Comments on: What is happening nowadays? http://pvanhoof.be/blog/index.php/2008/12/26/what-is-happening-nowadays From the mind of Philip Sun, 28 Aug 2011 12:59:07 +0000 hourly 1 http://wordpress.org/?v=3.0.5 By: pvanhoof http://pvanhoof.be/blog/index.php/2008/12/26/what-is-happening-nowadays#comment-969 pvanhoof Fri, 26 Dec 2008 14:20:24 +0000 http://pvanhoof.be/blog/index.php/2008/12/26/what-is-happening-nowadays#comment-969 @Kevin: Although that would make more sense, I decided to make it separate arrays because this was more simple to implement client-side in a higher language (demarshaling an 'a(ss)' is not as common as demarshaling two 'aas'). This doesn't mean this ain't open for discussion, of course. The purpose of the current proposed API and implementation is to serve as a proof of concept. Therefore if during the implementing of the concept things like this are more easy done differently, then we can of course adapt the proposal. What about making a vala-client sample (based on the existing vala sample that I posted in the blog item) and adapting my patch? Then we can see how your proposal works out in for example Vala, and make a decision. We should probably also make client-side examples using GJS (JavaScript), Python, C and C# just to cover most of the higher languages. Unlike the Xesam specification I try to have as simple DBus signatures as possible. It's not a contest of having the most complicated type signatures in my opinion. Rather one where it's as easy as possible to hook in and start consuming the service. Yet not cutting features. Agreed that 'a(ss)' is not terribly difficult either. Let's see? @Kevin: Although that would make more sense, I decided to make it separate arrays because this was more simple to implement client-side in a higher language (demarshaling an ‘a(ss)’ is not as common as demarshaling two ‘aas’). This doesn’t mean this ain’t open for discussion, of course.

The purpose of the current proposed API and implementation is to serve as a proof of concept. Therefore if during the implementing of the concept things like this are more easy done differently, then we can of course adapt the proposal.

What about making a vala-client sample (based on the existing vala sample that I posted in the blog item) and adapting my patch? Then we can see how your proposal works out in for example Vala, and make a decision. We should probably also make client-side examples using GJS (JavaScript), Python, C and C# just to cover most of the higher languages.

Unlike the Xesam specification I try to have as simple DBus signatures as possible. It’s not a contest of having the most complicated type signatures in my opinion. Rather one where it’s as easy as possible to hook in and start consuming the service. Yet not cutting features. Agreed that ‘a(ss)’ is not terribly difficult either. Let’s see?

]]> By: Kevin Krammer http://pvanhoof.be/blog/index.php/2008/12/26/what-is-happening-nowadays#comment-968 Kevin Krammer Fri, 26 Dec 2008 13:59:45 +0000 http://pvanhoof.be/blog/index.php/2008/12/26/what-is-happening-nowadays#comment-968 The D-Bus introspection XML and the example code seem to indicate that predicate and value arrays have a one-to-one mapping, i.e. value at some index belonging to predicate at same index. Doesn't it make much more sense to have one array of pairs? a(ss) in D-Bus types? The D-Bus introspection XML and the example code seem to indicate that predicate and value arrays have a one-to-one mapping, i.e. value at some index belonging to predicate at same index.

Doesn’t it make much more sense to have one array of pairs?
a(ss) in D-Bus types?

]]>