Re: [Evolution-hackers] A little E-D-S code reorganization

[Milan Crha <mcrha redhat com>]

On Wed, 2011-09-28 at 10:55 -0400, Matthew Barnes wrote:
> Now that I have this 3rd D-Bus service on my account-mgmt branch, I'm
> realizing there isn't really a natural place in our source tree to add
> it.  I kinda just have it lumped on at the moment.
> 
> What I'd like to do is add a new top-level folder called "services" and
> split the e-addressbook-factory and e-calendar-factory programs out from
> their respective "libedata" libraries.
> 
> NOT the EData[Book/Cal]Factory classes, mind you, but just the main()
> function, the migration gunk and the D-Bus .service files.
> 
> So the layout I'm proposing would look something like this:
> 
> 
>   evolution-data-server/
>    |
>    +- services/
>        |
>        +-- e-addressbook-factory/
>        |    |
>        |    +-- e-addressbook-factory.c
>        |    +-- e-addressbook-factory-migrate-basedir.c
>        |    +-- org.gnome.evolution.dataserver.AddressBook.service.in
>        |
>        +-- e-calendar-factory/
>             |
>             +-- e-calendar-factory.c
>             +-- e-calendar-factory-migrate-basedir.c
>             +-- org.gnome.evolution.dataserver.Calendar.service.in
> 
> 
> Everything else stays where it is.  Then on my branch I can just add a
> new "services/e-source-registry" folder.  The messy GConf-to-keyfile
> migration gunk will eventually wind up in these folders as well.
> 
> Sound okay?  Anyone morally opposed?

	Hi,
sounds good.

I think I mentioned it earlier, and maybe this is the right time, what
about replacing "e-" prefix with "evolution-" prefix for process names?
There was a discussion about it some time ago, Chen suggested it, and I
think it's a good idea. It has a benefit of searching for running
evolution-related process with one ps command only. I know it's more
system wide, not only for evolution itself, but because it comes from
evolution-data-server, then the change makes sense anyway.
	Bye,
	Milan