@dgt41 looks like you also need to update the php unit tests accordingly.. Travis is complaining

@yvesh yes that's the fun part ?

@dgt41 - I would suggest defining $(this) as a variable (var $self = $(this)) if it's used multiple times in the same function

Before:

var tooltipsSelector = $(this).data('selector'),
    animation = $(this).data('animation'),
    html  = $(this).data('html'),
    placement = $(this).data('placement'),
    title = $(this).data('title'),
    selector2 = $(this).data('other-selector'),
    trigger = $(this).data('trigger'),
    delay = $(this).data('delay'),
    container = $(this).data('container'),
    template = $(this).data('template'),
    $onshow = $(this).data('on-show'),
    $onshown = $(this).data('on-shown'),
    $onhide = $(this).data('on-hide'),
    $onhidden = $(this).data('on-hidden');

After:

var $self = $(this),
    tooltipsSelector = $self.data('selector'),
    animation = $self.data('animation'),
    html = $self.data('html'),
    placement = $self.data('placement'),
    title = $self.data('title'),
    selector2 = $self.data('other-selector'),
    trigger = $self.data('trigger'),
    delay = $self.data('delay'),
    container = $self.data('container'),
    template = $self.data('template'),
    $onshow = $self.data('on-show'),
    $onshown = $self.data('on-shown'),
    $onhide = $self.data('on-hide'),
    $onhidden = $self.data('on-hidden');

Sorry to be nit-picky, but a performance decrease springs to mind when I see creating the same jQuery object loads of times

@dgt41 - You've made some mistakes with $self. There will be some instances where you need to redefine it as it's out of scope:

var $self = $(this);
$('.modalTooltip').each(function(){
    var attr = $self.attr('data-placement');
});

As you can see from the above code, attr will not represent the data-placement attribute value for $('.modalTooltip') because you've defined $self outside each(function() { ...

If you want me to re-do it for you and put it on PasteBin, let me know and I'll send it over

@C-Lodder thanks, I didn't pay much attention I guess

This is exactly why JHtml has a registered callback system.

If you're hell bound and determined to move everything to layouts then JHtml::register(), JHtml::unregister(), JHtml::isRegistered(), JHtml::_(), JHtml::call(), and JHtml::extract() should be deprecated and everything refactored to just call the methods directly.

Until that happens, I honestly just see this PR as adding more complexity to an already complex and poorly constructed system in general. We've already merged a ton of crap code and even crappier layout structures, can we stop doing this? There is no freakin' reason at all to have a layout that has nothing but a closing div or nothing but pure PHP function calls.

@mbabker I reverted that part
@mbabker are you against this PR or was that only the part with the JHtml::_() ?

@dgt41 you know I am against such stuff, against use an empty element to keep the options
it have no profit, really

it force +1 DOM search for each feature, and work with DOM always slower.
Here is huge drawback if compare to inline script: All code bootstrap-init.js will be always executed. Example DOM query $('.joomla-accordion') will be always run, even if accordion not exist on the page, and there more such queries in the file.

And data-blabla="" should be assigned to the element where it used (instead of empty random element), otherwise use scriptOptions.

@Fedik Flexibility comes at a cost. This PR allows overriding of some hidden in the bootstrap.php code. Could it be done otherwise? Yes! In my code I override the bootstrap.php altogether. Maybe this could be broken to individual scripts per component as @yvesh suggested! Then we don't have useless empty elements. About the span/data: in some cases is really hard to not go this way. The idea here was to give some flexibility to template creators to move from 2.3.2 to 3 with minimum effort...

This PR allows overriding of some hidden in the bootstrap.php code

yes I understood it, but the cost is to height, for my opinion ?
There should be a better way, I try to show you some example, in near future, I hope ?

I tried similar in #6357, there "bootstrap feature" runs only if it requested

@Fedik I forgot about #6357, yes that's a better BUT requires the event system which is not available (and probably won't be for 3.x)

@dgt41 check #11671,
can take #6357 without Joomla.Behavior part,
mean only behavior-xxxx.js files and changes in /html/ and do something like:

var accordion = Joomla.getOptions('bootstrap.accordion');
if(accordion){
  // Set up Bootstrap Accordion feature
}

In result we do not have useless DOM query, and do not have inline JavaScript ?

@Fedik @andrepereiradasilva please check

If I add options to the carousel, like this JHtml::_('bootstrap.carousel', 'myCarousel', array("interval" => 500)); nothing change...

Before I changed the interval to 500 the slider slides approximately every 5 second. After I change the interval to 500 this is still the same.

@n9iels I think the problem is that the number is parsed as text, I will review the code tomorrow. Thanks

@n9iels myCarousel is an ID or a class? you have to specified that

I did not specified with a # or something like that. The case I described was based on you example module.

Oops something went horribly wrong here