Re: [Evolution-hackers] Regarding API breakage and lost test cases

[Matthew Barnes <mbarnes redhat com>]

On Thu, 2013-01-31 at 18:25 +0900, Tristan Van Berkom wrote:
> I'm not sure what the solution should be exactly, but we
> should discuss this a bit because the problem space is
> a little bigger than just the source existing in the
> client-side before e_source_registry_commit_source_sync()
> returns.

Note that e_source_registry_commit_source_sync() is just a convenience
wrapper.  If your test cases are just passing scratch sources then what
you're actually calling is e_source_registry_create_sources_sync().
That's the function I fixed in bug 685986.


> So I guess there are 2 potential ways of addressing this problem:

There's a 3rd approach.

If you want an immediate client connection to a newly created address
book, we could have evolution-addressbook-factory create the ESource on
your behalf -- since it too has an ESourceRegistry -- and immediately
instantiate a backend for it.  That would eliminate the race without
resorting to client-side hacks.

It would require a new method on the addressbook and calendar factory
interfaces, but that is easily done now that all our D-Bus interfaces
are generated from XML specifications.  That was the purpose of my
recent rash of commits to master.

Just this week I added e_book_client_connect_sync(), which combines the
now-deprecated e_book_client_new() and e_client_open_sync() functions
into one step.

I could enhance this function to detect whether the given ESource is a
scratch source, and invoke a different factory method which behaves as
described above.  No additional client-side APIs required.


To summarize:

The current addressbook factory D-Bus API looks like this:

   <method name="OpenAddressBook">
     <arg name="source_uid" direction="in" type="s"/>
     <arg name="object_path" direction="out" type="s"/>
   </method>

It only works for existing ESources already exported by the registry.

I'm proposing to add a new method that takes the raw key file content
from a scratch ESource in addition to a UID:

   <method name="CreateAddressBook">
     <arg name="source_uid" direction="in" type="s"/>
     <arg name="source_data" direction="in" type="s"/>
     <arg name="object_path" direction="out" type="s"/>
   </method>

Invoked by passing a scratch source to e_book_client_connect_sync().

The factory method would call e_source_registry_commit_source_sync() on
your behalf, and assuming that goes well, return the object path to its
newly-created EBookBackend.

With equivalent changes to the calendar factory, of course.

Matthew Barnes