Re: [RFC] [Discussion] Bound-Erased Generic Types
| From: | Seifeddine Gmati | Date: | Wed, 13 May 2026 19:52:05 +0000 |
| Subject: | Re: [RFC] [Discussion] Bound-Erased Generic Types | ||
| References: | 1 2 | Groups: | php.internals |
| Request: | Send a blank email to internals+get-130873@lists.php.net to get a copy of this message | ||
Hi Daniil,
Thanks for the support. That matters a lot.
> 1) Turbofish at callsite: this was already brought up before, while I understand the precedence
> issues that led to this choice, having written some rust myself, I *still* don’t like the syntax,
> and would very much prefer normal diamond syntax at callsite, like for declarations.
I personally like Turbofish; however, I understand that some people
might not find it "pretty". But the reason it has to stay as turbofish
is parser-level:
[A<B, C>(D)] is ambiguous between a single-element
array calling generic A and a two-element array of comparisons, and
there's no context-sensitive disambiguation that works inside
attributes, array expressions, and ternaries without introducing rules
that bite at exactly the worst places. Forcing parentheses around the
call ([(A<B, C>(D))]) was suggested earlier in the thread and is
uglier than turbofish; it also closes the door on tuple syntax I'd
like to keep open for a future RFC. Turbofish costs two characters and
keeps the parser non-context-sensitive, which is the trade I'd rather
take.
The [self::foo::<Bar>(1), self::foo::<Bar>(2)] case genuinely looks
the worst. The same call without turbofish (`[self::foo(1),
self::foo(2)]`) reads fine and is what the vast majority of call sites
will look like, since turbofish is opt-in for ambiguity resolution
(because inference is hard), and most calls won't need it.
> 2) -/+ for variant bounds: also brought up before, while it *can* be somewhat mnemonic (-
> consumes for input params, + produces for output params), in/out would indeed be much more
> descriptive IMO (or at the very least, both options could be provided at the same time).
Secondary vote added on this. +T/-T vs in T/out
T. I have no
strong preference; the vote will decide.
> I will implement support for this RFC in Psalm immediately after the RFC is approved
Side note: I'm optimistic ;)
https://github.com/carthage-software/mago/pull/1826
Thanks again for the careful read.
Cheers,
Seifeddine.