It complicates the grammar and doesn't fit with D's style of declaraing variables. (type then identifier) I realize that it was left in to make C programmers happy, but it doesn't even work like it does in C:

int a[10];
int b[10], b[10];
int[10] c, d;
int e[string];
int f[string], g[string];

test.d(2): Error: multiple declarations must have the same type, not int[10] and int[10]
test.d(5): Error: multiple declarations must have the same type, not int[string] and int[string]

On 2013-05-01 00:36, Brian Schott wrote:
> It complicates the grammar and doesn't fit with D's style of declaraing
> variables. (type then identifier) I realize that it was left in to make
> C programmers happy, but it doesn't even work like it does in C:
>
> int a[10];
> int b[10], b[10];
> int[10] c, d;
> int e[string];
> int f[string], g[string];
>
> test.d(2): Error: multiple declarations must have the same type, not
> int[10] and int[10]
> test.d(5): Error: multiple declarations must have the same type, not
> int[string] and int[string]

I would say no if it was removed.

-- 
/Jacob Carlborg

Brian Schott:

> It complicates the grammar and doesn't fit with D's style of declaraing variables. (type then identifier)

I suggested something more moderate: to just disallow mixing C and D syntax in the same declaration:
http://d.puremagic.com/issues/show_bug.cgi?id=5807


> I realize that it was left in to make C programmers happy,

For that I suggested a -cstyle switch, to be used only temporarily to port C code to D:
http://d.puremagic.com/issues/show_bug.cgi?id=4580

Bye,
bearophile

On Tuesday, 30 April 2013 at 22:36:44 UTC, Brian Schott wrote:
> I realize that it was left in to make C programmers happy

I remember being mentioned by someone that it is more about simplifying porting of C headers to D than actually making C programmers happy.

On Tuesday, 30 April 2013 at 22:36:44 UTC, Brian Schott wrote:
> I realize that it was left in to make C programmers happy

I remember being mentioned by someone that it is more about
simplifying porting of C headers to D than actually making C
programmers happy.

On Thursday, 2 May 2013 at 08:18:48 UTC, Dicebot wrote:
> On Tuesday, 30 April 2013 at 22:36:44 UTC, Brian Schott wrote:
>> I realize that it was left in to make C programmers happy
>
> I remember being mentioned by someone that it is more about simplifying porting of C headers to D than actually making C programmers happy.

Array are value type in D so I'm not sure this really helps.

On Wednesday, May 01, 2013 00:36:43 Brian Schott wrote:
> It complicates the grammar and doesn't fit with D's style of declaraing variables. (type then identifier) I realize that it was left in to make C programmers happy, but it doesn't even work like it does in C:
> 
> int a[10];
> int b[10], b[10];
> int[10] c, d;
> int e[string];
> int f[string], g[string];
> 
> test.d(2): Error: multiple declarations must have the same type,
> not int[10] and int[10]
> test.d(5): Error: multiple declarations must have the same type,
> not int[string] and int[string]

I think that deprecating the C-style syntax has been considered in the past, but I don't know what Andrei or Walter's stance on it is. I don't think that it adds much value though beyond making it slightly easier to port code from C to D, and I'm all for removing it. Most D programmers are probably using the D syntax anyway.

- Jonathan M Davis