Return-Path: MIME-Version: 1.0 In-Reply-To: References: <1335561285-3332-1-git-send-email-lucas.demarchi@profusion.mobi> From: Lucas De Marchi Date: Mon, 30 Apr 2012 10:37:00 -0300 Message-ID: Subject: Re: [BlueZ v3 00/10] gdbus: Better D-Bus introspection To: Luiz Augusto von Dentz Cc: linux-bluetooth@vger.kernel.org Content-Type: text/plain; charset=ISO-8859-1 Sender: linux-bluetooth-owner@vger.kernel.org List-ID: On Mon, Apr 30, 2012 at 7:34 AM, Luiz Augusto von Dentz wrote: > Hi Lucas, > > On Mon, Apr 30, 2012 at 12:40 AM, Lucas De Marchi > wrote: >> >> Since we are using '[' and ']' only internally. I don't see a problem >> with that. If D-Bus start using it we can change later. Of course it's >> better to convince D-Bus guys to reserve that, but IMO this shouldn't >> be a blocker. >> >> From the chars you suggested, only '@' seems reasonable and I can >> change to this char if we agree on it. > > ?Either way IMO it is ugly to mess around with the signature but > perhaps it is the only alternative we have. There are other ways, but they have shortcomings, too -> the other 2 approaches that I sent to this ML have the shortcoming that in C it's not possible to initialize flexible array members in a nested context. > > I wonder how other binding works in this regard, or they always have > to depend on a xml file that describes the interface? iirc that used > to be the case for GTK and QT, perhaps python is the only one that can > really make use of this in runtime as d-feet seems to be able to use > this information already. As far as I checked, what glib does is to maintain the arguments in a separate array (http://developer.gnome.org/gio/stable/gio-D-Bus-Introspection-Data.html#GDBusMethodInfo-struct). This implies that for each method we need to declare the arguments in a separate structure, not nested as we are doing now. Lucas De Marchi