Allow the types created for structs to be specified manually instead of always generating them from the struct name #134
Conversation
|
I like this idea, but I think it should probably be a new keyword option to |
|
Sure, that works. Just add |
|
I like |
|
@lexi-lambda ping |
b12374f to
b29edb4
Compare
|
@samth I updated this to use |
| ;; FIXME: unsound, but relied on in core libraries | ||
| ;; #:guard ought to be supportable with some work | ||
| ;; #:property is harder | ||
| (~optional (~seq #:guard guard:expr)) | ||
| (~seq #:property prop:expr prop-val:expr)) | ||
| ...))) | ||
| ...) | ||
| #:attr untyped #`(#,@(if (attribute mutable?) #'(#:mutable) #'()) |
There was a problem hiding this comment.
@samth This bit right here is sort of icky. I’d appreciate any tips on how to do this without the brittle repetition.
There was a problem hiding this comment.
The simplest way would be to use a parameterized splicing syntax class that has the keyword in a list as an attribute, if present, and an empty list otherwise. Also binding it to an attribute of ... depth 1 would let you write the results as something like:
mutable?.kw ...
transparent?.kw ...
and so on.
There was a problem hiding this comment.
I’m not sure I understand. Could you show me a simple example?
There was a problem hiding this comment.
Just composing this in the text window, but something like this:
(define-splicing-syntax-class (opt-kw kw)
[pattern (~seq k:keyword)
#:when (eq? kw (syntax-e #'k))
#:with (kw ...) #'(k)]
[pattern (~seq)
#:with (kw ...) #'()])
and then
(~var mutable (opt-kw '#:mutable))
|
Looks good with comments addressed. |
876e2ca to
81be660
Compare
This allows the types generated by the struct form, as well as #:struct clauses of require/typed, to be specified explicitly using a #:type-name option. This allows the name of a struct and the type it is assigned to be different. Closes racket#261
81be660 to
a3ca5ae
Compare
This change affects all the struct definition forms as well as
#:structclauses inrequire/typed. It's something I've been wanting to change for a while, but I've only now gotten around to doing it. The change is simple and can probably best be explained in code.(struct [point : Point] ([x : Real] [y : Real]))This creates a struct called
point, but the type generated for it isPoint. The change torequire/typedis the same.Capitalization is, unsurprisingly, the main motivator behind this change, which lets both naming conventions be satisfied without needing any type aliases.