Bluez DBUS API usuability regarding end user experience

Bluez-devel Fabien Chevalier 3 participants 3 messages Oct 25, 2007 – Oct 26, 2007 Bluetooth drivers

lore

Share & tools

lore

Fabien Chevalier

author

2007-10-25 18:23 UTC #0 lore raw

Marcel,

I have a (kind of) formal request to ask to the Bluez project.

I'm basically trying to use the Bluez API to design a user friendly device that will have bluetooth functionnality.
Speeking of a user friendly device, i'm thinking of two things i wanna have:
- the end user bluetooth feature should work most of the time, which
means it will have to be robust and not throw errors most of the time.
- When something goes wrong, the end user should only be presented
with error messages he can understand and actually act upon. If the message is too cryptic, then a generic placeholder would be use instead of a low level message. One of the first non cryptic messages i think of would be a "Device unreachable : please check your bluetooth device is powered on" message :-)

I've done a quick review of hcid an audio service code, and did some testing too. What i've found is:
- for hcid case: the closed match i found is the
org.audio.bluez.ConnectionAttemptFailed exception. This one seems to be raised when the L2CAP connection fails. But it is not specific on the
failure cause ( i think what i'm looking for is EHOSTDOWN posix error code, please correct me if i'm wrong ;-) )
- for Audio service the case is different. Sink object Connect() dbus
call would fail with org.bluez.audio.Error.Failed while the Headset object Connect() dbus call would fail with org.bluez.audio.Error.ConnectFailed. Those two are not even consistent :-( headset.c code looks even a bit hackish, as it tryes to regenerate the lost EHOSTDOWN error code by hand :

if (dbus_error_has_name(&derr, "org.bluez.Error.ConnectionAttemptFailed"))

	err_connect_failed(device->conn, c->msg,
					strerror(EHOSTDOWN));

My point of view is that this common case of error should be clearly identified and sent up to the application.

What i would like to do is propose a patch that would:
- Define org.bluez.Error.DeviceUnreachable exception for hcid. This
would be fired each time a connect on an l2cap socket would return EHOSTDOWN.
- Define org.bluez.Error.Audio.DeviceUnreachable for use by the audio
service. This one would be sent by the Headset and Sink objects instead of the org.bluez.audio.Error.Failed and org.bluez.audio.Error.ConnectFailed when the l2cap/rfcomm/sco connect() code returns EHOSTDOWN.

Marcel, would you agree this is the right way to proceed ? Would you be willing to accept patches that move the code in that direction ? ( I mean : shall i start coding now ? ;-) )

Regards,

Fabien

begin:vcard
fn:Fabien CHEVALIER
n:CHEVALIER;Fabien
org:SILICOM
adr:;;4 rue de Jouanet; RENNES ATALANTE;;35700;FRANCE email;internet:[email protected] title:Software & Studies Engineer tel;work:+33 (0) 2 99 84 17 17
version:2.1
end:vcard

_______________________________________________
Bluez-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/bluez-devel

Marcel Holtmann

maintainer

2007-10-25 22:43 UTC #1 lore raw

Hi Fabien,

Quoted 13 lines

I have a (kind of) formal request to ask to the Bluez project.
I'm basically trying to use the Bluez API to design a user friendly device that will have bluetooth functionnality.
Speeking of a user friendly device, i'm thinking of two things i wanna have:
- the end user bluetooth feature should work most of the time, which
means it will have to be robust and not throw errors most of the time.
- When something goes wrong, the end user should only be presented
with error messages he can understand and actually act upon. If the message is too cryptic, then a generic placeholder would be use instead of a low level message. One of the first non cryptic messages i think of would be a "Device unreachable : please check your bluetooth device is powered on" message :-)

feel free to improve the error messages. We haven't spend enough time with them yet. There were other issues that needed to be addressed first. Also the error support should made generic and reside in common/. So if you wanna work on it, send patches.

Regards

Marcel

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now >> http://get.splunk.com/
_______________________________________________
Bluez-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/bluez-devel

Fabien Chevalier

2007-10-26 09:25 UTC #2 lore raw

Quoted 16 lines · Marcel Holtmann

Hi Fabien,
Quoted 13 lines
I have a (kind of) formal request to ask to the Bluez project.
I'm basically trying to use the Bluez API to design a user friendly device that will have bluetooth functionnality.
Speeking of a user friendly device, i'm thinking of two things i wanna have:
- the end user bluetooth feature should work most of the time, which
means it will have to be robust and not throw errors most of the time.
- When something goes wrong, the end user should only be presented
with error messages he can understand and actually act upon. If the message is too cryptic, then a generic placeholder would be use instead of a low level message. One of the first non cryptic messages i think of would be a "Device unreachable : please check your bluetooth device is powered on" message :-)

Quoted 4 lines

feel free to improve the error messages. We haven't spend enough time with them yet. There were other issues that needed to be addressed first. Also the error support should made generic and reside in common/. So if you wanna work on it, send patches.

Hi Marcel,

I didn't though error handling was that bad. :-(
And... yes i volunteer to work on this, so i'm gonna start to work on generic error support in common/. :-)

Cheers,

Fabien