|
| Author: neiljp |
Date: 2002-11-04.22:12:48 |
This becomes important when we care whether we:
- failed during a call (TRANSIENT)
- failed to start the call (COMM_FAILURE)
This is not a major problem, since in most cases
we just care whether we failed or not, not how,
but it would be good to standardise. Different
things are standardised in different corba
versions; iirc COMM_FAILURE is only required with
the version supported in omni4. Omni3 supported an
earlier standard where TRANSIENT was an acceptable
response.
Basically: if we *can* distinguish, then we should.
|
| Author: neiljp |
Date: 2002-12-17.21:10:30 |
This is actually a server-stability issue if we
only catch for one and not the other.
Raising priority accordingly. cc'ing stefan since
we discussed this on irc.
|
| Author: neiljp |
Date: 2002-12-17.21:26:46 |
Stefan: add this to m2 please? Will you be
checking the wiki page? We need some better way
of milestone-bug/task suggesting than this!
|
| Author: tobias |
Date: 2003-01-04.18:14:03 |
Two things:
In general it should be enough to catch a
CORBA::SystemException. We don't really care which
one it is in most places anyway and this will
catch both, COMM_FAILURE and TRANSIENT (plus 27
others).
For debugging purposes it would be nice to have a
detailed report of all 29 exceptions that can
happen. The probelm I see is: How to make this
convinient for the developer? The only thing I can
think of to do that would be a macro
CATCH_AND_LOG_CORBA_EXEPTIONS which gets expanded
to that huge block of catch(es)... not nice, but
betten then rewritting that block all over the
place and forgetting something while doing so.
|
| Author: anonymous |
Date: 2004-01-12.23:28:56 |
For the sake of kitten-protection (based on
current #fresco topic ;), and perhaps also towards
getting M3 out before the *next* millenium, what
about simply adding a function which
dynamic_cast's to each possible SystemException in
a long if-else chain, outputting as appropriate?
Wouldn't that be better than a macro? Perhaps with
a parameter which specifies the ostream to log to. |
| Author: stefan |
Date: 2004-01-13.04:12:29 |
Don't CORBA::Exceptions have an output operator
defined that we could use instead of a home-baked
switch on a dynamic_cast ?
All we want is (polymorphically) logging the
exception to the console / log file.
(see also H&V chapter 7.15.6) |
| Author: anonymous |
Date: 2009-11-06.01:21:15 |
8jYWyh <a href="http://vsjdeltdiphd.com/">vsjdeltdiphd</a>, [url=http://sfngmixlkvbc.com/]sfngmixlkvbc[/url], [link=http://owxboymhznou.com/]owxboymhznou[/link], http://xinwefxbcbnn.com/ |
|
| Date |
User |
Action |
Args |
| 2002-11-04 22:12:48 | neiljp | create | files: (no value)
status: new
assignedto: (no value)
severity: minor
title: Catch COMM_FAILURE or TRANSIENT properly depending upon orb version
nosy:
neiljp
topics: (no value)
messages:
msg653
resolution: (no value)
priority: low
platforms: (no value)
bug_type: behavior
components:
general
dependson: (no value)
superseder: (no value) |
| 2002-12-17 21:10:31 | neiljp | set | priority: low
nosy:
+ stefan
messages:
+ msg1151
severity: minor |
| 2002-12-17 21:26:46 | neiljp | set | messages:
+ msg1155 |
| 2002-12-18 02:09:22 | admin | link | milestone2 bugs |
| 2003-01-04 18:14:03 | tobias | set | nosy:
+ tobias
messages:
+ msg1263 |
| 2003-03-04 06:15:35 | admin | link | milestone3 bugs |
| 2004-01-12 23:28:58 | anonymous | set | messages:
+ msg1932 |
| 2004-01-13 04:12:30 | stefan | set | messages:
+ msg1936 |
| 2009-11-06 01:21:25 | anonymous | set | platforms:
+ GNU/Linux, FreeBSD, Solaris, not applicable
messages:
+ msg4551
components:
+ build system, Babylon, Fresco, Berlin core, Berlin server, C++ clients, tracker test
nosy:
+ je_apostrophe, maddanio, njs, - stefan |
not logged in
register
Home