ng-include regression not fully fixed by #5247

[hippotz]

If I have a terminal directive on the same element as ng-include I cannot get access to element when linking occurs inside my directive.

In angular 1.2.1 -> 1.2.2 there was a regression with ng-include that caused this problem for all directives.
#5247 fixed the problem for directives with terminal: false

It is still a problem for directives with terminal: true.

example:
http://plnkr.co/edit/ehqS2VOFTvPLt5Z7iXs1

[petebacondarwin]

I spend a little time today looking at this issue and I haven't come up with a good solution yet.

[IgorMinar]

this is because the content for ngInclude is filled via ngIncludeFillContentDirective: https://github.com/angular/angular.js/blob/master/src/ng/directive/ngInclude.js#L267-L295

this directive has priority -400 which means that it doesn't run because of the terminal directive.

since this issue affects only ngInclude and not $compile I'm ok to push this out to be fixed post 1.3.0

[petebacondarwin]

@IgorMinar & @tbosch - It worries me that ngInclude creates its own scope to be used for the transcluded content. This means that nested transclusion could be broken when ngInclude is involved. Really, what these issues are highlighting is that there is a case for support in $compile of a different kind of "container" directive that effectively provides its content (and collocated lower priority directives) to its template and link function in such a way that we can control how and when this content is compiled.

I believe this can be achieved with a lot of boilerplate code in the ngInclude function by converting it back to a teminal directive and then manually collecting up its contents and lower priority directives into some form that can then be compiled at the correct time in such a way that these lower priority directives effectively appear on the included content, with the content already in place. But this would be a right pain to implement every time and it would be better if it was moved into $compile...

[petebacondarwin]

This seems like an uncommon corner case. I don't think we are going to get round to fixing it in the near future. Assigning to the Ice Box for now.