Deprecated parameter handling

[Clorith]

Per now the devhub doesn't play well with deprecated parameters for function calls, an example of this would be the get_the_author function which used to have a parameter.

The parser currently picks these up as Required, which is a little bit unlucky.

Deprecated parameters seem to consistently be defined with @param $var Deprecated, so detecting these shouldn't be an issue, but how should we relay that to the user in a clear manner, and in such a way that it doesn't accidentally get read as the whole function being deprecated which the current use of $deprecated as the parameter name might lead to.

[Rarst]

Two separate issues:

1. Required is clearly a bug, probably unrelated to deprecated.
2. Deprecated detection for these is tricky. _deprecated_argument() call has no connection with which argument is deprecated.

[Clorith]

For the 2nd bit here, if the Deprecated use in the inline docs is as consistent at it seems to be (i'm sure @DrewAPicture can chime in if this is so) then that would be the ideal thing to match against I'd think?

[DrewAPicture]

It's trivial to determine whether an entire function is deprecated, and that's not in contention here.

As @Rarst correctly pointed out, marking them as Required is clearly a bug, so it sounds like maybe the Optional keyword hasn't been consistently applied. To adjust our standard we need to propose the change at the inline docs chat on Wednesday, but I can see probably doing an audit for all deprecated arguments and coming up with a standard for how to properly mark them up.

Probably the most consistent way would be to add the Deprecated. keyword at the beginning of the description and treat that as Optional (which it is). Once we've made a decision, I can do a quick sweep through core and fix (likely) all of the deprecated argument descriptions at once.

[JDGrimes]

The best way to determine whether an argument is optional or required is to check for its default value. If it doesn't have one, then it is required.

[Rarst]

It's trivial to determine whether an entire function is deprecated, and that's not in contention here.

Yeah. the tricky part is deprecated argument in a non–deprecated function.

[DrewAPicture]

Yeah. the tricky part is deprecated argument in a non–deprecated function.

Yeah, see the entire paragraph I wrote about what we'll need to do to audit core and update the standards specific to deprecated arguments :)

[Rarst]

Not picking on you, just keeping the focus clear in the thread. :) We had topics wander away before. :D