Closed Bug 961195 Opened 10 years ago Closed 9 years ago

The acceleration sensor and the orientation is not working in Keon devices.

Categories

(Firefox OS Graveyard :: Vendcom, defect)

Product:
Firefox OS Graveyard
Component:
Vendcom
Platform:
x86_64
Linux
Type:
defect
Priority:
Not set
Severity:
major

Tracking

(Not tracked)

Status:
RESOLVED WONTFIX

People

(Reporter: madrid.crespo, Unassigned)

References

Details

(Keywords: regression)

Attachments

(1 file)

left is the test in a Peak device and right is the same test in a Keon device. Notice that they have different Z values.
62.73 KB, image/jpeg
Details 
I have run the following test app:
http://www.html5rocks.com/en/tutorials/device/orientation/devicemotionsample.html

That test only works for B2G 1.4, but similar tests for 1.2 and 1.3 should return the same values.

It shows that Z is 20. But Z should be around 9.8. And since it is not working the accelerometer sensor, the orientation doesn't work too. Keon devices can't change their screen orientation.

/** Keon devices using 1.2 and 1.3 can't change their orientation too, but I was not able to check the coordinates since the test doesn't work in those versions. **/

Discussion thread: https://groups.google.com/forum/#!topic/mozilla.dev.b2g/thPn7CmtCKk
Flags: needinfo?(vwang)
I have tested the data from acceleration sensor in Keon.
The z value is 20 m/s^2 which is a volatile movement to us.
we can not detect the correct orientation in the unstable status.
So it's a geekphone problem.
Flags: needinfo?(vwang)
Luis Cuervo, from Geeksphone said:
"In geeksphone we only compile the sources from mozilla, so if it's was working in 1.2 it should be working in 1.4, because we didnt change any internal data from 1.2 to 1.4. 

So i'm not sure that is a geeksphone problem. And more when you are using 1.4 that is the master branch that is the less stable. 

Test it in 1.3 and 1.2 and if you get the same results, contact me and will search for a fix."
It seems tat doesn't work in any version, not even in 1.2.
Post from Geeksphone developer saying that they didn't change anything:

https://groups.google.com/forum/#!msg/mozilla.dev.b2g/thPn7CmtCKk/ZA0OFcnI1DMJ

Viral, are you sure this is not due to a change by us? That post indicates that it works on 1.2 but not on 1.4, using same base image.
Flags: needinfo?(vwang)
Keywords: regression
I can confirm that this is NOT working in 1.2. It works in 1.1. The orientation lock setting first appeared in 1.2, so perhaps this bug is related to that setting in some way. Also, this bug is a duplicate of #925827:

https://bugzilla.mozilla.org/show_bug.cgi?id=925827

There are certain actions that are really tough to do in permanent portrait mode so we Keon users who hope to move to 1.2, 1.3, and/or 1.4 would greatly appreciate it if someone would solve this. Thanks in advance to anyone from Geeksphone, Mozilla, or the community who can get to the bottom of this issue.
We move screen rotation to use acceleration sensor instead of orientation sensor in bug 788975
I'm guessing z value of acceleration sensor is not correct in the beginning in Keon.
Using the orientation sensor (it comes from acceleration sensor most of time) may just hide the wrong data case in Keon.
Flags: needinfo?(vwang)
Component: General → Vendcom
Testing the v1.4 (build:20140224025811;Git commit:41dd8d1) I notice that compass sensor (Device orientation event) is receiving in fact the rotate acceleration (Device Motion Event)

How to test: 
1) Install any of the apps that show compass  [1]
2) test the app

What happen:
° In the stable 1.1 version it works fine
° In the v1.4 when you rotate the keon in the alpha vector [2] the compass pointer rotate with it


Reference
[1] https://marketplace.firefox.com/search?q=compass
[2] https://developer.mozilla.org/en-US/docs/Web/Guide/Events/Orientation_and_motion_data_explained#Alpha


PD: This bug should be listed like confirmed, no?
Flags: needinfo?
(In reply to Daniel Zorro from comment #10)
> Testing the v1.4 (build:20140224025811;Git commit:41dd8d1) I notice that
> compass sensor (Device orientation event) is receiving in fact the rotate
> acceleration (Device Motion Event)
> 
> How to test: 
> 1) Install any of the apps that show compass  [1]
> 2) test the app
> 
> What happen:
> ° In the stable 1.1 version it works fine
> ° In the v1.4 when you rotate the keon in the alpha vector [2] the compass
> pointer rotate with it
> 
> 
> Reference
> [1] https://marketplace.firefox.com/search?q=compass
> [2]
> https://developer.mozilla.org/en-US/docs/Web/Guide/Events/
> Orientation_and_motion_data_explained#Alpha
> 
> 
> PD: This bug should be listed like confirmed, no?

With the V 1.3 (build:2014031604437; Git commit:0ab8a9cd) The problem still happen
Status: UNCONFIRMED → NEW
Ever confirmed: true
(In reply to Dietrich Ayala (:dietrich) from comment #5)
> Post from Geeksphone developer saying that they didn't change anything:
> 
> https://groups.google.com/forum/#!msg/mozilla.dev.b2g/thPn7CmtCKk/
> ZA0OFcnI1DMJ
> 
> Viral, are you sure this is not due to a change by us?

Dietrich, they didn't change anything, but In fact their acceleration sensor was broken. So when you changed the orientation sensor by the acceleration sensor to rotate the screen, it started to fail because their acceleration sensor was broken and they didn't notice it until it was used.
Flags: needinfo?
(In reply to CodingFree from comment #12)
> (In reply to Dietrich Ayala (:dietrich) from comment #5)
> > Post from Geeksphone developer saying that they didn't change anything:
> > 
> > https://groups.google.com/forum/#!msg/mozilla.dev.b2g/thPn7CmtCKk/
> > ZA0OFcnI1DMJ
> > 
> > Viral, are you sure this is not due to a change by us?
> 
> Dietrich, they didn't change anything, but In fact their acceleration sensor
> was broken. So when you changed the orientation sensor by the acceleration
> sensor to rotate the screen, it started to fail because their acceleration
> sensor was broken and they didn't notice it until it was used.

Stupid question from an impacted user: in that case, would it be an option de start using the orientation sensor again?
(In reply to Jehan Bruggeman from comment #13)
> in that case, would it be an option
> de start using the orientation sensor again?

I guess it could work, but their sensor would be broken anyways. The screen orientation would work, but what about any other app using the acceleration sensor? :)
(In reply to CodingFree from comment #14)
> (In reply to Jehan Bruggeman from comment #13)
> > in that case, would it be an option
> > de start using the orientation sensor again?
> 
> I guess it could work, but their sensor would be broken anyways. The screen
> orientation would work, but what about any other app using the acceleration
> sensor? :)

Good question. It depends what you mean by "broken". If it means "useless", I'd rather as a user at least have the screen orientation feature working, if the acceleration sensor is no use to me anyway!
> Good question. It depends what you mean by "broken". If it means "useless",
> I'd rather as a user at least have the screen orientation feature working,
> if the acceleration sensor is no use to me anyway!

It makes sense, but maybe it was changed by the accel sensor by some reason :\
Any update on this? I think a good fix would look like the following: 

If the phone is a Geeksphone KEON (with broken acceleration sensor), then use the old 1.1 code to detect orientation via the orientation sensor. If the phone is not a Geeksphone KEON (and thus has a working acceleration sensor), then continue to use the current post-1.1 code path. 

Can someone familiar with the codebase and with a working build environment please take a crack at doing this? CodingFree? Any help on this would be very much appreciated. 

Thanks,
-JP
I don't expect this to be fixed. If Geeksphone hasn't do it until now, that means that nobody cares about it...
The Geeksphone staff is very small and overwhelmed. I don't expect them to fix this. But you are wrong that "nobody cares about it." Keon owners care about it, as evidenced by the votes for this bug and by the many comments about the issue on forum.geeksphone.com going back the the first FFOS 1.1 release. Since a Mozilla code change broke the phone (the change from 1.0 to 1.1), and since Mozilla has way more resources than Geeksphone, it would be very much appreciated if someone from Mozilla could step up and support one of their early hardware partners and the users of that phone who took a chance on the platform. I hope you or someone else with the knowledge to fix this will consider doing so. Thanks for listening and thanks for the work you do on this open mobile platform.

-JP
Dear petersens@rocketmail.com,

Maybe do they really care at Geeksphone. But why the hell did they decided to make a new phone (the Revolution? Meaning back to android) while their first one has that kind of bug?
I actually feel betreated by both Geeksphoine and Mozilla.
Ok it's a developer preview, but early adopters should be handled with care.
(In reply to fredericseverino from comment #20)
> Dear petersens@rocketmail.com,
> 
> Maybe do they really care at Geeksphone. But why the hell did they decided
> to make a new phone (the Revolution? Meaning back to android) while their
> first one has that kind of bug?
> I actually feel betrayed by both Geeksphoine and Mozilla.
> Ok it's a developer preview, but early adopters should be handled with care.
Status: NEW → RESOLVED
Closed: 9 years ago
Resolution: --- → WONTFIX
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: