C++ Logo


Advanced search

Re: Down with "template"! (up with the turbofish)

From: Jean-Baptiste Vallon Hoarau <jeanbaptiste.vallon_at_[hidden]>
Date: Sat, 9 Oct 2021 19:20:00 +0200
Jason :

> I believe his point is that, even though it's "simpler" than
> `template`, you still have to do *something* that you wouldn't
> ordinarily have to do. That is, you still have to know that you need
> disambiguation syntax. As such, your proposal doesn't fix the C++
> parser to take the expected code. It merely provides a different
> spelling of the disambiguated code.

If I could magically make the C++ grammar context-free, I would. However
it's too late for that, the grammar is undecidable and having to use a
disambiguator is the price to pay in some cases. The question is what kind
of disambiguator do we want : one so bad that everyone puts extra work
to avoid needing it, or one that's reasonably usable and can be used
coherently in non ambiguous code.

The problem you're talking about is one that the compiler basically
> tells you how to fix. "It looks like you tried to write a template
> specialization, but you need to put 'template' there." You encountered
> a problem and understand what you need to do to solve it.

I think there's a misunderstanding about the nature of the problem. The
problem is not that people try to write this kind of code and it fails :
this happens once. The problem is that people now have to go out of their
way to workaround something so basic, which happens everyday. How many
times do programmers accidentally step on the "most vexing parse"? Perhaps,
once or twice before they know what's going on. The fix is easy : use "{}"
or "=" to declare and initialize variables. You can use that and forget
about it because the solution is unobtrusive. You can't forget about having
to type ".template" everywhere or turns everything into a
deducible parameters : it's both ugly and requires cognitive load, that you
could have used to do more interesting things than fixing the C++ grammar.

> If having to write 9 extra characters
> is enough to make you fundamentally rewrite code, I don't see how
> having to write 2 characters would make you stop.
Isn't 2 a lot less than 9? I have no trouble seeing how not interspersing
my code with a 9 letter word when most of my identifiers are less than
that, and replacing that with a 2 characters punctuator would improve

 C++ didn't make
> `std::get` a free function because making it a template member of
> `tuple` would make it too ugly to call. We did it because it creates a
> uniform interface for extracting members which other types (like
> `pair`) can participate in.
How would this have been impossible with a member function? Have you ever
written an invalid call to "get<>()" ? A member function would be welcome
here, I don't want the compiler telling me about the 50 overloads in the
std namespace.

Even something like member `get` could be solved by having an overload
> that takes a `std::integral_constant` parameter and using template
> argument deduction to extract the value from it. You'd probably also
> need a special literal suffix to make using integer literals easier to
> use, so you could just use `x.get(0_ce)`.

Sure, I know all about the possible workarounds. You can always hack your
way around language issues. But I don't want a language that makes me work
around itself, I want a language that handles common use cases elegantly.

Le sam. 9 oct. 2021 à 16:16, Gašper Ažman via Std-Proposals <
std-proposals_at_[hidden]> a écrit :

> I've started asking people to pass a std::type_identity<T>{} parameter
> instead of explicit template params for my templated overloads. Way easier
> than requiring people to disambiguate with `template`.
> Then again, I call std::type_identity<T>{} `type<T>`. Way easier to work
> with.
> --
> Std-Proposals mailing list
> Std-Proposals_at_[hidden]
> https://lists.isocpp.org/mailman/listinfo.cgi/std-proposals

Received on 2021-10-09 12:20:17