feat(carousel): hide prev/next if not loop and indexes at bounds

[plondon]

No description provided.

[icfantv]

these changes add two extra watches per slide - can you please investigate combining the ng-show/ng-hide?

[RobJacobs]

I'd prefer ng-if over ng-hide/show. We also need to consider, does it make for a jarring experience with elements being removed rather than just disabling them?

[plondon]

Also a fan of disable, that way users can choose the more jarring experience if they want with CSS. Since disabled is not a valid attribute for an anchor tag is it alright to swap <a role="button"> for the regular <button>?

[icfantv]

@plondon are there accessibility issues with doing that or is that the preferred mechanism? IIRC, using the btn-link class had accessibility ramifications.

[wesleycho]

These need to be condensed into a single function for the show/hide - too much logic in the view here.

[plondon]

@icfantv <button> has the aria role="button" by default, so it should act as <a role="button"> from an accessibility perspective. @realalexhomer anything to add here regarding <button> and accessibility? Thanks

[realalexhomer]

@plondon @icfantv Yes that is correct, screen readers prefer over the a tag so it is better to use it when possible

[plondon]

Since previous and next buttons are "disabled" by the noWrap logic I think we can leave them as anchor tag's.

Instead of button disabled="true" as we were discussing above I used isPrevDisabled() and isNextDisabled() to toggle a disabled class, leaving it up to the user to hide, show, fade, or whatever the buttons him/herself.

I'll squash these commits if you all agree with this approach.

[wesleycho]

This looks solid to me - anyone else have thoughts?

[RobJacobs]

LGTM