Hi

This is the monthly status for the unofficial d wish list: http://all-technology.com/eigenpolls/dwishlist/

Right now the wish list looks like this:

114  array initialization/literals
91  Reflection API
75  vectorization
73  Stack tracing
72  Faster GC
51  Short syntax for new
48  Improved Foreach
47  Multiple opCast per class
45  unit test after compilation
44  Multiple return values (tuples
44  readonly/const enforcement
42  extra compiler values
39  Native AMD64 codegen
36  Unit test isolation
36  Explicit out/inout
35  Stack allocated classes.
33  Posix threads support native
33  Array masking
33  Debug check for null reference
32  Auto new-ing of classes
31  !in
28  Explicit type initializers
28  Foreach on first/on last
26  Return-type overloading
24  unit test & code separation
24  black box unit testing
24  associative arrays by index
24  Weak references/pointers
23  unit test coverage
23  coherent assoc. array syntax
23  Pass value params byref
22  Renaming ctor/dtor
22  better syntax for cast
22  Consistent struct/class sizeof
22  auto-member objects
20  Unit test measurements
19  Array pushback/popback
18  proper cast operators
18  Header File generation by DMD
17  User-defined sync function
16  Templates in classes
16  Explicit module `friendship`
15  Non-Static isExpression
15  Experimental d compiler
13  Conditional syncronized
13  Eigenpoll fix
13  opIn
10  imag and comp FP types.
9  OS X Build
8  Built-in variant type
8  Iterators and Generators
8  D library contest
8  inout variable and return
8  Precise names for floats
8  Pascal casing for methods, mod
7  Call log
7  Pascal like sets
6  Against class instance sizeof
6  modules must not rely on files
5  Add native string type
5  Meta Information
5  Explicit property keyword
4  Improve module architecture
4  L-Value return
3  Relational class/array algebra
3  conv() and opConv
3  inline expansion
3  if, while, true, false, int
2  interface to C++
2  Finite sets
2  Multistep return
2  struct literal/initialization
2  No Postfix Array Declarations
1  Manage .resources files
1  Declaration in function calls
1  Inline enum declaration
0  allow change self interface(?)
0  consistant new
0  Explicit 'property' keyword
0  Parallel Scavenging GC
0  opCast overloading

So we have array literals now.  I guess non-static array initialization is possible now that we have this?


On Sat, 30 Sep 2006 17:47:34 -0700, <12tkvvb02@sneakemail.com> wrote:

>
> Hi
>
> This is the monthly status for the unofficial d wish list:
> http://all-technology.com/eigenpolls/dwishlist/
>
> Right now the wish list looks like this:
>
> 114  array initialization/literals
> 91  Reflection API
> 75  vectorization
> 73  Stack tracing
> 72  Faster GC
> 51  Short syntax for new
> 48  Improved Foreach
> 47  Multiple opCast per class
> 45  unit test after compilation
> 44  Multiple return values (tuples
> 44  readonly/const enforcement
> 42  extra compiler values
> 39  Native AMD64 codegen
> 36  Unit test isolation
> 36  Explicit out/inout
> 35  Stack allocated classes.
> 33  Posix threads support native
> 33  Array masking
> 33  Debug check for null reference
> 32  Auto new-ing of classes
> 31  !in
> 28  Explicit type initializers
> 28  Foreach on first/on last
> 26  Return-type overloading
> 24  unit test & code separation
> 24  black box unit testing
> 24  associative arrays by index
> 24  Weak references/pointers
> 23  unit test coverage
> 23  coherent assoc. array syntax
> 23  Pass value params byref
> 22  Renaming ctor/dtor
> 22  better syntax for cast
> 22  Consistent struct/class sizeof
> 22  auto-member objects
> 20  Unit test measurements
> 19  Array pushback/popback
> 18  proper cast operators
> 18  Header File generation by DMD
> 17  User-defined sync function
> 16  Templates in classes
> 16  Explicit module `friendship`
> 15  Non-Static isExpression
> 15  Experimental d compiler
> 13  Conditional syncronized
> 13  Eigenpoll fix
> 13  opIn
> 10  imag and comp FP types.
> 9  OS X Build
> 8  Built-in variant type
> 8  Iterators and Generators
> 8  D library contest
> 8  inout variable and return
> 8  Precise names for floats
> 8  Pascal casing for methods, mod
> 7  Call log
> 7  Pascal like sets
> 6  Against class instance sizeof
> 6  modules must not rely on files
> 5  Add native string type
> 5  Meta Information
> 5  Explicit property keyword
> 4  Improve module architecture
> 4  L-Value return
> 3  Relational class/array algebra
> 3  conv() and opConv
> 3  inline expansion
> 3  if, while, true, false, int
> 2  interface to C++
> 2  Finite sets
> 2  Multistep return
> 2  struct literal/initialization
> 2  No Postfix Array Declarations
> 1  Manage .resources files
> 1  Declaration in function calls
> 1  Inline enum declaration
> 0  allow change self interface(?)
> 0  consistant new
> 0  Explicit 'property' keyword
> 0  Parallel Scavenging GC
> 0  opCast overloading



-- 
Using Opera's revolutionary e-mail client: http://www.opera.com/mail/

John Reimer wrote:
> So we have array literals now.  I guess non-static array initialization is possible now that we have this?

No, Walter seems to have forgotten to remove the check that produced the error message for non-static array initialization:

    D:\Temp> cat test.d
    void main()
    {
        int[] foo = [1, 2, 3];
    }
    D:\Temp> dmd test.d
    test.d(3): variable test.main.foo is not a static and cannot have static initializer

IIRC he has acknowledged this is a bug. So hopefully it'll be fixed in the next DMD version.

Workaround until then:

    void main()
    {
        int[] foo; foo = [1, 2, 3];
    }

Hi John Reimer, you wrote:
> So we have array literals now.  I guess non-static array initialization is possible now that we have this?

How about associative array literals?

-- 
serg.

Frits van Bommel wrote:
> John Reimer wrote:
>> So we have array literals now.  I guess non-static array initialization is possible now that we have this?
> 
> No, Walter seems to have forgotten to remove the check that produced the error message for non-static array initialization:
> 
>     D:\Temp> cat test.d
>     void main()
>     {
>         int[] foo = [1, 2, 3];
>     }
>     D:\Temp> dmd test.d
>     test.d(3): variable test.main.foo is not a static and cannot have static initializer
> 
> IIRC he has acknowledged this is a bug. So hopefully it'll be fixed in the next DMD version.

And just over 2 hours later, DMD 0.168 came out and the bug is fixed. Guess I should have waited to reply ;).

So now "array initialization/literals" have been implemented except for associative arrays (as Serg Kovrov mentioned).

Serg Kovrov wrote:
> Hi John Reimer, you wrote:
>> So we have array literals now.  I guess non-static array initialization is possible now that we have this?
> 
> How about associative array literals?

Nobody's ever happy <g>.

Hi Walter Bright, you wrote:
> Nobody's ever happy <g>.

No, I'm happy in general, although it may not seems so =)
Im not complain that language is bad, just wish for it to be even better (at least for me).
And again, big thank you Walter for your efforts. Really.
I believe most of people here mostly happy with D =)

Everyone have small issues. But small things matters, you know... For me AA literals would be useful, but its more aesthetic thing (although big one).

What I really NEED is GC to return memory to OS...

And a cool thing would be for GC to have an interface to provide some stats like current GC-allocated memory and GC-pool size, etc... Yeah =)

-- 
serg.

Serg Kovrov wrote:
> And a cool thing would be for GC to have an interface to provide some stats like current GC-allocated memory and GC-pool size, etc... Yeah =)

Sorry if you already know this, but there is std.gc.getStats

Hi Tom S, you wrote:
> Sorry if you already know this, but there is std.gc.getStats

No, I did not. Thank you Tom.

I can't find any docs for it. Seems I have to dig sources then...

-- 
serg.

On Tue, 03 Oct 2006 01:06:33 -0700, Frits van Bommel <fvbommel@REMwOVExCAPSs.nl> wrote:

> Frits van Bommel wrote:
>> John Reimer wrote:
>>> So we have array literals now.  I guess non-static array initialization is possible now that we have this?
>>  No, Walter seems to have forgotten to remove the check that produced the error message for non-static array initialization:
>>      D:\Temp> cat test.d
>>     void main()
>>     {
>>         int[] foo = [1, 2, 3];
>>     }
>>     D:\Temp> dmd test.d
>>     test.d(3): variable test.main.foo is not a static and cannot have static initializer
>>  IIRC he has acknowledged this is a bug. So hopefully it'll be fixed in the next DMD version.
>
> And just over 2 hours later, DMD 0.168 came out and the bug is fixed. Guess I should have waited to reply ;).
>
> So now "array initialization/literals" have been implemented except for associative arrays (as Serg Kovrov mentioned).


Yes, I noticed that too. I think it was just a coincidence. ;D

Thanks, Walter.  You've been doing a phenomenal job.

-JJR