Fix app menu calculation for random size of the right header

[juliusknorr]

Fixes#10857

This PR will ensure that at least 210px of space is kept available for the right header icons. The issue in #10857 was that the number of icons to show was calculated before all right icons where loaded. This leads to different behaviour when resizing the window/loading an app that doesn't have a search.

[skjnldsv]

Couldn't we just make sure the calculation is always done properly?
Or just so that icons always have the same width, so that way, even if they're now loaded, the width would be ok?

[weeman1337]

As @juliushaertl told me adding icons afterwards is working. It's just for the initial loading.

[skjnldsv]

Yes, but I don't like having hardcoded values. We shouldn't have to do that :)
Can we watch the header changes with an event, or just add another trigger to ensure we wait for things to be loaded before having to launch the calculation?

[juliusknorr]

Yes, but I don't like having hardcoded values. We shouldn't have to do that :)
Can we watch the header changes with an event, or just add another trigger to ensure we wait for things to be loaded before having to launch the calculation?

I have no idea how we could do that properly. The issue is that the two divs in the header (left/right) are resized when some icons are added. Do you have an idea how we could properly calculate the icon count every time @weeman1337@skjnldsv ?

[weeman1337]

Would be easier if there is a explicit API for adding removing stuff. Maybe we should go for a fix now and improve the header bar in the future (followup)?

[rullzer]

@skjnldsv@juliushaertl can we get a decision here. I have to agree with @weeman1337 to get this in now and fix properly for 15. We are in RC so no huge changes please 😉

[skjnldsv]

LGTM then :)

[MorrisJobke]

Tested and works 👍

[MorrisJobke]

JSUnit fails:

PhantomJS 2.1.1 (Linux 0.0.0) Core base tests Main menu mobile toggle Clicking menu toggle toggles navigation in FAILED
Expected false to equal true.
core/js/tests/specs/coreSpec.js:556:46

[rullzer]

@juliushaertl see #10890 (comment)

[MorrisJobke]

@juliushaertl Ping ;)

[juliusknorr]

Tests should be fixed, let's see.

[rullzer]

Seems good right @juliushaertl ?

[juliusknorr]

Yep, remaining failures are unrelated. 😉

[MorrisJobke]

@juliushaertl Mind to backport this?