Return-Path: MIME-Version: 1.0 In-Reply-To: <20121112192607.GA6300@x220.P-661HNU-F1> References: <1352385015-2127-1-git-send-email-mikel.astiz.oss@gmail.com> <20121112172346.GA6267@x220.P-661HNU-F1> <20121112192607.GA6300@x220.P-661HNU-F1> Date: Tue, 13 Nov 2012 11:44:03 +0100 Message-ID: Subject: Re: [RFC v0 0/7] Manager/Adapter transition to ObjectManager From: Mikel Astiz To: Mikel Astiz , linux-bluetooth@vger.kernel.org, Mikel Astiz Content-Type: text/plain; charset=ISO-8859-1 Sender: linux-bluetooth-owner@vger.kernel.org List-ID: Hi Johan, On Mon, Nov 12, 2012 at 8:26 PM, Johan Hedberg wrote: > Hi Mikel, > >> > Mikel Astiz (7): >> > manager: Expose default adapter using property >> > manager: Remove redundant D-Bus signals >> > manager: Remove redundant Adapters property >> > adapter: Remove redundant D-Bus signals >> > adapter: Remove DevicesFound signal >> > adapter: Remove redundant Devices property >> > adapter: Remove FindDevice method from D-Bus API >> > >> > doc/adapter-api.txt | 32 +-------- >> > doc/manager-api.txt | 28 +------- >> > src/adapter.c | 187 +--------------------------------------------------- >> > src/manager.c | 103 +++++++---------------------- >> > 4 files changed, 29 insertions(+), 321 deletions(-) >> >> All patches except 5/7 (which needs a bit more discussion) have been >> applied. Thanks. > > I had to revert all of these since I didn't realize you hadn't updated > the test scripts (all of which broke with these patches). Also, it seems > the DefaultAdapter patch is also broken since I never saw that property > (e.g. with test/get-managed-objects). Please send a new patch set once > you've got the scripts converted and the DefaultAdapter patch fixed. > Thanks. You're right, we should keep the test scripts in sync all the time. I tried to warn you about this in the cover-letter but no problem, I will send v2 with the changes you suggest. Regarding the DefaultAdapter property, it was working for me, but I was using the Properties interface. The test script you mention is indeed not reporting the property, but this is because the root object is not a "managed object", and this property is associated to the root. This is no real issue IMO but still, for convenience reasons: do you think we should move the org.bluez.Manager interface to some other object path such as /adapters? As a side effect, this could make it possible that clients filter out signals (i.e. InterfacesAdded) based on the object path. Cheers, Mikel