using XHANDLE = std::unique_ptr<void, decltype([]( void *h ) { h !=
INVALID_HANDLE_VALUE && h && CloseHandle( h ); })>;

A Win32 HANDLE-wrapper in just one line in C++20.
C++20 allows lambdas in an unevaluated context.

On 2023-11-05, Bonita Montero <Bonita.Montero@gmail.com> wrote:
> using XHANDLE = std::unique_ptr<void, decltype([]( void *h ) { h !=
> INVALID_HANDLE_VALUE && h && CloseHandle( h ); })>;
>
> A Win32 HANDLE-wrapper in just one line in C++20.
> C++20 allows lambdas in an unevaluated context.

So what?

1960 Lisp allows lambdas in unevaluated context.

(QUOTE (LAMBDA (H)
(AND (NOT (INVALIDHANDLE H))
(CLOSEHANDLE H))))

This is just C++ catching up to something that existed when
Stroustrup was in elementary school.

--
TXR Programming Language: http://nongnu.org/txr
Cygnal: Cygwin Native Application Library: http://kylheku.com/cygnal
Mastodon: @Kazinator@mstdn.ca
NOTE: If you use Google Groups, I don't see you, unless you're whitelisted.

Kaz Kylheku <864-117-4973@kylheku.com> writes:

> On 2023-11-05, Bonita Montero <Bonita.Montero@gmail.com> wrote:
>> using XHANDLE = std::unique_ptr<void, decltype([]( void *h ) { h !=
>> INVALID_HANDLE_VALUE && h && CloseHandle( h ); })>;
>>
>> A Win32 HANDLE-wrapper in just one line in C++20.
>> C++20 allows lambdas in an unevaluated context.
>
> So what?
>
> 1960 Lisp allows lambdas in unevaluated context.
>
> (QUOTE (LAMBDA (H)
> (AND (NOT (INVALIDHANDLE H))
> (CLOSEHANDLE H))))
>
> This is just C++ catching up to something that existed when
> Stroustrup was in elementary school.

I was going to say the usage here is not a good look for C++. As a
language with compile-time type checking why is decltype needed? It's
usually because C++ can't express the type involved any other way and
that's not something to shout about.

But then it turns out I don't know what's going on. This program:

#include <iostream>
#include <memory>

using uptr_int = std::unique_ptr<int, decltype([](int *ip) {
std::cout << "delete (" << *ip << ")\n";
})>;

int main()
{
int x = 42;
uptr_int ip(&x);
}

prints "delete (42)". Why? I thought decltype was just being used to
give a type for the std::unique_ptr template instantiation, but it seems
to supply the function too. What's going on?

Having used C++ from the early CFront days, and watched balloon in size
and complexity, I knew the day would come when even a short C++ program
would be mystery to me. That day has come!

--
Ben.

On 05/11/2023 22:42, Ben Bacarisse wrote:
> Kaz Kylheku <864-117-4973@kylheku.com> writes:
>
>> On 2023-11-05, Bonita Montero <Bonita.Montero@gmail.com> wrote:
>>> using XHANDLE = std::unique_ptr<void, decltype([]( void *h ) { h !=
>>> INVALID_HANDLE_VALUE && h && CloseHandle( h ); })>;
>>>
>>> A Win32 HANDLE-wrapper in just one line in C++20.
>>> C++20 allows lambdas in an unevaluated context.
>>
>> So what?
>>
>> 1960 Lisp allows lambdas in unevaluated context.
>>
>> (QUOTE (LAMBDA (H)
>> (AND (NOT (INVALIDHANDLE H))
>> (CLOSEHANDLE H))))
>>
>> This is just C++ catching up to something that existed when
>> Stroustrup was in elementary school.
>
> I was going to say the usage here is not a good look for C++. As a
> language with compile-time type checking why is decltype needed? It's
> usually because C++ can't express the type involved any other way and
> that's not something to shout about.
>
> But then it turns out I don't know what's going on. This program:
>
> #include <iostream>
> #include <memory>
>
> using uptr_int = std::unique_ptr<int, decltype([](int *ip) {
> std::cout << "delete (" << *ip << ")\n";
> })>;
>
> int main()
> {
> int x = 42;
> uptr_int ip(&x);
> }
>
> prints "delete (42)". Why? I thought decltype was just being used to
> give a type for the std::unique_ptr template instantiation, but it seems
> to supply the function too. What's going on?
>
> Having used C++ from the early CFront days, and watched balloon in size
> and complexity, I knew the day would come when even a short C++ program
> would be mystery to me. That day has come!
>

I think the type of a lambda encodes the definition of the lambda (so
that each lambda function has a unique type, perhaps). But I don't know
why it appears to be executing the lambda.

Am 05.11.2023 um 21:02 schrieb Kaz Kylheku:
> On 2023-11-05, Bonita Montero <Bonita.Montero@gmail.com> wrote:
>> using XHANDLE = std::unique_ptr<void, decltype([]( void *h ) { h !=
>> INVALID_HANDLE_VALUE && h && CloseHandle( h ); })>;
>>
>> A Win32 HANDLE-wrapper in just one line in C++20.
>> C++20 allows lambdas in an unevaluated context.
>
> So what?
>
> 1960 Lisp allows lambdas in unevaluated context.
>
> (QUOTE (LAMBDA (H)
> (AND (NOT (INVALIDHANDLE H))
> (CLOSEHANDLE H))))

This is figuratively the definition of a lambda in C++,
but not the derivation of the type, which then maintains
the identity of the call operator.

On 2023-11-06, Bonita Montero <Bonita.Montero@gmail.com> wrote:
>> (QUOTE (LAMBDA (H)
>> (AND (NOT (INVALIDHANDLE H))
>> (CLOSEHANDLE H))))
>
> This is figuratively the definition of a lambda in C++,
> but not the derivation of the type, which then maintains
> the identity of the call operator.

This smells like AI-generated word salad?

Aha!

MORON AI BOT NET
||||| || B|| |||
||||| || O| |||
||||| || | N||
||||| |I | ||
||||| | | |T
||||| A | |
M|||| | |
O||| | |
||N | |
|| T |
|| E
R|
O

--
TXR Programming Language: http://nongnu.org/txr
Cygnal: Cygwin Native Application Library: http://kylheku.com/cygnal
Mastodon: @Kazinator@mstdn.ca
NOTE: If you use Google Groups, I don't see you, unless you're whitelisted.

Am 06.11.2023 um 18:50 schrieb Kaz Kylheku:
> On 2023-11-06, Bonita Montero <Bonita.Montero@gmail.com> wrote:
>>> (QUOTE (LAMBDA (H)
>>> (AND (NOT (INVALIDHANDLE H))
>>> (CLOSEHANDLE H))))
>>
>> This is figuratively the definition of a lambda in C++,
>> but not the derivation of the type, which then maintains
>> the identity of the call operator.
>
> This smells like AI-generated word salad?
>
> Aha!
>
> MORON AI BOT NET
> ||||| || B|| |||
> ||||| || O| |||
> ||||| || | N||
> ||||| |I | ||
> ||||| | | |T
> ||||| A | |
> M|||| | |
> O||| | |
> ||N | |
> || T |
> || E
> R|
> O
>

If you don't understand the meaning its moronic. You've shown
a pure lambda, I've shown how the identity of the lambda is
injected into a class without an actual function object.

On 2023-11-06, Bonita Montero <Bonita.Montero@gmail.com> wrote:
> If you don't understand the meaning its moronic. You've shown
^^^

pron. possessive + be -> contr.

he his he is -> he's
she her(s) she is -> she's
you your(s) you are -> you're

it its it is -> it's

What language model are you using?

--
TXR Programming Language: http://nongnu.org/txr
Cygnal: Cygwin Native Application Library: http://kylheku.com/cygnal
Mastodon: @Kazinator@mstdn.ca
NOTE: If you use Google Groups, I don't see you, unless you're whitelisted.

Am 06.11.2023 um 19:59 schrieb Kaz Kylheku:
> On 2023-11-06, Bonita Montero <Bonita.Montero@gmail.com> wrote:
>> If you don't understand the meaning its moronic. You've shown
> ^^^
>
> pron. possessive + be -> contr.
>
> he his he is -> he's
> she her(s) she is -> she's
> you your(s) you are -> you're
>
> it its it is -> it's
>
> What language model are you using?

I can imagine that you didn't understand what I said because of that.

On 2023-11-06, Bonita Montero <Bonita.Montero@gmail.com> wrote:
> I can imagine

I've not seen evidence to support this.

On 11/6/2023 9:50 AM, Kaz Kylheku wrote:
> On 2023-11-06, Bonita Montero <Bonita.Montero@gmail.com> wrote:
>>> (QUOTE (LAMBDA (H)
>>> (AND (NOT (INVALIDHANDLE H))
>>> (CLOSEHANDLE H))))
>>
>> This is figuratively the definition of a lambda in C++,
>> but not the derivation of the type, which then maintains
>> the identity of the call operator.
>
> This smells like AI-generated word salad?
>
> Aha!
>
> MORON AI BOT NET
> ||||| || B|| |||
> ||||| || O| |||
> ||||| || | N||
> ||||| |I | ||
> ||||| | | |T
> ||||| A | |
> M|||| | |
> O||| | |
> ||N | |
> || T |
> || E
> R|
> O
>

No shit!

On 06/11/2023 20:10, Kaz Kylheku wrote:
> On 2023-11-06, Bonita Montero <Bonita.Montero@gmail.com> wrote:
>> I can imagine
>
> I've not seen evidence to support this.

Do all threads in this newsgroup have to descend to a kindergarten level
of insults?

Whatever people may think of Bonita's original post in this thread, it
inspired a rather profound and interesting question from Ben. I for one
would be happy if the knowledgable regulars here spent more time
answering questions like his and less time bickering.