• Dennis Ritchie (6/6) Mar 09 2015 Hi.
• Adam D. Ruppe (9/10) Mar 09 2015 that's not an array, use [1,2,3] for that.
• John Colvin (12/18) Mar 09 2015 I think this is a misunderstanding about UFCS.
• Dennis Ritchie (4/15) Mar 09 2015 Thanks.
• ketmar (4/6) Mar 09 2015 or at least warn about it. producing a warning for such expression will=...
• Jonathan M Davis via Digitalmars-d-learn (9/15) Mar 09 2015 There's not much point in that. If someone is using the comma operator
• ketmar (7/25) Mar 09 2015 i remember that deprecation was rejected. maybe this is false memory,=20
• Meta (4/12) Mar 09 2015 I think the last time there was a big discussion about this
• Jonathan M Davis via Digitalmars-d-learn (7/20) Mar 10 2015 Yeah. Even Java has that. But IIRC, it was decided that the comma operat...
• ketmar (4/7) Mar 10 2015 sometimes i'm irritated by this, but with `delete`... ah, i hope nobody=...

"Dennis Ritchie" <dennis.ritchie mail.ru> writes:

Hi.
Why prints only the last element?

import std.stdio;

void main() {

	(1, 2, 3).writeln;	// prints 3
}

"Adam D. Ruppe" <destructionator gmail.com> writes:

On Monday, 9 March 2015 at 14:38:41 UTC, Dennis Ritchie wrote:
 	(1, 2, 3).writeln;	// prints 3

that's not an array, use [1,2,3] for that.

What you wrote is a parenthesis expression with comma expressions 
inside. The comma operator is a bit weird: it works like a 
semicolon, but in different contexts.

So 1,2,3, in this context, means "calculate 1, then calculate 2, 
then calculate 3, the result is the final one: 3". The 1 and 2 
are just discarded because their calculations don't save anything 
outside.

"John Colvin" <john.loughran.colvin gmail.com> writes:

On Monday, 9 March 2015 at 14:38:41 UTC, Dennis Ritchie wrote:
 Hi.
 Why prints only the last element?

 import std.stdio;

 void main() {

 	(1, 2, 3).writeln;	// prints 3
 }

I think this is a misunderstanding about UFCS.

1.writeln(2, 3);

is the same as

writeln(1, 2, 3);

and the same as

(1).writeln(2, 3);

but

(1, 2, 3).writeln;

is completely different. UFCS is only for the first argument.

Yet another excellent example of why we should abolish the comma 
operator in D.

"Dennis Ritchie" <dennis.ritchie mail.ru> writes:

On Monday, 9 March 2015 at 14:42:35 UTC, Adam D. Ruppe wrote:
 On Monday, 9 March 2015 at 14:38:41 UTC, Dennis Ritchie wrote:
 	(1, 2, 3).writeln;	// prints 3

 that's not an array, use [1,2,3] for that.

 What you wrote is a parenthesis expression with comma 
 expressions inside. The comma operator is a bit weird: it works 
 like a semicolon, but in different contexts.

 So 1,2,3, in this context, means "calculate 1, then calculate 
 2, then calculate 3, the result is the final one: 3". The 1 and 
 2 are just discarded because their calculations don't save 
 anything outside.

Thanks.

On Monday, 9 March 2015 at 14:48:20 UTC, John Colvin wrote:
 I think this is a misunderstanding about UFCS.

Yes.

ketmar <ketmar ketmar.no-ip.org> writes:

On Mon, 09 Mar 2015 14:48:19 +0000, John Colvin wrote:

 Yet another excellent example of why we should abolish the comma
 operator in D.

or at least warn about it. producing a warning for such expression will=20
ring a bell in programmers head: "compiler is unhappy. i definitely doing=20
something wrong here!"=

Jonathan M Davis via Digitalmars-d-learn writes:

On Monday, March 09, 2015 15:45:58 ketmar via Digitalmars-d-learn wrote:
 On Mon, 09 Mar 2015 14:48:19 +0000, John Colvin wrote:

 Yet another excellent example of why we should abolish the comma
 operator in D.

 or at least warn about it. producing a warning for such expression will
 ring a bell in programmers head: "compiler is unhappy. i definitely doing
 something wrong here!"

There's not much point in that. If someone is using the comma operator
correctly, then the warnings would be incredibly annoying. And if we value
avoiding comma operator-related bugs over letting folks take advantage of
its usefulness enough to consider such a warning, we might as well just
deprecate the comma operator and move towards removing it from the language.
Either it's there and useful but error-prone, or it's gone. In between just
annoys both groups.

- Jonathan M Davis

ketmar <ketmar ketmar.no-ip.org> writes:

On Mon, 09 Mar 2015 13:18:58 -0700, Jonathan M Davis via
Digitalmars-d-learn wrote:

 On Monday, March 09, 2015 15:45:58 ketmar via Digitalmars-d-learn wrote:
 On Mon, 09 Mar 2015 14:48:19 +0000, John Colvin wrote:

 Yet another excellent example of why we should abolish the comma
 operator in D.

 or at least warn about it. producing a warning for such expression will
 ring a bell in programmers head: "compiler is unhappy. i definitely
 doing something wrong here!"

=20
 There's not much point in that. If someone is using the comma operator
 correctly, then the warnings would be incredibly annoying. And if we
 value avoiding comma operator-related bugs over letting folks take
 advantage of its usefulness enough to consider such a warning, we might
 as well just deprecate the comma operator and move towards removing it
 from the language.
 Either it's there and useful but error-prone, or it's gone. In between
 just annoys both groups.

i remember that deprecation was rejected. maybe this is false memory,=20
though.

btw, there are legit uses of comma, in c-style `for`, for example. this=20
should be left intact, i think (oh, can c-style `for` be deprecated too?!
).=

"Meta" <jared771 gmail.com> writes:

On Monday, 9 March 2015 at 22:00:46 UTC, ketmar wrote:
 i remember that deprecation was rejected. maybe this is false 
 memory,
 though.

 btw, there are legit uses of comma, in c-style `for`, for 
 example. this
 should be left intact, i think (oh, can c-style `for` be 
 deprecated too?!
 ).

I think the last time there was a big discussion about this 
everyone agreed that the comma operator in a for loop was 
acceptable.

Jonathan M Davis via Digitalmars-d-learn writes:

On Monday, March 09, 2015 22:29:23 Meta via Digitalmars-d-learn wrote:
 On Monday, 9 March 2015 at 22:00:46 UTC, ketmar wrote:
 i remember that deprecation was rejected. maybe this is false
 memory,
 though.

 btw, there are legit uses of comma, in c-style `for`, for
 example. this
 should be left intact, i think (oh, can c-style `for` be
 deprecated too?!
 ).

 I think the last time there was a big discussion about this
 everyone agreed that the comma operator in a for loop was
 acceptable.

Yeah. Even Java has that. But IIRC, it was decided that the comma operator
in general would be deprecated. I could be wrong though, and it's the sort
of thing that could easily be decided and then not actually happen for ages
(e.g. it was decided years ago that delete would be deprecated, but it still
isn't).

- Jonathan M Davis

ketmar <ketmar ketmar.no-ip.org> writes:

On Tue, 10 Mar 2015 01:31:39 -0700, Jonathan M Davis via
Digitalmars-d-learn wrote:

 it's the sort of thing that could easily be decided and then not
 actually happen for ages (e.g. it was decided years ago that delete
 would be deprecated, but it still isn't).

sometimes i'm irritated by this, but with `delete`... ah, i hope nobody=20
will remember about for for next 20 years.=