On Tue, Jul 10, 2001 at 08:01:04AM -0500, Frank Singleton wrote:
> Frank Singleton wrote:
>
> > Just for my own sanity, if we apply this patch then does it mean it
> > will now build giop subdissectors that are not usable because of
No. I have used it and it seems to work.
However, It generates one plugin for each giop subdissector (one plugin
for each IDL file). We'll have to modify it if we want to generate one
big plugin for all the giop subdissectors :
> > > Yes, you can put more than one protocol into a plugin. There's
> > > currently no *automatic* provision for linking multiple
> > > independent files, each with its own dissector and register
> > > routines, into a single plugin, so you'd have to have a common
> > > "plugin_init()" routine call all of their registration and
> > > initialization routines and a common "plugin_reg_handoff()"
> > > routine call all of their handoff routines.
> >
> > I am not up to speed on plugins yet, so any response would be
> > appreciated. It would appear that packet-giop.[ch] would need
> > updating as described above before the patch should be applied.
> >
>
> I guess I am trying to say that (IMO) its better for the forthcoming
> release to keep giop and idl2eth using "normal" dissectors, so
> end users can make use of it , then after the release, do the
> plugin thing ? I promise not to pollute the src tree with
> giop subdissectors until we get the plugin thing happening after
> the 0.8.19 release this week :-)
>
> I will however try to sneek in the USER exception code in the
> next 2 days .
>
> End users are using idl2eth now, and I just dont want to break anything
> for an official release.
I agree with you. The plugin conversion can wait.
Olivier
--
In order to live free and happily, you must sacrifice boredom.
It is not always an easy sacrifice.