[Bernd Becker <bb@xxxxxxxxxxxxxxx>]

Hi,
here is a patch I sent in a while back that never made it into the release.
Could somebody check it in?
Thanks,
Bernd

--On Wednesday, April 23, 2003 13:53:43 +0200 Bernd Becker 
<bb@xxxxxxxxxxxxxxx> wrote:

Here is a patch I had promised to Nicolas De Montbel a while ago in
http://www.ethereal.com/lists/ethereal-dev/200303/msg00161.html

The response_flags in the Request_Header should not be mapped
one to one to the SyncScope values. Certain SyncScope values
just seem to imply certain response_flags according to a
comment in the Corba Spec v3.0, 15-35.
So now I print the numeric value of response_flags followed
by a string with the possible corresponding SyncScope values
in parenthesis.
If this is still confusing we should probably print only the
numeric value.
> Bernd
>
> --On Tuesday, March 11, 2003 10:26:22 +0100 Bernd Becker
> <bb@xxxxxxxxxxxxxxx> wrote:
>
> > Hi Nicolas,
> >
> > I think you are right. We should not map the response_flags to
> > sync_scope.
> > The sync_scope defined in packet-giop.c obviously came from the
> > following definition in Corba Messaging Quality of Service:
> >
> > typedef short SyncScope;
> > const SyncScope SYNC_NONE = 0;
> > const SyncScope SYNC_WITH_TRANSPORT = 1;
> > const SyncScope SYNC_WITH_SERVER = 2;
> > const SyncScope SYNC_WITH_TARGET = 3;
> >
> > We should leave the sync_scope value_string unchanged in case somebody
> > implements the Messaging module, although that would probably be another
> > plugin.
> > We should either just display the numeric value for the response_flags
> > or maybe display something more like a comment as in the spec, e.g.
> >
> > static const value_string response_flags[] = {
> >     { 0x0, "SyncScope NONE or WITH_TRANSPORT" },
> >     { 0x1, "SyncScope WITH_SERVER"},
> >     { 0x3, "SyncScope WITH_TARGET"},
> >     { 0, NULL}
> > };
> >
> > I can supply a patch if you (and maybe others) agree.
> >
> > I don't really know if anybody has looked at the changes in the Corba 3.0
> > Spec to see what has to be added/changed in packet-giop.c. Some additions
> > were made regarding Service Contexts a short time ago.
> >
> > It might be better to ask such questions on the developer list, as I
> > suspect that not many of the few giop developers check the users list. I
> > will send this to the developers list.
> >
> > Regards,
> > Bernd
> >
> > --On Monday, March 10, 2003 17:25:00 +0100 Nicolas.De_Montbel@xxxxxxxxxx
> > wrote:
> >
> > > Hello everybody,
> > > I am new on this tool which seems to be very convenient and complete
> > > but I have a problem on decoding one giop request header (response
> > > flags).
> > >
> > > from CORBA3.0 specification :
> > > "response_flags is set to 0x0 for a SyncScope of NONE and
> > > WITH_TRANSPORT. The flag is set to 0x1 for a SyncScope of
> > > WITH_SERVER. A non exception reply to a request message containing a
> > > response_flags value of 0x1 should contain an empty body, i.e. the
> > > equivalent of
> > > a void operation with no out/inout parameters. The flag is set to 0x3
> > > for a SyncScope of WITH_TARGET. These values ensure interworking
> > > compatibility between this and previous versions of GIOP."
> > > In packet-giop.c :
> > >         static const value_string sync_scope[] = {
> > >               { 0x0, "SYNC_NONE" },
> > >               { 0x1, "SYNC_WITH_TRANSPORT"},
> > >               { 0x2, "SYNC_WITH_SERVER"},
> > >               { 0x3, "SYNC_WITH_TARGET"},
> > >               { 0, NULL}};
> > >
> > > This is quite different.
> > >
> > > Can you confirm that
> > > and say to me how can I correct that in Ethereal
> > > sincerly
> > >
> > > Nicolas de Montbel

Attachment: packet-giop.patch
Description: Binary data