From: Tadas D. <ta...@da...> - 2009-01-31 00:21:28
|
On Thu, 2009-01-29 at 15:30 +0000, Bastien Nocera wrote: > On Thu, 2009-01-29 at 13:55 +0200, Alex Kanavin wrote: > > 2009/1/29 Bastien Nocera <ha...@ha...>: > > > Using the same code as openobex for the device discovery, and fixing a > > > few memleaks in the process. > > > > > > Patch for obex-data-server attached. > > > > Quite a hack: > > > > +/* Information about a USB OBEX interface present on the system */ > > +struct obex_usb_intf_transport_t { > > > > You copied an internal structure from openobex to ods, how are you > > going to keep it in sync? It's going to change, and I can't make any > > promises about keeping it stable. > > We don't export that structure, and we don't make any use of it through > openobex. See the ods_manager_create_usb_session() function which still > uses the old-style. > > The only thing we do is have our own discovery functions, and export > that data through D-Bus, including the interface number. > > The only thing we have to make sure of is that the discovery function > matches that of openobex, so the devices are enumerated in the same > order. > > I've changed the code to not use the same structures as openobex does > internally, to avoid such confusion. > > > Also, the external structure still > > needs the path member, how do you solve this without patching > > openobex? > > The external structure needs a path member if somebody wants to use > openobex directly for enumeration. > > Cheers This is now in ods svn (also added --disable-usb configure option). Check if everything works as expected since I don't have any devices to check with atm. Cheers |