[omniORB] Re: omniidl segfaults

[Juergen Weber]

Hi,

I am trying to access a Geronimo based EJB from an omniorb client,
just to see, if this is possible (if it works, I'd opensource the code
for an omniORB and Geronimo sample).
I got 1.6 rmic to create the IDLs,
rmic -idl -d rmic -classpath
lib\geronimo-ejb_3.0_spec-1.0.jar;CustomerService-ejb.jar
ejb.CustomerService
had to tweak them a bit, but now omniidl segfaults compiling them.

In case somebody might try out this, I append the changed IDL sources.

Thanks,
J?rgen

***@ubuntu:~/c/CustomerService$ omniidl -bcxx -Wbinline -I.
-I/home/weberjn/omniORB-4.1.1/idl ejb/CustomerService.idl
javax/ejb/EJBObject.idl:61: Warning: Identifier 'primaryKey' clashes
with CORBA 3 keyword 'primarykey'
omniidl: 1 warning.
Segmentation fault


***@ubuntu:~/c/CustomerService$ uname -a
Linux ubuntu 2.6.22-14-server #1 SMP Sun Oct 14 23:34:23 GMT 2007 i686 GNU/Linux

***@ubuntu:~/c/CustomerService$ which omniidl
/home/weberjn/omniORB-4.1.1/bin/omniidl

#!/usr/bin/python
***@ubuntu:~/c/CustomerService$ file /usr/bin/python
/usr/bin/python: symbolic link to `python2.5'
***@ubuntu:~/c/CustomerService$ file /usr/bin/python2.5
/usr/bin/python2.5: ELF 32-bit LSB executable, Intel 80386, version 1
(SYSV), for GNU/Linux 2.6.8, dynamical ly linked (uses shared
libs), stripped
-------------- next part --------------
A non-text attachment was scrubbed...
Name: CustomerService-IDL.zip
Type: application/zip
Size: 17506 bytes
Desc: not available
Url : http://www.omniorb-support.com/pipermail/omniorb-list/attachments/20071203/72bec61d/CustomerService-IDL-0001.zip

[Wernke zur Borg]

Subject: [omniORB] Re: omniidl segfaults
Hi,
I am trying to access a Geronimo based EJB from an omniorb client,
just to see, if this is possible (if it works, I'd opensource the code
for an omniORB and Geronimo sample).
I got 1.6 rmic to create the IDLs,
rmic -idl -d rmic -classpath
lib\geronimo-ejb_3.0_spec-1.0.jar;CustomerService-ejb.jar
ejb.CustomerService
had to tweak them a bit, but now omniidl segfaults compiling them.
In case somebody might try out this, I append the changed IDL sources.
Thanks,
J?rgen
-I/home/weberjn/omniORB-4.1.1/idl ejb/CustomerService.idl
javax/ejb/EJBObject.idl:61: Warning: Identifier 'primaryKey' clashes
with CORBA 3 keyword 'primarykey'
omniidl: 1 warning.
Segmentation fault

I can't tell you about the segfault but have you tried to replace the clashing identifier in EJBObject.idl with a non-clashing one?

Wernke

[Juergen Weber]

Post by Wernke zur Borg

Subject: [omniORB] Re: omniidl segfaults
Hi,
I am trying to access a Geronimo based EJB from an omniorb client,
just to see, if this is possible (if it works, I'd opensource the code
for an omniORB and Geronimo sample).
I got 1.6 rmic to create the IDLs,
rmic -idl -d rmic -classpath
lib\geronimo-ejb_3.0_spec-1.0.jar;CustomerService-ejb.jar
ejb.CustomerService
had to tweak them a bit, but now omniidl segfaults compiling them.
In case somebody might try out this, I append the changed IDL sources.
Thanks,
J?rgen
-I/home/weberjn/omniORB-4.1.1/idl ejb/CustomerService.idl
javax/ejb/EJBObject.idl:61: Warning: Identifier 'primaryKey' clashes
with CORBA 3 keyword 'primarykey'
omniidl: 1 warning.
Segmentation fault

I can't tell you about the segfault but have you tried to replace the clashing identifier in EJBObject.idl with a non-clashing one?
Wernke

When I replace the clashing identifier omniidl still segfaults.

But anyway, CustomerServiceSK.cc and CustomerService.hh are written
before the segfault (correctly, I hope), so the segfault shouldn't
matter too much.

Juergen

[Duncan Grisby]

Post by Juergen Weber
I am trying to access a Geronimo based EJB from an omniorb client,
just to see, if this is possible (if it works, I'd opensource the code
for an omniORB and Geronimo sample).
I got 1.6 rmic to create the IDLs,
rmic -idl -d rmic -classpath
lib\geronimo-ejb_3.0_spec-1.0.jar;CustomerService-ejb.jar
ejb.CustomerService
had to tweak them a bit, but now omniidl segfaults compiling them.

Thanks for the bug report. The problem was that omniidl incorrectly
cleaned up after seeing a forward declaration of a valuetype that had
previously been fully declared. I've fixed it in CVS.

Cheers,

Duncan.