Re: [anjuta-devel] The new parser-engine Plugin

[Abderrahim Kitouni <a kitouni gmail com>]

Hello Moritz,

The purpose of my questions isn't to challenge you or something like
that, I just want to make sure I understand what you're doing so I can
help you. (And I think you may also understand things better by
explaining how they work). So you shouldn't feel attacked, right? :-)

                  في خ، 12-07-2012 عند 12:48 +0200 ، كتب Moritz Lüdecke:
> > So the question is how a language support plugin is supposed to work
> > after your changes? And how are plugins are going to communicate?
>
> If I move IAnjutaCalltipProvider to IAnjutaProvider, a language support 
> plugins will work as before my changes.

What do you mean by "before my changes"? The same way it's working right
now in anjuta master? (i.e. having plugins register for a watch on the
current document, and register the provider on the editor) or some
intermediate state on your repository (that I don't know about)?

My understanding is that you wanted to have a new architecture, by
moving some code between plugins, and that's why I ask about how the
language plugins are supposed to work.

>  Only the IAnjutaProvider 
> interface will change and integrate following methods:
>   * populate: As before to show completion
>   * get_context: Searches for a calltip context
>   * get_boolean: Get setting like PREF_CALLTIP_ENABLE
>   * clear_context: Clears the calltip context and brings it back into a 
> save state
>   * query: Starts an async query for the calltip

Is there an implementation of something similar to this already in your
repository? (btw, the methods could use a better name, what is a
"calltip context". I'm also concerned that you may be leaking some
implementation details into the interface.)

Another question: How about your problem with communication between
plugins (the purpose of your original post), is it solved by using this
interface?

Regards,
Abderrahim