[Behat] Improve CommonActions and Authentication context Css selectors

[miguelcleverti]

In order to stop fetching the elements using their text value the css selectors needed some changes.

This PR adds some specific tag to the PlatformUI markup in order to better fetch the elements. Also some unused methods were cleaned up, since their were not being used anymore.

[miguelcleverti]

ping @bdunogier @dpobel @andrerom @joaoinacio

[andrerom]

looks like a good direction, as for the details here @bdunogier would have more insight on bdd part and @dpobel / @yannickroger on templates parts.

[dpobel]

what is the point of that ? ok you are not searching by text content, but you are still searching by text which can be translated or changed for whatever reason. Actually, each navigation item has a unique identifier so we should probably expose it somewhere so that you can use it in the tests.

[dpobel]

Created https://jira.ez.no/browse/EZP-26120 for that.

[miguelcleverti]

@dpobel okay, I have already done this, should I include it in this PR or in a separate one?

[miguelcleverti]

I made it in a separate PR so it's easier to review #654

[miguelcleverti]

ping @dpobel, @bdunogier. After trying some different way I found that in order to use the node-id in content tree elements would need a lot of code scattered by multiple contexts. I made a some tests and moving the title attribute to the parent element does not seem to break anything and this simplify the code greatly. What do you think?

[bdunogier]

Could you link to the change you're referring to @miguelcleverti ?

[miguelcleverti]

@bdunogier, @dpobel these changes

[dpobel]

what's the point of moving the title from the link to the list item ? I mean you can find the link and just retrieve the parent node to get the list item. I'm saying that because with this change, you'll end up with something like:

<li title="A nice Content in the tree">
    <a href="#" <!-- ... -->>A nice Content in the tree</li>
    <ul class="ez-tree-level">
        <li title="A subitm of a nice Content in the tree">
        <!-- ... -->
        </li>
    </ul>
</li>

then the title is also associated with the subitems (the ul) of a given Location which is a bit weird semantically speaking. Also the li represents a Location in the tree while the a represents the Content item so the name (title) should stay at the Content item level since a Location has no name.

[bdunogier]

Looks okay to me. @dpobel ?

[miguelcleverti]

ping @dpobel @bdunogier @joaoinacio. Removed the previous changes to the title attribute in the content tree elements. Now the selecting of the elements is done using the data-node-id and the node id is fetched using the search service.

[andrerom]

Seems you might want to rebase the PR or something, as there are some behat failures here.

[miguelcleverti]

All the Behat test failures are due to Deprecation warning. It is ready to be reviewed.

[andrerom]

Master is not failing on behat, could you rebase on top of latests? and if you introduce deprecation issues please fix them ;)

Besides that +1 so far :)

[miguelcleverti]

PR approved by QA