961675 - Send device information to Everything.me Partners API

Status: RESOLVED FIXED

(Firefox OS Graveyard :: Gaia::Everything.me, defect)

Comment 1

[Amir Nissim (:amirn)]

work started, waiting for keyboard.current replacement for 1.4 (bug 930402)
https://github.com/EverythingMe/gaia/commit/5834196de238f7f4e3298d8f29027b67c2046c9c

Comment 2

[Tim Guan-tin Chien [:timdream] (please needinfo)]

Hi Amir!

There isn't any proposed replacement for now. Probing |keyboard.current| will continue to work with built-in keyboard. Not sure about the new keyboard through, looping :djf.

Comment 3

[David Flanagan [:djf]]

Tim,

Thanks for looping me in. As we move to 3rd party keyboards that are OOP and not certified, the keyboard app will not be able to set keyboard.current.  Would system/js/keyboard_manager.js be the logical place to set this?

Amir, 

Does e.me really find keyboard.current useful?  If an English speaker in in San Francisco switches to a Chinese keyboard to send a text to a friend in Taiway, do you really want e.me results to change based on that transient keyboard change? What if the user forgets to switch their keyboard back to English and are then puzzled when e.me gives strange results?

Comment 4

[Amir Nissim (:amirn)]

(In reply to David Flanagan [:djf] from comment #3)
> Does e.me really find keyboard.current useful?  If an English speaker in in
> San Francisco switches to a Chinese keyboard to send a text to a friend in
> Taiway, do you really want e.me results to change based on that transient
> keyboard change? What if the user forgets to switch their keyboard back to
> English and are then puzzled when e.me gives strange results?

Keyboard.current is a very helpful hint for providing better results.
Some of the uses are:
1. autocomplete suggestions
2. spelling correction
3. stopwords identification
4. disambiguate words that are used in multiple languages

In your example, in case the user switches to Chinese keyboard and forgets to switch back, she would be searching in Chinese - so showing Chinese results would be suitable indeed.

Comment 5

[Amir Nissim (:amirn)]

Attachment: PR

Comment 6

[Amir Nissim (:amirn)]

David, is |keyboard.current| deprecated in master? is there an alternative?

Comment 7

[Amir Nissim (:amirn)]

Comment on attachment 8430041 [details] [review]
PR

Please see comments on GH.

Comment 8

[Kevin Grandon :kgrandon]

Comment on attachment 8430041 [details] [review]
PR

It's a bit hard to find the relevant parts due to a few extra commits in there. I think you should rebase against upstream master to get rid of the extra commits. Thanks!

Comment 9

[Amir Nissim (:amirn)]

dropped keyboard.current and tracking carrier change:
https://github.com/EverythingMe/gaia/commit/7d743d89c1848bb0c4646cf7e910c2e160c90ebf

Comment 10

[Amir Nissim (:amirn)]

:gandalf is using |language.current| from settings any different than |navigator.mozL10n.language.code| ?

Comment 11

[Amir Nissim (:amirn)]

landed: https://github.com/mozilla-b2g/gaia/commit/e198f882f3eac0eeff84ecb5f90265056d9e0c2c

Comment 12

[Zibi Braniecki [:zbraniecki][:gandalf]]

(In reply to Amir Nissim (Everything.me) from comment #10)
> :gandalf is using |language.current| from settings any different than
> |navigator.mozL10n.language.code| ?

It should not be, but if you want to observe language changes you should use the new API onlanguagechange+navigator.language instead.

Landed in bug 889335. In bug 1013929 we transitioned to use it in mozL10n. I'd recommend you transitioning away from mozSettings' language.current toward it.

Comment 13

[Kevin Grandon :kgrandon]

Mass modify - set status-b2g-v2.0 fixed for fixed bugs under vertical homescreen dependency tree.