PhpdocAnnotationWithoutDotFixer - Handle empty line in comment

[kubawerlos]

Solves #3857 (comment) - yet, feels too easy - I'm pessimistic about this one - can we easily trick it? @SpacePossum @dmvdbrugge what do you think?

[dmvdbrugge]

Learning from all the comments I've gotten before, I don't think this is the way to go, it indeed feels a bit hacky.

Also, @SpacePossum said in #3857 he considers this case a priority-issue, for which I made #3879 which should fix not only PhpdocAnnotationWithoutDotFixer but all Phpdoc*Fixers

[dmvdbrugge]

We can see that the fixer works by the removal of the dot, but this should be string $str instead of string string, and then Surprisingly will be lowercased as well.

[SpacePossum]

The solutions, this one and resolving the priority, don't have to be exclusive, sorry if I caused confusion about that.

For this PR, I like the direction 👍
How would something like

    /**
     * @return bool|null Returns `true` if the class has a single-column ID
                        and Returns `false` otherwise.
     */'

be fixed in the new situation?

[kubawerlos]

@SpacePossum not nicely - PhpCsFixer\DocBlock\DocBlock::getAnnotations will lose 2nd line at all.

[SpacePossum]

can you add this to the tests?

    /**
     * @return bool|null Returns `true` if the class has a single-column ID.
                         Returns `false` otherwise.
     */

If the fixer removes the dot after ID but not after otherwise I'm happy, if it doesn't removes any dot than it is fine as well for me. It is nice to have a test to see how the fixer handles invalid PHPDocs, that is all :)

[SpacePossum]

Thanks for the test, this PR looks good to me.
Maybe another member can have look as well?

[keradus]

@kubawerlos , can you rebase please ? we have some conflicts due to recent merging session

[keradus]

Thank you @kubawerlos.