903401 - [Contacts] The close button on Merge duplicates screen is kept like pressed on

Status: RESOLVED DUPLICATE of bug 908100

(Firefox OS Graveyard :: Gaia::Contacts, defect)

Description

[isabelrios]

Attachment: 2013-08-09-15-00-30.png

Unagi device master 08/09 build:
Gecko-9fc67ed
Gaia-845abbb
ref ril

STR
At least there should be two contacts in the contacts list with for example same Name and same Phone number
1. Open contacts app
2. Tap on one of the contacts
3. On details view, tap on 'Find duplicates'
4. When the Merge duplicates screen is shown tap on the 'x' button, without lifting the finger from the screen move it to the right.

EXPECTED
The 'x' button should go back to normal after it has been tapped 

ACTUAL
The 'x' button looks like it is pressed althouth it is not. 
Please see screenshot attached

Comment 1

[Joe Cheng [:jcheng] (please needinfo)]

koi+ as it links to koi-comms bug

Comment 2

[Francisco Jordano [:arcturus] [:francisco]]

Adding Salva to the bug :)

Comment 3

[Salvador de la Puente González [:salva]]

This is a very weird bug. This is happening for the buttons inside those views being loaded via the `extensionFrame` in `service_extension.js`. You can reproduce it in the import formulary from Gmail as well.

Comment 4

[Salvador de la Puente González [:salva]]

Hey, Isma, I need some of your magic. The problem is in the presence of classes `view-right` and `view-bottom`.

You can test it by:

1- Comment all the JavaScript in `index.html` and `matching_contacts.html`
2- Add an <iframe> to `index.html` with the src attribute set to `matching_contacts.html`
3- Use CSS to show it over the rest of views.

Now you can see the `matching_contacts.html` layout, check the issue is reproducing. Tap on the X button, drag the finger outside the X and leave the screen to see how the button keeps pressed.

Now remove `view-right` and `view-bottom` classes and check again.

What could it be? Thank you.

Comment 5

[Salvador de la Puente González [:salva]]

Wooow. This was tricky but I think I found the root cause

It is related with position: absolute or position: fixed. When using the CSS snippet:

* {
  margin: 0;
  padding: 0;
}

.view {
  position: absolute;
  top: 0; left: 0; right: 0; bottom: 0;
}

We can already move the .view a little bit. It seems when moving the underlying div and releasing, Gecko is not removing the pseudo-class :active for the buttons. This can be avoided by using position: fixed instead. It has the same practical result in this context.

I'm going prepare a workaround for this bug and opening a platform issue with a simple test application.

Comment 6

[Ismael Gonzalez [:basiclines]]

You didn't give the chance to test it Salva! Glad to see you've find a workaround.
But yes seems to be a platform issue :D thx!

Comment 7

[Salvador de la Puente González [:salva]]

Attachment: Fixing markup and CSS

Comment 8

[Francisco Jordano [:arcturus] [:francisco]]

Comment on attachment 793892 [details]
Fixing markup and CSS

r+ just a small comment on github.

Thank Salva!

Comment 9

[Joe Cheng [:jcheng] (please needinfo)]

ready for landing? thanks

Comment 10

[Salvador de la Puente González [:salva]]

Master: 96b2b6b87433bc0f818de431b63e70c3f3d79135

Sorry for the delay!

Comment 11

[rafael.marquez]

Bug verified with:
Device: unagi
Branch: master 
Gecko: d445929
Gaia: 0395f26

Comment 12

[Vivien Nicolas (:vingtetun) (:21) - (NOT reading bugmails, needinfo? please)]

https://github.com/vingtetun/gaia/commit/29fc735cf923564ceb856d2a2a25f6f66d9467a4

Sorry guys I had to backout this one because of fps scrolling performance regression on the main contact list. Those are due to the switch from |absolute| to |fixed|.

Comment 13

[Salvador de la Puente González [:salva]]

Then please, review the bug 908100.

Comment 14

[Salvador de la Puente González [:salva]]

Apart, )'m getting curious, can you explain to me (or refer me to some discussion / article) why fixed position is affecting performance, please?

Comment 15

[Jose Manuel Cantera]

German,

How is this bug?

Can we close it?

Comment 16

[Germán Toro del Valle (:gtorodelvalle)]

Nope, it still reproduces... :-) I do not know if Salva got the info he requested from Vivien regarding the performance issue, did you Salva? Thanks!

I'll keep as assigned to me until we decide the way to go and who implements the agreed solution.

Comment 17

[Noemí Freire (:noemi)]

Once Bug 908100 (platform one) is fixed no additional changes should be needed on gaia side, at that stage, we could mark this issue as RESOLVED FIXED or WORKSFORME so let's wait till Bug 908100 is fixed in order to close this one.

Comment 18

[Jason Smith [:jsmith]]

(In reply to Noemí Freire (:noemi) from comment #17)
> Once Bug 908100 (platform one) is fixed no additional changes should be
> needed on gaia side, at that stage, we could mark this issue as RESOLVED
> FIXED or WORKSFORME so let's wait till Bug 908100 is fixed in order to close
> this one.

Let's actually dupe in that case - the root problem is platform level, which is tracked in bug 908100.