Re: [Evolution-hackers] State of the Bonobo removal effort for Evolution

[Srinivasa Ragavan <sragavan novell com>]

On Mon, 2008-11-17 at 01:36 -0500, Matthew Barnes wrote:
> Hi folks,
> 
> One of the goals for Evolution 2.26 is to finally remove the use of
> Bonobo and BonoboUI from Evolution [1] in favor of equivalent
> functionality now provided by the GTK+ stack.  This is happening in
> parallel with Ross Burton's effort to port Evolution-Data-Server from
> using Bonobo to D-Bus as the communication medium.  The two efforts are,
> at this point, still progressing independently.
> 
> I've been working on removing Bonobo from Evolution since late August
> and have been publishing my source code changes to a Subversion branch
> named "kill-bonobo" [2].  I've also been maintaining a wiki page by the
> same name [3] with screen shots and somewhat regular updates, though the
> updates tend not to go into much technical detail.
> 
> With the U.S. on daylight savings time now, our weekly IRC meetings are
> at 05:00 EST and, not being much of a morning person, I've been waking
> up in time even less frequently than I did when they were at 06:00 EST.
> So I feel like I've been a little incommunicado lately about how things
> are progressing.
> 
> I hope to correct that with this, the first of a series of postings
> about where things stand and also some technical details about the
> direction I've taken with the newly-rewritten EShell.
> 
> The wiki page [3] already has a brief overview of the new shell design,
> along with a link to some incomplete and slightly out-of-date API docs
> for it [4].  I'll try to improve the completeness and quality of the
> documentation over the next week.
> 
> 
> So, for the remainder of this first post I'll cover schedule.
> 
> Is this going to make it in time for 2.26?
> 
> Short answer is: I still don't know.
> 
> Red Hat was gracious enough to allocate two months for me to work on
> this full time.  In that time I managed to complete the new EShell, more
> or less finish the Contacts, Tasks and Memos modules, and get a good
> start on the Calendars and Mail modules.  But with November already here
> I'm obligated to focus on other issues of interest to Red Hat during
> business hours, which leaves evenings and weekends (when I can find
> time) to work on this.  Suffice it to say, progress has slowed.
> 
> Unfortunately, no matter how much I test the "kill-bonobo" branch
> beforehand, the truth is when we finally do merge it to trunk there
> _will_ be regressions.  Lots of 'em.  Hopefully no major show-stoppers
> but lots of little things I missed.  The scope of the changes here is
> simply too massive for me to catch everything.  Heck, I'm still fixing
> little bugs I missed in the composer rewrite from earlier this year and
> that was of much smaller scope than this.  We'll need a few months of
> heavy testing and bug reporting from developers and interested members
> of the community after the merge and before the stable release date.
> 
> So, I've drawn a line in the sand: 2.25.3.  If the merge doesn't happen
> in time for that release, which is scheduled for mid-December, I'm
> inclined to re-target the feature to 2.27.1.
> 
> As I said, at this point I don't know if I'm going to make it in time.
> There's still a lot of work left.  Finish the Mailer and Calendar for
> starters, and then go through and rework most of the plugins (EMenu and
> EPopup are going bye-bye).  Plus I have a TODO list a mile long of loose
> ends I have to go back and address.  Evolution-Exchange will likely need
> some rework, though to what degree I'm not yet sure.  And I also don't
> know what impact all this will have on the MAPI work.
> 
> It's getting there.  Just not as fast as I would've liked.  In the next
> posting I'll talk about the new shell design and touch on some of the
> things it now handles for you that the current shell does not.
> 
> If you have questions, suggestions or concerns, please don't hesitate to
> shoot me or the list an email.  After all, I'm writing this in hopes of
> getting some feedback.
> 

Thanks for the great update on this Matt. Its going to be a tough time
for many of us for the next couple of months. That's one reason, I asked
you to look at Camel/Gobject stuff for 2.27.x. May be, we have too many
things to do, which would take more than a 6-month cycle? May be a
bigger GNOME 3.0 schedule might fit us well to revamp most of Evolution
[Split it, Kill bonobo, Write Accounts well etc]. Anyways, just some
thoughts.

-Srini.