Pagecontainer: Viewport's height

[Palestinian]

If pagecontainer is set to a different wrapper, page scrolls. height: 99.9% should be assigned to ui-mobile-viewport to avoid this issue. http://plnkr.co/edit/qMHTRFIBTSQkMX3Qbog8?p=preview

[Palestinian]

Fixes #7842

[arschmitz]

@Palestinian your change is cause a test failure. Either the test assumptions are no longer valid and needs to be updates or your change has broken something.

[Palestinian]

@arschmitz Shall I close this and create a new PR?

[arschmitz]

No just find out why the tests are failing and fix them though we need to discuss still if this is a change we want in #7842

[arschmitz]

@Palestinian Bump any progress on this?

[Palestinian]

@arschmitz No, honestly not. I tried once and it failed. I didn't know what to do next.

What do you recommend?

[arschmitz]

It looks like tests related to silent scroll i would look at the failing test and see if you can see why your change would make it fail. Then determine if the test just needs to be updated or if your change really is breaking something. Running the tests locally is probably the best way to debug this.

[gabrielschulhof]

@arschmitz this test fails on Chrome desktop even on master.

[gabrielschulhof]

Chrome has a rounding error. If you window.scrollTo( 0, 1 ), $( window ).scrollTop() returns 0.9091.

[gabrielschulhof]

OK, that rounding error seems to be fixed in a bleeding-edge version of Chrome.

[arschmitz]

@gabrielschulhof but it does not fail on master on phantom so something is still wrong here

[gabrielschulhof]

@arschmitz Yeah, it was comparing the result of silentScroll() to 0, whereas it should've been $.mobile.defaultHomeScroll. My modified version of the test works on both master and this branch.

[arschmitz]

@jaspermdegroot can you think of any potential problems with this?

[jaspermdegroot]

@arschmitz

We should not land this PR, because it's not the right fix for the problem.

The reason we set height for the html and body element is to make min-height: 100%; for the page (when it's an immediate child of body) work, to give the page gets same height as the viewport even before the JS sets a min-height with pixel value as inline-style.
We shouldn't make any changes that could result in setting it for anything else than html and body.

The actual problem in @Palestinian his test page is the default browser style for body (margin top/bottom, padding left/right). What we could do is change it to:

.ui-mobile,
.ui-mobile body {
    height: 99.9%;
    margin: 0;
    padding: 0;
}

I am closing this PR. @Palestinian can you open a ticket for the issue, please? Do you want to create a new PR?

[Palestinian]

@jaspermdegroot thank you! padding and margin set to 0 is the solution as you've mentioned in your previous comment.

@arschmitz @gabrielschulhof Thank you both for sparing some time to look into this issue, I appreciate it :)