DACP in Rhythmbox: Weekly Report 10

[Alexandre Rosenfeld <alexandre rosenfeld gmail com>]

Hi everyone,

Weekly report 10 (already?!). As usual, check the wiki at�http://live.gnome.org/SummerOfCode2010/AlexandreRosenfeld_Rhythmbox

To all of you that were in GUADEC, hope you had a good time. And it's another year that I wished I had gone to GUADEC ;)

- What I did this week?

This week was very productive. Earlier this week I sent patches to bugzilla with my code (https://bugzilla.gnome.org/show_bug.cgi?id=625214�and�https://bugzilla.gnome.org/show_bug.cgi?id=625213). It's nice getting some feedback, especially from Jonathan, the Rhythmbox mantainer. It will be easier to get it accepted later on. After fixing most of the problems pointed out I then fixed some other long-lasting bugs.

What I'm able to do now is browse the entire Rhythmbox library with the Remote. And after figuring out why everything was out of order (libdmapsharing wasn't sorting listings), I had to change some code to add sorted output to libdmapsharing. There are other places that also needs to support sorting, but I'll be adding them as needed.

The early work in the query parser are paying off now, as more and more complicated queries are being used by the Remote. Pretty much every command involves a query (which got me thinking about optimizations again, but it's out of scope now. And in my computer the queries are pretty fast). And led me to implement groups quite easily, now the list of albums appears in the Remote just as they should.

- What will I do next week?

Somehow the Remote hangs on an infinite loop trying to get the current volume, which is being sent correctly. On further investigation I fixed other responses that were being incorrectly sent, but nothing changed this issue. The Remote probably timeouts sometimes and then other requests are processed.

I have started to implement the CUE command, which allows to queue and play songs. I'll probably be using a playlist like iTunes instead of the Rhythmbox queue, because their behaviour are different. So I'll spent some time on this.

- Was my planning accurate?

The TODO list helped a lot, I finished the first two items and the signals are not a priority because they can't be tested right now (the Remote don't send their commands. Hopefully after implementing CUE they will work).

- Cool things?

More and more I'm finding that my initial planning was completely off and DACP is a way more complicated protocol then I thought. I'm having to check other implementations more and more to see if my implementation is correct.�

Also, which is pretty cool, my project is stressing a lot the Rhythmbox DAAP implementation, which will make it a lot better, even if originally this wasn't planned at all.

That's all folks, until next report,

Alexandre