On Tue, Jan 17, 2012 at 5:18 PM, Chris Kühl <blixtra(a)gmail.com> wrote:
On Tue, Jan 17, 2012 at 4:03 PM, Patrick Ohly
> On Mo, 2012-01-16 at 17:19 +0100, Patrick Ohly wrote:
>> After also fixing the handling of asynchronous method
>> implementation in GDBus GIO, local sync works with it.
> After fixing some more regressions the D-Bus tests ran well
), so I
> pushed all of the changes for gdbus C++ and fork/exec to the master
> Chris, this is different from the previous fork/exec branch, so please
> do a "git rebase -i" and pick just your own changes when moving to the
> latest code.
Ok, I'll take a look after I've pushed the changes I'm working on now
to get the full one-to-one dbus interface implemented.
I've renamed my branch to concurrent-sync-sessions and rebased onto
master. I'm now going through and making required changes to get tests
Btw, similar to the way the code in syncevolution.cpp makes a dbus
method call then runs a mainloop till the call back has returned, I
had to create such a hack but slightly differently. I've
implemented this by polling the dbusconnection's fd and on any fd
activity, do a context iteration, then seeing if the callback has
finished. Like I said it's a hack and I'll see how much mileage this
gets me but I'm not convinced it's any worse than what's being done in
syncevolution.cpp. Feel free to convince me, though. ;)