Always try form registry first

[Spea]

In 2.8 the getName method from the FormTypeInterface is deprecated, instead the FQDN should be used. When doing so, the FormAnnotationResolver always tries to instantiate the class instead of using the formRegistry.
Since the FormRegistry already handles this case since 2.8, I check the registry if the type exists and get it from the there instead of trying to instantiate it.

This should not break any compatibility.

[mmoreram]

@Spea Thanks for that issue + PR!
Tomorrow morning I will definitely review, merge and tag this PR :)
I appreciate that! ^^

[Spea]

@mmoreram do you need anything else to get this merged?

[mmoreram]

I'm trying to fix some issues related to v3.0.0 compatibility, and there is any tests failing because of this...
Can you add a test covering this?
I will create a PR now with some tests, turning green all tests, and you'll be able to rebase.
After this, I will tag.

Thanks :)

[Spea]

Allright, thx :)

[Spea]

Just out of curiosity. Do you want a functional test, an unit test or both?

[mmoreram]

Well, you've playing more than me on that issue, so feel free. The important thing here is to find an scenario failing without your patch, and passing with it :)

[mmoreram]

Check this :)
#98
You will notice that I've updated the minimum versions in order to work with 2.8 and 3.0. This is not BC Break so, people using these versions, will be able to update. Otherwise, no.
What do you think?

[mmoreram]

@Spea rebase with master and as soon as you find that scenario that fails, we can review this ^^

Thanks :)

[Spea]

@mmoreram ping

I removed the class_exists check completely since the master now only supports > 2.8

[Spea]

@mmoreram I don't want to be pushy, but I kinda depend on this PR. Is it possible to merge this today?

[mmoreram]

Oh, the ping was because the test is done!
Great :)

[mmoreram]

@Spea Excellent, man! Great job! :D

Thanks !

[Spea]

Thanks for merging :)

[mmoreram]

Thanks you for doing :D