orbitcpp-list

[orbitcpp-list] iiop example wanted

[David W. <dw...@sy...>]

Hi.  I am just starting to work with corba. I am using orbitcpp 0.28.1
(current source version on the web with a debian woody installation of
orbit.)


The application I am working on requires communication between
different machines over the Internet.  My understanding is that in
order to create the connection between the client and the server iiop
is the natural method.

I  am not clear from the  material that I have   read either michi and
steve's  book  or omg  docs  whether there  is a  local requirement to
connect to a name server which  provides the service of "directing the
connection from the local host to the  remote name server or orb" then
to the remote server,  or  if there is   a standardized protocol   for
dynamically requesting and establishing a connection.

I am guessing that there  is a step missing in  the model I understand
so far, unless  iiop  invisibly handles the  underlying  socket/server
communication and connection. Below is a generic understanding of what
I think is happening for  a message to be  passed from a client to the
server.

I am missing an  understanding of how the  initial ior is found by the
client.  That is how  does the client establish  the connection to the
remote server?  Does  the server publish  itself to a name  server via
it's local name server(called remote in the diagram below), which then
maintains  a permanent published   address at a  port  for  iiop based
`consumer' clients to find?



                 [iiop methods]                    [server]
                 +-----------+                        ^
                 |           |                        |
[client] -> [local orb]  [remote name server] -> [remote orb] 
                     
So I guess the questions that I have are:

1. Who is responsible for providing the initial connection information
from a client to a server over the internet?
        a. name servers 

        b. another middleware application which is done on an as
        needed basis

        c. iiop calls which (either through name-servers or each
        systems local orbs manages the connections transparently)



What  I would like is a  simple example  which demonstrates creating a
connection to  a   remote  host (like echo-*   for   iiop) and  allows
bidirectional communication between systems.

Is there a working example either in C++ or in  C which can be used to
`grok the source   luke' or if  there  is documentation  which details
this, I have missed it, my apologies, and pointers welcome.

(I  could cross post to comp.*corba  as well but  wonder if in general
they will know any issues related to orbit* features in development --
so I ask here. apologies if this is inappropriate.)


Thanks so much for any advice / pointers.

David

Re: [orbitcpp-list] ./configure problem with Orbitcpp 0.28.1

[Mike B. <mb...@co...>]

Or you can temporarily remove the lines in configure that look like:

#ifdef __cplusplus
extern "C" void exit(int);
#endif

I believe in 0.28.1 there were lines 2535-2537, but not sure, using CVS.

On Thu, Sep 14, 2000 at 12:56:42PM +0100, Phil Dawes wrote:
> Hi Knut,
> 
> This will be fixed in the upcoming 0.29 release, which has been
> progressing slowly because I'm suffering from a spot of RSI :-(
> 
> This should be out in a week or so, all being well. In the meantime
> I'm afraid you must use gcc 2.95.2 (even with the cvs version)
> 
> Cheers,
> 
> Phil
> 
> 
> Knut Olav Fjellheim writes:
>  > I tried to install Orbitcpp v0.28.1 but I get the following error message when I
>  > execute ./configure
>  > 
>  > [lots of messages, etc..]
>  > configure: error: Could not compile c++ new block size length test program
>  > 
>  > I am using gcc v 2.95.3  19991030 (prerelease). Anybody know what I need to do
>  > in order to make the project compile?
>  > 
>  > Knut Fjellheim
>  > _______________________________________________
>  > orbitcpp-list mailing list
>  > orb...@li...
>  > http://lists.sourceforge.net/mailman/listinfo/orbitcpp-list
>  > 
> 
> _______________________________________________
> orbitcpp-list mailing list
> orb...@li...
> http://lists.sourceforge.net/mailman/listinfo/orbitcpp-list

-- 
TTFN
MikeB

Re: [orbitcpp-list] ./configure problem with Orbitcpp 0.28.1

[Phil D. <ph...@us...>]

Hi Knut,

This will be fixed in the upcoming 0.29 release, which has been
progressing slowly because I'm suffering from a spot of RSI :-(

This should be out in a week or so, all being well. In the meantime
I'm afraid you must use gcc 2.95.2 (even with the cvs version)

Cheers,

Phil


Knut Olav Fjellheim writes:
 > I tried to install Orbitcpp v0.28.1 but I get the following error message when I
 > execute ./configure
 > 
 > [lots of messages, etc..]
 > configure: error: Could not compile c++ new block size length test program
 > 
 > I am using gcc v 2.95.3  19991030 (prerelease). Anybody know what I need to do
 > in order to make the project compile?
 > 
 > Knut Fjellheim
 > _______________________________________________
 > orbitcpp-list mailing list
 > orb...@li...
 > http://lists.sourceforge.net/mailman/listinfo/orbitcpp-list
 > 

[orbitcpp-list] ./configure problem with Orbitcpp 0.28.1

[Knut O. F. <knu...@pr...>]

I tried to install Orbitcpp v0.28.1 but I get the following error message when I
execute ./configure

[lots of messages, etc..]
configure: error: Could not compile c++ new block size length test program

I am using gcc v 2.95.3  19991030 (prerelease). Anybody know what I need to do
in order to make the project compile?

Knut Fjellheim

Re: [orbitcpp-list] Question clarified...

[Michael R. <mi...@ru...>]

Hi,

Mike Bond wrote:
> 
> I just realized my question on mapping C++ data to corba doesn't make
> complete sense. What I want to be able to do is pass a C++ object by
> value as a parameter to an interface method. Is this possible using
> ORBit C++?

No, not yet (?). The feature is called Object By Value (OBV) and has been 
introduced with CORBA 2.3. ORBit-C++ is a wrapper around ORBit (implemented in C) 
which is (mostly) aiming at CORBA 2.2. The reason for this is that the C-language 
mapping has not been aligned to CORBA 2.3 yet. 
I heard that there is currently some OMG discussion going on to align 
the C-language mapping to CORBA 2.x (x>=3) but there are no official statements yet.

Meanwhile, you must live without this feature, 
or make a reference implementation for C which can be used as input to the OMG.... ;-)

Michael

> 
> --
> TTFN
> MikeB
> _______________________________________________
> orbitcpp-list mailing list
> orb...@li...
> http://lists.sourceforge.net/mailman/listinfo/orbitcpp-list

[orbitcpp-list] Question clarified...

[Mike B. <mb...@co...>]

I just realized my question on mapping C++ data to corba doesn't make
complete sense. What I want to be able to do is pass a C++ object by
value as a parameter to an interface method. Is this possible using
ORBit C++?

-- 
TTFN
MikeB

[orbitcpp-list] Beginner questions...

[Mike B. <mb...@co...>]

I'm relatively new to corba in general, but not to network programming. Some
of my questions may be generic corba questions more than ORBit C++ questions,
if so I apologize.

First and most important to me at this time. Is it currently possible, using
ORBit C++ or some other mechanism, to map existing C++ classes directly to
a corba class or struct such that it does not have to be copied. If so, how
do I go about doing this.

This one is a generic corba question, but I don't really know anywhere else
to ask it at the moment. Is it possible to issue asynchronous calls using
corba, or do I have to make use of threads to do this?

Is all or parts of ORBit C++ threadsafe?

-- 
TTFN
MikeB

Re: [orbitcpp-list] installing orbitcpp

[Kelvin L. <kel...@ma...>]

I tried that after un-tarring the archived file ORBitcpp0.28.1 that's the lastest version I think. And they gave me an error warning. 
That goes something like this:
IDL.h (line 66) cannot find 64 bit integer or not avaliable, using 32 bit

And after that it just aborted. 
However, I managed to install the ORBit 0.5.1 patched for ORBit-cpp with no warnings or errors.
--

On Sun, 20 Aug 2000 14:17:38   Phil Dawes wrote:
>
>Hi Kelvin,
>
>The usual method with gnu software is:
>
>./configure
>make
>make install
>
>Check out the INSTALL file for details.
>
>Cheers,
>
>Phil.
>
>Kelvin Lim writes:
> > Hi everybody, I've been using corba for some time and just recently started using ORBit on my Red Hat Linux 6.1 system.
> > 
> > I currently have ORBit 0.4.95 and I want to be able to program using C++, I understand that the original ORBit only supports C language mappings. 
> > 
> > I have downloaded the patched ORBit 0.5.1 file
> > and ORBitcpp 0.28.1
> > 
> > Could you tell me how to upgrade my ORBit ?
> > Any advice woule be great, thanks !
> > 
> > 
> > Get your FREE Email at http://www.mailcityasia.com
> > 
> > _______________________________________________
> > orbitcpp-list mailing list
> > orb...@li...
> > http://lists.sourceforge.net/mailman/listinfo/orbitcpp-list
> > 
>
>
>_______________________________________________
>orbitcpp-list mailing list
>orb...@li...
>http://lists.sourceforge.net/mailman/listinfo/orbitcpp-list
>


Get your FREE Email at http://www.mailcityasia.com

Re: [orbitcpp-list] warnings

[Laszlo M. <las...@et...>]

> I need to do some work in this area. Any info on what sort of hacks
> mico and omniorb use (or where to look for them) would be very
> helpful.

OK, mico does this in include/var.h:

template<class T>
class ObjVar
{
  // [...]
// g++ const overload problem
#ifdef HAVE_CONST_OVERLOAD
  operator T*() const
  {
      return _ptr;
  }
  operator T*&()
  {
      check();
      return _ptr;
  }
#else
  operator T*&() const
  {
      ((ObjVar<T>*)this)->check();
      return (T*&)_ptr;
  }
#endif
}

and here is the test from its aclocal.m4

AC_DEFUN(AC_CONST_OVERLOAD,
[AC_MSG_CHECKING(for const overload)
AC_CACHE_VAL(ac_cv_have_const_overload,
[ac_cv_tmp_old_cxxflags="$CXXFLAGS"
if test X"$GXX" = Xyes; then
  CXXFLAGS="-Werror $CXXFLAGS"
fi
AC_TRY_COMPILE_GLOBAL(,[
  struct S {
    int *i;
    operator const int *() const { return i; }
    operator int *&() { return i; }
    operator const int &() const { return *i; }
    operator int &() { return *i; }
  };

  void bar (const int &, int &, const int *, int *, int *&);

  void foo (S &s, const S &cs)
  {
     int i = s; i = cs; i = *s; i = *cs;
     bar (s, s, s, s, s);
     bar (cs, s, cs, s, s);
  }
],
eval "ac_cv_have_const_overload=yes",
eval "ac_cv_have_const_overload=no",)
CXXFLAGS=$ac_cv_tmp_old_cxxflags
])


>  >    operator Base&()
>  > So I guess, this code could be safely removed.
> Hmmm - I suspect not. The thing is that we aren't using real C++
> objects, we're casting ORBit C objects into C++ ones, and using these
> cast operators to insure that the compiler doesn't try to do smart
> thunking in the case of multiple inheritance downcasts.  This provides
> a good speed increase over wrapping the C objects in C++ ones.

The problem is that as the warning says, gcc will NOT USE that operator
when it does the casting. Consider this example:

struct Base
{};

Base dummy;

struct Derived  : public Base
{
    operator Base &() { return dummy; }
};

gcc gives:
x.cc:8: warning: conversion to a reference to a base class will never
use a type conversion operator

And gcc is right, because it's totally silly what I'm doing.
I have no idea, why this is only a warning, IMHO it should be an error.

> Does the 'everything' test in orbitcpp CVS work with gcc 2.96?

Well, I just checked whether this const/non const warning is still
there, and didn't compile anything useful (I doubt I could - 2.96
is not binary compatible with my libraries).

Laszlo

Re: [orbitcpp-list] warnings

[Phil D. <ph...@us...>]

Laszlo Molnar writes:
 > Hi Folks!

Hi Laslo,


 > 
 > Some days ago I tried to investigate why gcc gives tons of warnings
 > during compilation of orbitc++ generated stuff. After I found the main
 > cause (those const/non const cast operators) I looked at some other
 > orbs how they solve this problem. I found special ifdefs in both mico
 > and omniorb, unfortunately none of those solutions were 100% perfect.
 > Anyway when I do something similar for orbitc++, I was not able to
 > compile everything :(
 > 
 > And unfortunately gcc 2.96 issues these warnings too :(
 > 

I need to do some work in this area. Any info on what sort of hacks
mico and omniorb use (or where to look for them) would be very
helpful.


 > But there is another kind of warnings, which is caused by this
 > stuff in pass_stubs.cc:
 > 
    [example snipped]

 > Here gcc says, that there is no point to provide
 > 
 >    operator Base&()
 > 
 > for the derived classes, because it won't execute them, but do the
 > right thing itself ;)
 > 
 > So I guess, this code could be safely removed.
 > 

Hmmm - I suspect not. The thing is that we aren't using real C++
objects, we're casting ORBit C objects into C++ ones, and using these
cast operators to insure that the compiler doesn't try to do smart
thunking in the case of multiple inheritance downcasts.  This provides
a good speed increase over wrapping the C objects in C++ ones.

Does the 'everything' test in orbitcpp CVS work with gcc 2.96? 

Thanks for your work,

Cheers,

Phil.

Re: [orbitcpp-list] installing orbitcpp

[Phil D. <ph...@us...>]

Hi Kelvin,

The usual method with gnu software is:

./configure
make
make install

Check out the INSTALL file for details.

Cheers,

Phil.

Kelvin Lim writes:
 > Hi everybody, I've been using corba for some time and just recently started using ORBit on my Red Hat Linux 6.1 system.
 > 
 > I currently have ORBit 0.4.95 and I want to be able to program using C++, I understand that the original ORBit only supports C language mappings. 
 > 
 > I have downloaded the patched ORBit 0.5.1 file
 > and ORBitcpp 0.28.1
 > 
 > Could you tell me how to upgrade my ORBit ?
 > Any advice woule be great, thanks !
 > 
 > 
 > Get your FREE Email at http://www.mailcityasia.com
 > 
 > _______________________________________________
 > orbitcpp-list mailing list
 > orb...@li...
 > http://lists.sourceforge.net/mailman/listinfo/orbitcpp-list
 > 

[orbitcpp-list] installing orbitcpp

[Kelvin L. <kel...@ma...>]

Hi everybody, I've been using corba for some time and just recently started using ORBit on my Red Hat Linux 6.1 system.

I currently have ORBit 0.4.95 and I want to be able to program using C++, I understand that the original ORBit only supports C language mappings. 

I have downloaded the patched ORBit 0.5.1 file
and ORBitcpp 0.28.1

Could you tell me how to upgrade my ORBit ?
Any advice woule be great, thanks !


Get your FREE Email at http://www.mailcityasia.com

[orbitcpp-list] warnings

[Laszlo M. <las...@et...>]

Hi Folks!

Some days ago I tried to investigate why gcc gives tons of warnings
during compilation of orbitc++ generated stuff. After I found the main
cause (those const/non const cast operators) I looked at some other
orbs how they solve this problem. I found special ifdefs in both mico
and omniorb, unfortunately none of those solutions were 100% perfect.
Anyway when I do something similar for orbitc++, I was not able to
compile everything :(

And unfortunately gcc 2.96 issues these warnings too :(

But there is another kind of warnings, which is caused by this
stuff in pass_stubs.cc:

void IDLPassStubs::doInterface(IDLInterface &iface)

[...]

       // make cast operators for all mi base classes
        IDLInterface::BaseList::const_iterator
        first = iface.m_all_mi_bases.begin(),last = iface.m_all_mi_bases.end();

        while (first != last) {
                m_header
                << indent << "operator " << (*first)->getQualifiedCPPIdentifier() << " &() {" << endl;
                m_header
                << ++indent << "return *" << (*first)->getQualifiedCPPCast("this") << ';' << endl;
                m_header
                << --indent << '}' << endl;
                first++;
        }

Here gcc says, that there is no point to provide

   operator Base&()

for the derived classes, because it won't execute them, but do the
right thing itself ;)

So I guess, this code could be safely removed.

Laszlo

Re: [orbitcpp-list] A quick question

[Phil D. <ph...@us...>]

Andreas Kloeckner writes:
 > On Fri, Aug 11, 2000 at 03:29:56PM -0600, Bruce Ide wrote:
 > > Hey all!
 > > 
 > > Does anyone know if the ORBit guys ever rolled the ORBitcpp patches into
 > > the ORB? Can I upgrade ORBit now without breaking ORBitcpp?
 > 
 > Yup, they are in current CVS 
 > and
 > No, they haven't found their way into any release version.
 > 

More to the point, the CVS HEAD version of ORBit doesn't work with
orbitcpp either because of some bugs in the pidl generated stuff. I am
working to fix these ASAP.

Cheers,

Phil.

[orbitcpp-list] delete operation in CORBA_request.idl

[Phil D. <ph...@us...>]

Hi ORBit-list,

CORBA_request.idl and CORBA_context.idl both contain interfaces with a
'delete' operation in them. Under the new ORBit scheme, orbit-idl
generates code in corba_defs.h which is included by everything. This
causes a problem for C++, since 'delete' is a reserved word.

Does anybody have any smart ideas for how we can get around this? The
problem is that the C mapping specifies that the operation name occurs
un-mangled in the epv (everywhere else it has the module & interface
name tacked on before the operation name).

Ideas:

1) Divert from the spec, and mangle the operation sig in the
epv. (e.g. something like _epv_delete).

2) Divert from the spec and use the recommended c++ mangling in epvs
for c++ reserved words. (e.g. _cxx_delete)

3) Don't include corba_defs.h from the client headers.

4) Generate a special C++ corba_defs.h which can be included prior to
including "orbit.h". (would this cause link errors?)

Any thoughts on the above? 
(I'd like this resolved quick, since it stops orbitcpp working with
CVS ORBit).

Cheers,

Phil.

P.S. I've included a patch which does (1) for people who need to
compile CVS head with ORBit right now ;-)

[orbitcpp-list] CosNaming service?

[Bruce I. <nr...@us...>]

Is it possible to enumerate all objects in the CosNaming service? How
would one go about that?

-- 
Bruce Ide                   gr...@pa...
http://www.paratheoanametamystikhood.net

Re: [orbitcpp-list] A quick question

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

On Fri, Aug 11, 2000 at 03:29:56PM -0600, Bruce Ide wrote:
> Hey all!
> 
> Does anyone know if the ORBit guys ever rolled the ORBitcpp patches into
> the ORB? Can I upgrade ORBit now without breaking ORBitcpp?

Yup, they are in current CVS 
and
No, they haven't found their way into any release version.

Hope this is what you wanted to know,
Andy

-- 
Don't innovate, intimidate.

[orbitcpp-list] A quick question

[Bruce I. <bru...@ec...>]

Hey all!

Does anyone know if the ORBit guys ever rolled the ORBitcpp patches into
the ORB? Can I upgrade ORBit now without breaking ORBitcpp?

-- 
Bruce Ide                                           
bru...@ec...
http://www.paratheoanametamystikhood.net

Re: [orbitcpp-list] bug in the c++ backend

[Phil D. <ph...@us...>]

Yep - I'm afraid that the compiler doesn't do 'inlined sequences' yet.

The work around is to typedef them like you say. (this means you cant do
recursive structs etc.. yet.

Cheers,

Phil.

Laszlo Molnar wrote:
> 
> Hi Folks!
> 
> Here is a little bug I found some days ago in orbitcpp.
> 
> module X
> {
>     struct Y
>     {
>         sequence<double> Z;
>     };
> };
> 
> This dumps core for me during compilation. Using a typedef for the
> sequence works around the problem.
> 
> Laszlo
> 
> _______________________________________________
> orbitcpp-list mailing list
> orb...@li...
> http://lists.sourceforge.net/mailman/listinfo/orbitcpp-list

Re: [orbitcpp-list] Getting setup

[Phil D. <ph...@us...>]

Andreas Kloeckner wrote:
> 
> > > Current CVS features everything except:
> > > * attributes
> >
> > Actually this is implemented.
> 
> Great thing to hear. Anything else implemented that was on my list?
> I'll be having time from ~Aug 3 to ~Aug 20 to hack stuff. Have unions
> finally been done?
> 

I'm afraid not. A while ago you mentioned that there was something to do
with arrays which made them a special case when used with unions. Can
you remember what that was? 
(I may have mis-remembered this myself)

Cheers,

Phil.

[orbitcpp-list] bug in the c++ backend

[Laszlo M. <las...@et...>]

Hi Folks!

Here is a little bug I found some days ago in orbitcpp.

module X
{
    struct Y
    {
	sequence<double> Z;
    };
};

This dumps core for me during compilation. Using a typedef for the
sequence works around the problem.

Laszlo

Re: [orbitcpp-list] Getting setup

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

> > Current CVS features everything except:
> > * attributes
> 
> Actually this is implemented.

Great thing to hear. Anything else implemented that was on my list?
I'll be having time from ~Aug 3 to ~Aug 20 to hack stuff. Have unions
finally been done?

bye
andy
-- 
Don't innovate, intimidate.

Re: [orbitcpp-list] Understanding orbit-c++

[Adam Y. <ay...@fo...>]

THis is the compile command that fails.

c++ -DHAVE_CONFIG_H -I. -I. -I. -I.. -I.. -I/usr/local/lib/glib/include
-I/usr/local/include -I/usr/include -g -O0 -Wall
-Wp,-MD,.deps/orbitcpp_tools.pp -c orbitcpp_tools.cc  -fPIC -DPIC -o
.libs/orbitcpp_tools.lo


Is --with-orbit-prefix=/usr/local/devel required?  I'm a little unclear
through all the emails.  Are there other params to configure that are
required (for ORBit to find glib, etc)

BTW 
I'm running:
 gcc --version 
2.95.2

I get a lot of these :
configure.in:93: warning: AC_TRY_RUN called without default to allow
cross compiling


> > IN the configure output I found this snippet:
> > *** Your copy of orbit-idl doesn't support --backenddir
> > *** This means that you will have to install the C++ idl compiler
> > back-end
> > *** (with 'make install') before you can build the tests.
> >
> > Is this indicitave of it not finding the right ORBit code?
> >
> 
> Yes! The CVS head version of ORBit definitely has --with-orbit-prefix.
> 
> Hmmm...
> <gets off his arse and attempts to build ORBit/orbitcpp from scratch>
> 
> I may have the answer:
> I've just attempted to do this on a fresh redhat 6.2 system with helix
> gnome, and helix gnome comes with ORBit-0.5.2. The problem is that the
> ORBit in CVS head is fixed at 0.5.1 for some god-only-knows reason. This
> means that the configure script gets confused (and prints out a verbose
> warning msg), but ultimately doesn't stick the correct include and lib
> directives into the makefiles.
> 
> I fixed this by editing the ORBIT_MICRO_VERSION variable in ORBit
> configure.in from 1 to 2, which fixed the problem.
> 



Hmm, Well that describes my configuration.  But I tried that too and it
didn't run. (I also tried setting it to 3) SOmething is fooling
configure into not picking up the stuff from the /usr/local/devel
directory. 


I'll also try it with the latest CVS of ORBit.

Adam

Re: [orbitcpp-list] ORBit-0.5.3

[Bruce I. <nr...@us...>]

Phil Dawes wrote:
> 
> Doh! Those following the ORBit list will have noticed that Elliot's just
> released ORBit-0.5.3.
> 
> Here we go again! ;-)

Swell. Will that break OrbitCPP? Gnome?

-- 
Bruce Ide                        nr...@us...
If trees screamed would we be as quick to cut them down? Possibly,
if they screamed all the time.

Re: [orbitcpp-list] Understanding orbit-c++

[Bruce I. <nr...@us...>]

Adam Young wrote:
> (SNIP)...
>
> 6)and finally for orbitcpp
> 
> cd ../orbitcpp
> ./autogen.sh -prefix /usr/local/devel/
> make
> 
> This is the step thaht fails for me.  I get a build error in
> orbitcpp_tools looking MemHow.  I know this is in
> PREFIX/include/orb/allocators.h
> (SNIP)...

That problem went away for me when I upgraded to GCC 2.95.2. I believe
the FAQ mumbles something about it. Maybe it should mumble it a little
more loudly...

-- 
Bruce Ide                        nr...@us...
If trees screamed would we be as quick to cut them down? Possibly,
if they screamed all the time.