Directive scope variables using hungarian type notation

[Haemp]

Using variable names prefixed by a single letter seems to have changed in behaviour from 1.1.5 to 1.2.0. In 1.1.5 I used to be able to set oStart as a property name and reference it with o-start, but that seems to now have broken.

Using 1.2.0 http://plnkr.co/edit/EgRNDE?p=preview
Using 1.1.5 http://plnkr.co/edit/r8Q9bj?p=preview

Skimmed through the breaking changes and could't find anything about this and no other related issues as far as I was able to search. Any thoughts?

[caitp]

You're right, the key in the Attributes object is o, even though the attribute name should be oStart. I'm poking through the blame to see what caused that. I can probably have a patch ready shortly.

[caitp]

You know why this is? It's because of the special directive-start and directive-end attributes. It won't apply to attributes which don't end with Start or End... although I can see how that's sort of inconvenient.

              var directiveNName = ngAttrName.replace(/(Start|End)$/, '');
              if (ngAttrName === directiveNName + 'Start') {
                attrStartName = name;
                attrEndName = name.substr(0, name.length - 5) + 'end';
                name = name.substr(0, name.length - 6);
              }

...so, I don't know if this really needs a patch, unless people think this rule needs to be tightened up a bit or something

It is actually mentioned in the breaking changes here

• due to e46100f, existing directives with name ending with "-start" or "-end" will stop working.

  This change was necessary to enable multi-element directives. The best fix is to rename existing directives, so that they don't end with these suffixes.

I agree that this isn't necessarily good, as you may wish to have onDragStart or onDragEnd or similar names, so I think it's probably justifiable to only enforce these rules if the directiveNName actually matches an existing directive... but that's how it is for now I guess

[Haemp]

Ah, totally missed that - well that solves that then. I'm fine with it being that way, just caught me off guard. Thanks for looking into it.

[jeffbcross]

I'm closing this issue, as I think @caitp has captured the problem and solution more accurately in #5372.