Re: [anjuta-devel] RFC: Killing Anjuta-Extras

[James Liggett <jrliggett cox net>]

Hi Seb,

On Thu, 2011-01-20 at 21:32 +0100, Sébastien Granjoux wrote:


> > 1. Having some plugins in a separate module is a serious maintenance
> > headache.
>
> I'm agree. On the other hand, moving these plugins out of the main 
> anjuta module have made the changes in anjuta easier.
>
> Overall, I think we probably have still gain some time doing it like 
> this specially at the moment with all changes in Gtk. I don't really see 
> a problem if anjuta-extras is broken for some time.
I have to be honest, I really don't see that. While it might mean that
you don't have to do as much work to make changes to Anjuta's core
functionality, it also means that some otherwise useful plugins just
fall into disrepair because they're out of sight for most users. And
having things that are broken released to the public and passed off as
stable makes us look bad. 

If we are to accept that anjuta-extras plugins might be broken for long
periods, then we probably should avoid making stable series releases for
them and maybe just do development releases. 

> Moreover, it allows us to keep some useful code without all gnome 
> constraints.
In these cases, we're probably better off trying to work around those
constraints, like we did with the GnomeCanvas-foocanvas switch for the
class inheritance plugin. There's no reason we couldn't do the same for
the profiler and any other plugins that might have similar problems. 


> Basically, I think we could still make a 3.0 release, if at least 
> scintilla is ready. The scratchbox plugin is useful but not worths the 
> release by itself. I still see some advantages of having anjuta-extras 
> in separate module.
OK, I'll tell you what: if you can get at least some of the plugins
workable with gtk 3 by freeze, I'll make the release. I've also made
that branch you asked for. Let's see what you can come up with... :)

Thanks,
James