Bug 486 - IDL Cubit Test - Collocation Part
Summary: IDL Cubit Test - Collocation Part
Status: RESOLVED
Alias: None
Product: TAO
Classification: Unclassified
Component: Example (show other bugs)
Version: 1.0.13
Hardware: x86 Linux
: P1 enhancement
Assignee: Nanbor Wang
URL:
Depends on:
Blocks:
 
Reported: 2000-03-10 10:57 CST by Chanaka Liyanaarachchi
Modified: 2000-03-27 23:53 CST (History)
0 users

See Also:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Chanaka Liyanaarachchi 2000-03-10 10:57:03 CST
The TAO_IDL_FLAGS does not seem to be correct for  "direct" collocation
strategy.  The example compiles the IDL file with "thru_poa" collocation
strategy (Default) and the README gives the idea that you can use this for both
"direct" and "thru_poa" strategies. My understanding is for "direct"
the -Gd (-Sp) are needed by the idl compiler, meaning the same stubs and
skeletons cannot be used for both the above strategies.
(Am i correct ?)

Both these invocations would result in:

./collocation_test -s "-ORBCollocationStrategy direct"

or

./collocation_test -s "-ORBCollocationStrategy direct -f server.ior" -c "-k
file://server.ior"


(15737|1024) EXCEPTION, Cubit::init
system exception, ID 'IDL:omg.org/CORBA/BAD_PARAM:1.0'
TAO exception, minor code = 54410000 (unknown location; unspecified errno),
completed = NO

This makes sense since that the "-ORBCollocationStrategy direct" is invalid for
the generated stuf.


Furthermore, when I compile cubit for "direct" collocation,

/vol1/1.0.13/ACE_wrappers/TAO/TAO_IDL/tao_idl -Gd -Sp cubit.idl

The collocation example cannot be compiled:

In file included from Cubit_Server.h:33,
                 from Cubit_Server.cpp:3:
Cubit_i.h:146: cannot declare field `Cubit_Factory_i::my_cubit_' to be of type
`Cubit_i'
Cubit_i.h:146:   since the following virtual functions are abstract:
Cubit_i.h:146:  void POA_Cubit::cube_oneway()
Cubit_i.h:146:  void POA_Cubit::cube_void()

and more....
Comment 1 Nanbor Wang 2000-03-10 15:31:59 CST
mine
Comment 2 Nanbor Wang 2000-03-10 15:31:59 CST
mine, again
Comment 3 Nanbor Wang 2000-03-27 23:53:59 CST
I didn't see the problem reported.  The IDL file is compiled with both thru_poa
and direct collocated stubs available.  TAO_IDLFLAGS is defined as "-Ge 1 -Gd"
so both it looks fine to me.  Please let me know if you see any other problem
with the test.