What to Call a "NULL means you need TRY" Annotation?

Development Naming
1

With the old policy known as "BLANK-in, NULL-out", it was desirable to keep you from having to manually write this pattern:

something: function [arg [blank! file!]] [
     if blank? arg [return null]
     ...  ; actual code
]

So an annotation with this behavior was needed. What to call it?

  • <nullifies>

  • <revokes>

  • <retracts>

  • If it were Zen it could be <unask> or <mu> (or <无>) for "unasking a question".

Because something had to be picked, the suboptimal choice of <blank> has been used for the past 3 years:

something: function [arg [<blank> file!]] [
     ...  ; actual code
]

It did the job, though it would leave anyone unfamiliar scratching their head about what the difference between blank! and <blank> was.

Now It's "NULL in...NULL out" (but the call needs TRY)

This is about as exciting a change as we get around here--and frees BLANK! up for more interesting dialecting without getting entangled in this.

What you want to save time writing is now:

something: function [arg [<opt> file!]] [
     if null? arg [return raise make error! [id: 'try-if-null-meant]]
     ...  ; actual code
]

(Let's put aside that <opt> should probably be <null>.)

One fairly obvious-seeming name for the annotation would be <try>.

something: function [arg [<try> file!]] [
     ...  ; actual code
]

It's a little bit weird, because it's saying "if someone passes a NULL to to ARG, they're going to have to use TRY SOMETHING"

Although it's a weird connection, it definitely makes the connection. So I'm going to go with it.

Efficiency Note: The Function Isn't Called At All

Something to know about this trick is that it actually is good for performance. The function is bypassed completely--it typechecks the arguments but then the frame is discarded!

2

No objection here. Seems fine but I'll reserve judgment for those like @rgchris who are busy absorbing and applying these new conventions.

1 Like