,9 @@
>  
>  
>  
> -#include <orb/orbitcpp_types.hh>
> -#include <orb/orbitcpp_tools.hh>
> -#include <orb/orbitcpp_smartptr.hh>
> +#include "orbitcpp_types.hh"
> +#include "orbitcpp_tools.hh"
> +#include "orbitcpp_smartptr.hh"
>  
>  
>  
> @@ -61,7 +61,7 @@
>  		friend void release(Object_ptr o);
>  	
>  	protected:
> -		CORBA_Object_struct m_target;
> +		CORBA_Object_type m_target;
>  		
>  	public:
>  		// begin ORBit-C++ extension
> Index: src/orb-cpp/orbitcpp_orb.cc
> ===================================================================
> RCS file: /cvs/gnome/ORBit2/src/orb-cpp/Attic/orbitcpp_orb.cc,v
> retrieving revision 1.1.2.1
> diff -u -r1.1.2.1 orbitcpp_orb.cc
> --- src/orb-cpp/orbitcpp_orb.cc	2001/12/11 07:59:42	1.1.2.1
> +++ src/orb-cpp/orbitcpp_orb.cc	2002/01/05 10:21:01
> @@ -27,9 +27,9 @@
>  
>  
>  #include <iostream>
> -#include <orb/orbitcpp_object.hh>
> -#include <orb/orbitcpp_orb.hh>
> -#include <orb/orbitcpp_exception.hh>
> +#include "orbitcpp_object.hh"
> +#include "orbitcpp_orb.hh"
> +#include "orbitcpp_exception.hh"
>  
>  
>  
> Index: src/orb-cpp/orbitcpp_orb.hh
> ===================================================================
> RCS file: /cvs/gnome/ORBit2/src/orb-cpp/Attic/orbitcpp_orb.hh,v
> retrieving revision 1.1.2.1
> diff -u -r1.1.2.1 orbitcpp_orb.hh
> --- src/orb-cpp/orbitcpp_orb.hh	2001/12/11 07:59:42	1.1.2.1
> +++ src/orb-cpp/orbitcpp_orb.hh	2002/01/05 10:21:01
> @@ -32,7 +32,7 @@
>  
>  
>  
> -#include <orb/orbitcpp_object.hh>
> +#include "orbitcpp_object.hh"
>  
>  namespace CORBA {
>  	CORBA::ORB_ptr ORB_init(int& argc, char** argv,
> Index: src/orb-cpp/orbitcpp_poa.cc
> ===================================================================
> RCS file: /cvs/gnome/ORBit2/src/orb-cpp/Attic/orbitcpp_poa.cc,v
> retrieving revision 1.1.2.1
> diff -u -r1.1.2.1 orbitcpp_poa.cc
> --- src/orb-cpp/orbitcpp_poa.cc	2001/12/11 07:59:42	1.1.2.1
> +++ src/orb-cpp/orbitcpp_poa.cc	2002/01/05 10:21:04
> @@ -27,7 +27,7 @@
>  
>  
>  
> -#include <orb/orbitcpp_poa.hh>
> +#include "orbitcpp_poa.hh"
>  
>  
>  
> Index: src/orb-cpp/orbitcpp_poa.hh
> ===================================================================
> RCS file: /cvs/gnome/ORBit2/src/orb-cpp/Attic/orbitcpp_poa.hh,v
> retrieving revision 1.1.2.1
> diff -u -r1.1.2.1 orbitcpp_poa.hh
> --- src/orb-cpp/orbitcpp_poa.hh	2001/12/11 07:59:42	1.1.2.1
> +++ src/orb-cpp/orbitcpp_poa.hh	2002/01/05 10:21:04
> @@ -34,19 +34,18 @@
>  
>  
>  
> -#include <orb/orbitcpp_smartptr.hh>
> -#include <orb/orbitcpp_orb.hh>
> -#include <orb/orbitcpp_object.hh>
> -#include <orb/orbitcpp_exception.hh>
> -#include <orb/orbitcpp_sequence.hh>
> +#include "orbitcpp_smartptr.hh"
> +#include "orbitcpp_orb.hh"
> +#include "orbitcpp_object.hh"
> +#include "orbitcpp_exception.hh"
> +#include "orbitcpp_sequence.hh"
>  
>  
>  
> -
>  namespace PortableServer {
>  	ORBITCPP_DECLARE_SIMPLE_SEQUENCE(CORBA::Octet,
> -		CORBA_sequence_octet__alloc,
> -		CORBA_octet_allocbuf,
> +		CORBA_sequence_CORBA_octet__alloc,
> +		CORBA_sequence_CORBA_octet_allocbuf,
>  		PortableServer_ObjectId,
>  		ObjectId)
>  
> Index: src/orb-cpp/orbitcpp_sequence.hh
> ===================================================================
> RCS file: /cvs/gnome/ORBit2/src/orb-cpp/Attic/orbitcpp_sequence.hh,v
> retrieving revision 1.1.2.1
> diff -u -r1.1.2.1 orbitcpp_sequence.hh
> --- src/orb-cpp/orbitcpp_sequence.hh	2001/12/11 07:59:42	1.1.2.1
> +++ src/orb-cpp/orbitcpp_sequence.hh	2002/01/05 10:21:04
> @@ -31,7 +31,8 @@
>  
>  #include <cstddef>
>  #include <cassert>
> -#include <orb/orbitcpp_smartptr.hh>
> +#include "orbitcpp_smartptr.hh"
> +#include <orbit/orb-core/orb-core.h>
>  
>  
>  
> Index: src/orb-cpp/orbitcpp_smartptr.hh
> ===================================================================
> RCS file: /cvs/gnome/ORBit2/src/orb-cpp/Attic/orbitcpp_smartptr.hh,v
> retrieving revision 1.1.2.1
> diff -u -r1.1.2.1 orbitcpp_smartptr.hh
> --- src/orb-cpp/orbitcpp_smartptr.hh	2001/12/11 07:59:42	1.1.2.1
> +++ src/orb-cpp/orbitcpp_smartptr.hh	2002/01/05 10:21:04
> @@ -33,8 +33,8 @@
>  
>  
>  
> -#include <orb/orbitcpp_types.hh>
> -#include <orb/orbitcpp_tools.hh>
> +#include "orbitcpp_types.hh"
> +#include "orbitcpp_tools.hh"
>  
>  namespace CORBA {
>  	class Object;
> Index: src/orb-cpp/orbitcpp_tools.cc
> ===================================================================
> RCS file: /cvs/gnome/ORBit2/src/orb-cpp/Attic/orbitcpp_tools.cc,v
> retrieving revision 1.1.2.1
> diff -u -r1.1.2.1 orbitcpp_tools.cc
> --- src/orb-cpp/orbitcpp_tools.cc	2001/12/11 07:59:42	1.1.2.1
> +++ src/orb-cpp/orbitcpp_tools.cc	2002/01/05 10:21:06
> @@ -30,8 +30,9 @@
>  
>  #include <cstdio>
>  #include <cstdlib>
> -#include <orb/orbitcpp_tools.hh>
> -#include <orb/orbitcpp_exception.hh>
> +#include "orbitcpp_tools.hh"
> +#include "orbitcpp_exception.hh"
> +#include <orbit/orb-core/corba-orb.h>
>  
>  
>  
> @@ -64,10 +65,7 @@
>  }
>  
>  CORBA_TypeCode _orbitcpp::TypeCode_allocate() {
> -	CORBA_TypeCode tc = g_new0(struct CORBA_TypeCode_struct,1);
> -	CEnvironment ev;
> -	ORBit_pseudo_object_init(&tc->parent,ORBIT_PSEUDO_TYPECODE,ev);
> -	ev.propagate_sysex();
> +	CORBA_TypeCode tc = CORBA_TypeCode_allocate();
>  	return tc;
>  }
>  
> Index: src/orb-cpp/orbitcpp_tools.hh
> ===================================================================
> RCS file: /cvs/gnome/ORBit2/src/orb-cpp/Attic/orbitcpp_tools.hh,v
> retrieving revision 1.1.2.1
> diff -u -r1.1.2.1 orbitcpp_tools.hh
> --- src/orb-cpp/orbitcpp_tools.hh	2001/12/11 07:59:42	1.1.2.1
> +++ src/orb-cpp/orbitcpp_tools.hh	2002/01/05 10:21:06
> @@ -30,10 +30,10 @@
>  #define __ORBITCPP_TOOLS_HH
>  
>  
> -#include <orb/orbit.h>
> +#include <orbit/orbit.h>
>  
> -#include <orb/orbitcpp_config.hh>
> -#include <orb/orbitcpp_types.hh>
> +#include "orbitcpp_config.hh"
> +#include "orbitcpp_types.hh"
>  
>  
>  
> Index: src/orb-cpp/orbitcpp_typecode.cc
> ===================================================================
> RCS file: /cvs/gnome/ORBit2/src/orb-cpp/Attic/orbitcpp_typecode.cc,v
> retrieving revision 1.1.2.1
> diff -u -r1.1.2.1 orbitcpp_typecode.cc
> --- src/orb-cpp/orbitcpp_typecode.cc	2001/12/11 07:59:42	1.1.2.1
> +++ src/orb-cpp/orbitcpp_typecode.cc	2002/01/05 10:21:06
> @@ -22,53 +22,59 @@
>   *  Author: Sam Couter <sa...@to...>
>   */
>  
> -#include <orb/orbitcpp_typecode.hh>
> -#include <orb/orbitcpp_exception.hh>
> +#include "orbitcpp_typecode.hh"
> +#include "orbitcpp_exception.hh"
> +#include <orbit/orb-core/corba-typecode.h>
>  
> +
>  using namespace _orbitcpp;
>  
> +
>  CORBA::TypeCode_ptr CORBA::_tc_null =
>  	reinterpret_cast <CORBA::TypeCode_ptr> (TC_null);
>  CORBA::TypeCode_ptr CORBA::_tc_void =
>  	reinterpret_cast <CORBA::TypeCode_ptr> (TC_void);
>  CORBA::TypeCode_ptr CORBA::_tc_short =
> -	reinterpret_cast <CORBA::TypeCode_ptr> (TC_short);
> +	reinterpret_cast <CORBA::TypeCode_ptr> (TC_CORBA_short);
>  CORBA::TypeCode_ptr CORBA::_tc_long =
> -	reinterpret_cast <CORBA::TypeCode_ptr> (TC_long);
> +	reinterpret_cast <CORBA::TypeCode_ptr> (TC_CORBA_long);
>  CORBA::TypeCode_ptr CORBA::_tc_longlong =
> -	reinterpret_cast <CORBA::TypeCode_ptr> (TC_longlong);
> +	reinterpret_cast <CORBA::TypeCode_ptr> (TC_CORBA_long_long);
>  CORBA::TypeCode_ptr CORBA::_tc_ushort =
> -	reinterpret_cast <CORBA::TypeCode_ptr> (TC_ushort);
> +	reinterpret_cast <CORBA::TypeCode_ptr> (TC_CORBA_unsigned_short);
>  CORBA::TypeCode_ptr CORBA::_tc_ulong =
> -	reinterpret_cast <CORBA::TypeCode_ptr> (TC_ulong);
> +	reinterpret_cast <CORBA::TypeCode_ptr> (TC_CORBA_unsigned_long);
>  CORBA::TypeCode_ptr CORBA::_tc_ulonglong =
> -	reinterpret_cast <CORBA::TypeCode_ptr> (TC_ulonglong);
> +	reinterpret_cast <CORBA::TypeCode_ptr> (TC_CORBA_unsigned_long_long);
>  CORBA::TypeCode_ptr CORBA::_tc_float =
> -	reinterpret_cast <CORBA::TypeCode_ptr> (TC_float);
> +	reinterpret_cast <CORBA::TypeCode_ptr> (TC_CORBA_float);
>  CORBA::TypeCode_ptr CORBA::_tc_double =
> -	reinterpret_cast <CORBA::TypeCode_ptr> (TC_double);
> +	reinterpret_cast <CORBA::TypeCode_ptr> (TC_CORBA_double);
>  CORBA::TypeCode_ptr CORBA::_tc_longdouble =
> -	reinterpret_cast <CORBA::TypeCode_ptr> (TC_longdouble);
> +	reinterpret_cast <CORBA::TypeCode_ptr> (TC_CORBA_long_double);
>  CORBA::TypeCode_ptr CORBA::_tc_boolean =
> -	reinterpret_cast <CORBA::TypeCode_ptr> (TC_boolean);
> +	reinterpret_cast <CORBA::TypeCode_ptr> (TC_CORBA_boolean);
>  CORBA::TypeCode_ptr CORBA::_tc_char =
> -	reinterpret_cast <CORBA::TypeCode_ptr> (TC_char);
> +	reinterpret_cast <CORBA::TypeCode_ptr> (TC_CORBA_char);
>  CORBA::TypeCode_ptr CORBA::_tc_wchar =
> -	reinterpret_cast <CORBA::TypeCode_ptr> (TC_wchar);
> +	reinterpret_cast <CORBA::TypeCode_ptr> (TC_CORBA_wchar);
>  CORBA::TypeCode_ptr CORBA::_tc_octet =
> -	reinterpret_cast <CORBA::TypeCode_ptr> (TC_octet);
> +	reinterpret_cast <CORBA::TypeCode_ptr> (TC_CORBA_octet);
>  CORBA::TypeCode_ptr CORBA::_tc_any =
> -	reinterpret_cast <CORBA::TypeCode_ptr> (TC_any);
> +	reinterpret_cast <CORBA::TypeCode_ptr> (TC_CORBA_any);
>  CORBA::TypeCode_ptr CORBA::_tc_TypeCode =
> -	reinterpret_cast <CORBA::TypeCode_ptr> (TC_TypeCode);
> +	reinterpret_cast <CORBA::TypeCode_ptr> (TC_CORBA_TypeCode);
> +/* There doesn't seem to be a TypeCode defined for CORBA_Principal yet... */
> +#if 0
>  CORBA::TypeCode_ptr CORBA::_tc_Principal =
> -	reinterpret_cast <CORBA::TypeCode_ptr> (TC_Principal);
> +	reinterpret_cast <CORBA::TypeCode_ptr> (TC_CORBA_Principal);
> +#endif
>  CORBA::TypeCode_ptr CORBA::_tc_Object =
> -	reinterpret_cast <CORBA::TypeCode_ptr> (TC_Object);
> +	reinterpret_cast <CORBA::TypeCode_ptr> (TC_CORBA_Object);
>  CORBA::TypeCode_ptr CORBA::_tc_string =
> -	reinterpret_cast <CORBA::TypeCode_ptr> (TC_string);
> +	reinterpret_cast <CORBA::TypeCode_ptr> (TC_CORBA_string);
>  CORBA::TypeCode_ptr CORBA::_tc_wstring =
> -	reinterpret_cast <CORBA::TypeCode_ptr> (TC_wstring);
> +	reinterpret_cast <CORBA::TypeCode_ptr> (TC_CORBA_wstring);
>  
>  CORBA::Boolean
>  CORBA::TypeCode::equal(TypeCode_ptr tc) const
> @@ -290,7 +296,7 @@
>  CORBA::ValueModifier
>  CORBA::TypeCode::type_modifier() const
>  {
> -	CORBA::ValueModifier retval;
> +	CORBA::ValueModifier retval = 0;
>  	CEnvironment ev;
>  
>  	/* not implemented by ORBit */
> Index: src/orb-cpp/orbitcpp_typecode.hh
> ===================================================================
> RCS file: /cvs/gnome/ORBit2/src/orb-cpp/Attic/orbitcpp_typecode.hh,v
> retrieving revision 1.1.2.1
> diff -u -r1.1.2.1 orbitcpp_typecode.hh
> --- src/orb-cpp/orbitcpp_typecode.hh	2001/12/11 07:59:42	1.1.2.1
> +++ src/orb-cpp/orbitcpp_typecode.hh	2002/01/05 10:21:06
> @@ -25,10 +25,10 @@
>  #ifndef __ORBITCPP_TYPECODE_HH
>  #define __ORBITCPP_TYPECODE_HH
>  
> -#include <orb/corba_typecode.h>
> -#include <orb/orbitcpp_object.hh>
> -#include <orb/orbitcpp_exception.hh>
> -#include <orb/orbitcpp_sequence.hh>
> +#include <orbit/orb-core/corba-typecode.h>
> +#include "orbitcpp_object.hh"
> +#include "orbitcpp_exception.hh"
> +#include "orbitcpp_sequence.hh"
>  
>  namespace CORBA {
>  
> Index: src/orb-cpp/orbitcpp_types.cc
> ===================================================================
> RCS file: /cvs/gnome/ORBit2/src/orb-cpp/Attic/orbitcpp_types.cc,v
> retrieving revision 1.1.2.1
> diff -u -r1.1.2.1 orbitcpp_types.cc
> --- src/orb-cpp/orbitcpp_types.cc	2001/12/11 07:59:42	1.1.2.1
> +++ src/orb-cpp/orbitcpp_types.cc	2002/01/05 10:21:06
> @@ -26,7 +26,7 @@
>  
>  
>  
> -#include <orb/orbitcpp_types.hh>
> +#include "orbitcpp_types.hh"
>  
>  
>  
> Index: src/orb-cpp/orbitcpp_types.hh
> ===================================================================
> RCS file: /cvs/gnome/ORBit2/src/orb-cpp/Attic/orbitcpp_types.hh,v
> retrieving revision 1.1.2.1
> diff -u -r1.1.2.1 orbitcpp_types.hh
> --- src/orb-cpp/orbitcpp_types.hh	2001/12/11 07:59:42	1.1.2.1
> +++ src/orb-cpp/orbitcpp_types.hh	2002/01/05 10:21:06
> @@ -27,7 +27,7 @@
>  #ifndef __ORBITCPP_BASICTYPES_HH
>  #define __ORBITCPP_BASICTYPES_HH
>  
> -#include <orb/orbit.h>
> +#include <orbit/orbit.h>
>  
>  
>  
-- 
Murray Cumming
mu...@us...
www.murrayc.com

Re: [orbitcpp-list] Re: cpp branch

[<mu...@t-...>]

On Sat, 2002-01-05 at 06:06, Sam Couter wrote:
> Michael Meeks <mi...@xi...> wrote:
> > 
> >         When I last tried this invocation produces just empty files; so
> > ... something is bust somewhere, I was hoping Sam would find what it was.
> > Either way, any fixes there would be greatly appreciated; feel free to
> > commit any fixes, and send a patch here as you do.
> 
> I found the cause of this problem: The C++ backend doesn't know how to
> handle the IDLN_SRCFILE IDL type, hence the error about something being
> unexpected. The error would be slightly more helpful if the C++ backend
> also knew how to get the correct type name for a node from libIDL.

I've already fixed this problem. Maybe you could revise your patch.

-- 
Murray Cumming
mu...@us...
www.murrayc.com

[orbitcpp-list] Re: cpp branch

[Sam C. <sa...@to...>]

Sam Couter <sa...@to...> wrote:
> 2) Actually getting the C++ ORB to compile at all. No guarantees about
> the code actually being useful. :)

Okay, the attached patch will make the C++ ORB build, but it won't be a
useful ORB. Some of the internals have changed a fair bit, so the few
problems I couldn't solve quickly got #if 0'd out of existence. Getting
it to build is the first step towards making it work.
-- 
Sam "Eddie" Couter  |  mailto:sa...@to...     |  I need a short and
Internet Engineer   |  jabber:edd...@ja...  |  clever comment for
tSA Consulting      |  http://www.topic.com.au/    |  my .signature file
OpenPGP fingerprint:  A46B 9BB5 3148 7BEA 1F05  5BD5 8530 03AE DE89 C75C

[orbitcpp-list] Re: cpp branch

[Sam C. <sa...@to...>]

[ Sorry for the reply to my own message ]

Sam Couter <sa...@to...> wrote:
> 1) Conditional compilation of the C++ compiler and ORB. This is pretty
> simple, and just needs tidying up.

Patch is attached.
-- 
Sam "Eddie" Couter  |  mailto:sa...@to...     |  I need a short and
Internet Engineer   |  jabber:edd...@ja...  |  clever comment for
tSA Consulting      |  http://www.topic.com.au/    |  my .signature file
OpenPGP fingerprint:  A46B 9BB5 3148 7BEA 1F05  5BD5 8530 03AE DE89 C75C

[orbitcpp-list] Re: cpp branch

[Sam C. <sa...@to...>]

Michael Meeks <mi...@xi...> wrote:
> 
>         When I last tried this invocation produces just empty files; so
> ... something is bust somewhere, I was hoping Sam would find what it was.
> Either way, any fixes there would be greatly appreciated; feel free to
> commit any fixes, and send a patch here as you do.

I found the cause of this problem: The C++ backend doesn't know how to
handle the IDLN_SRCFILE IDL type, hence the error about something being
unexpected. The error would be slightly more helpful if the C++ backend
also knew how to get the correct type name for a node from libIDL.

Anyway, I'll attach a patch to fix this particular problem, as well as
fixing the way the backend grabs the node type name, so error messages
will be more meaningful.

When (if?) the backend is re-written in C, maybe it should be made
intelligent enough to warn about IDL types it doesn't know about, rather
than bombing out? Or could that just hide problems further down the
track?

In the mean time, I'm working on two things:
1) Conditional compilation of the C++ compiler and ORB. This is pretty
simple, and just needs tidying up.
2) Actually getting the C++ ORB to compile at all. No guarantees about
the code actually being useful. :)
-- 
Sam "Eddie" Couter  |  mailto:sa...@to...     |  I need a short and
Internet Engineer   |  jabber:edd...@ja...  |  clever comment for
tSA Consulting      |  http://www.topic.com.au/    |  my .signature file
OpenPGP fingerprint:  A46B 9BB5 3148 7BEA 1F05  5BD5 8530 03AE DE89 C75C

[orbitcpp-list] [ orbitcpp-Patches-498857 ] Love orbitcpp_constants.hh

[<no...@so...>]

Patches item #498857, was opened at 2002-01-02 23:10
You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=300646&aid=498857&group_id=646

Category: None
Group: None
>Status: Closed
>Resolution: Accepted
Priority: 5
Submitted By: Braden McDaniel (braden)
>Assigned to: Sam Couter (eddiesam)
Summary: Love orbitcpp_constants.hh

Initial Comment:
This patch does the following for/to
"orbitcpp_constants.hh":

 * Uses preprocessor definitions (defined in configure)
for the version numbers, which allows
"orbitcpp_constants.hh" be a regular header instead of
being generated by configure. This means that the file
doesn't get touched every time config.status is run,
which in turn means that there are some things that
nolonger get rebuilt unnecessarily.

 * The contents of "orbitcpp_constants.hh" are all in a
private namespace, so there's no need for it to be a
public header. Thus, it is moved to the
"noinst_HEADERS" variable in the Makefile.am so that it
will not be installed.

"orbitcpp_constants.patch" contains the unified diffs
for the modified files.

"orbitcpp_constants.hh" is a new file that should be
placed in the "orb" subdirectory.

This patch obsoletes the file
"orb/orbitcpp_constants.hh.in".


----------------------------------------------------------------------

You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=300646&aid=498857&group_id=646

[orbitcpp-list] [ orbitcpp-Patches-498857 ] Love orbitcpp_constants.hh

[<no...@so...>]

Patches item #498857, was opened at 2002-01-02 23:10
You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=300646&aid=498857&group_id=646

Category: None
Group: None
Status: Open
Resolution: None
Priority: 5
Submitted By: Braden McDaniel (braden)
Assigned to: Nobody/Anonymous (nobody)
Summary: Love orbitcpp_constants.hh

Initial Comment:
This patch does the following for/to
"orbitcpp_constants.hh":

 * Uses preprocessor definitions (defined in configure)
for the version numbers, which allows
"orbitcpp_constants.hh" be a regular header instead of
being generated by configure. This means that the file
doesn't get touched every time config.status is run,
which in turn means that there are some things that
nolonger get rebuilt unnecessarily.

 * The contents of "orbitcpp_constants.hh" are all in a
private namespace, so there's no need for it to be a
public header. Thus, it is moved to the
"noinst_HEADERS" variable in the Makefile.am so that it
will not be installed.

"orbitcpp_constants.patch" contains the unified diffs
for the modified files.

"orbitcpp_constants.hh" is a new file that should be
placed in the "orb" subdirectory.

This patch obsoletes the file
"orb/orbitcpp_constants.hh.in".


----------------------------------------------------------------------

You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=300646&aid=498857&group_id=646

[orbitcpp-list] [ orbitcpp-Patches-498431 ] Remove "disable-debug" configure arg

[<no...@so...>]

Patches item #498431, was opened at 2002-01-01 21:10
You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=300646&aid=498431&group_id=646

Category: None
Group: None
>Status: Closed
>Resolution: Accepted
Priority: 5
Submitted By: Braden McDaniel (braden)
>Assigned to: Sam Couter (eddiesam)
>Summary: Remove "disable-debug" configure arg

Initial Comment:
This patch removes the "disable-debug" arg from configure.

This is actually something that should have been
included in my previous patch, but I overlooked it.
Users who want compiler flags other than the default
("-g -O2") can provide them by setting CXXFLAGS and
CFLAGS at configure time, so this option isn't necessary.


----------------------------------------------------------------------

You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=300646&aid=498431&group_id=646

[orbitcpp-list] [ orbitcpp-Patches-498431 ] Remove "disable-debug" configure arg

[<no...@so...>]

Patches item #498431, was opened at 2002-01-01 21:10
You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=300646&aid=498431&group_id=646

Category: None
Group: None
Status: Open
Resolution: None
Priority: 5
Submitted By: Braden McDaniel (braden)
Assigned to: Nobody/Anonymous (nobody)
Summary: Remove "disable-debug" configure arg

Initial Comment:
This patch removes the "disable-debug" arg from configure.

This is actually something that should have been
included in my previous patch, but I overlooked it.
Users who want compiler flags other than the default
("-g -O2") can provide them by setting CXXFLAGS and
CFLAGS at configure time, so this option isn't necessary.


----------------------------------------------------------------------

You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=300646&aid=498431&group_id=646