dsimcha wrote:
> == Quote from Yigal Chripun (yigal100@gmail.com)'s article
>> dsimcha wrote:
>>> == Quote from Max Samukha (spambox@d-coding.com)'s article
>>>> On Sat, 21 Nov 2009 18:51:40 +0000 (UTC), dsimcha <dsimcha@yahoo.com>
>>>> wrote:
>>>>> == Quote from Max Samukha (spambox@d-coding.com)'s article
>>>>>> On Fri, 20 Nov 2009 15:30:48 -0800, Walter Bright
>>>>>> <newshound1@digitalmars.com> wrote:
>>>>>>> Yigal Chripun wrote:
>>>>>>>> what about foreach_reverse ?
>>>>>>> No love for foreach_reverse? <tear>
>>>>>> And no mercy for opApply
>>>>> opApply **must** be kept!!!!  It's how my parallel foreach loop works.  This
> would
>>>>> be **impossible** to implement with ranges.  If opApply is removed now, I will
>>>>> fork the language over it.
>>>> I guess it is possible:
>>>> uint[] numbers = new uint[1_000];
>>>> pool.parallel_each!((size_t i){
>>>>         numbers[i] = i;
>>>>     })(iota(0, numbers.length));
>>>> Though I agree it's not as cute but it is faster since the delegate is
>>>> called directly. Or did I miss something?
>>> I'm sorry, but I put a lot of work into getting parallel foreach working, and I
>>> also have a few other pieces of code that depend on opApply and could not (easily)
>>> be rewritten in terms of ranges.  I feel very strongly that opApply and ranges
>>> accomplish different enough goals that they should both be kept.
>>>
>>> opApply is good when you **just** want to define foreach syntax and nothing else,
>>> with maximum flexibility as to how the foreach syntax is implemented.  Ranges are
>>> good when you want to solve a superset of this problem and define iteration over
>>> your object more generally, giving up some flexibility as to how this iteration
>>> will be implemented.
>>>
>>> Furthermore, ranges don't allow for overloading based on the iteration type.  For
>>> example, you can't do this with ranges:
>>>
>>> foreach(char[] line; file) {}  // Recycles buffer.
>>> foreach(string line; file) {}  // Doesn't recycle buffer.
>>>
>>> They also don't allow iterating over more than one variable, like:
>>> foreach(var1, var2, var3; myObject) {}
>>>
>>> Contrary to popular belief, opApply doesn't even have to be slow.  Ranges can be
>>> as slow as or slower than opApply if at least one of the three functions (front,
>>> popFront, empty) is not inlined.   This actually happens in practice.  For
>>> example, based on reading disassemblies and the code to inline.c, neither front()
>>> nor popFront() in std.range.Take is ever inlined.  If the range functions are
>>> virtual, none of them will be inlined.
>>>
>>> Just as importantly, I've confirmed by reading the disassembly that LDC is capable
>>> of inlining the loop body of opApply at optimization levels >= O3.  If D becomes
>>> mainstream, D2 will eventually also be implemented on a compiler that's smart
>>> enough to do stuff like this.  To remove opApply for performance reasons would be
>>> to let the capabilities of DMD's current optimizer influence long-term design
>>> decisions.
>>>
>>> If anyone sees any harm in keeping opApply other than a slightly larger language
>>> spec, please let me know.  Despite its having been superseded by ranges for a
>>> subset of use cases (and this subset, I will acknowledge, is better handled by
>>> ranges), I actually think the flexibility it gives in terms of how foreach can be
>>> implemented makes it one of D's best features.
>> There are three types of iteration: internal to the container, external
>> by index, pointer, range, etc, and a third design with co-routines
>> (fibers) in which the container internally iterates itself and yields a
>> single item on each call.
>> Ranges accomplish only the external type of iteration. opApply allows
>> for internal iteration. All three strategies have their uses and should
>> be allowed in D.
> 
> Exactly.  I've said this before, but I think you said it much better.  Now that
> Walter has agreed to keep opApply, this should really be explained somewhere in
> TDPL and in the online docs to clarify to newcomers why someone would choose
> opApply over ranges or vice-versa.  External iteration is more flexible for the
> user of the object, internal iteration is more flexible for the designer of the
> object.

Copied this message to my todo list, thanks.

Andrei

On Sat, 21 Nov 2009 20:21:54 +0000 (UTC), dsimcha <dsimcha@yahoo.com>
wrote:

>> I guess it is possible:
>> uint[] numbers = new uint[1_000];
>> pool.parallel_each!((size_t i){
>>         numbers[i] = i;
>>     })(iota(0, numbers.length));
>> Though I agree it's not as cute but it is faster since the delegate is
>> called directly. Or did I miss something?
>
>I'm sorry, but I put a lot of work into getting parallel foreach working, and I also have a few other pieces of code that depend on opApply and could not (easily) be rewritten in terms of ranges.  I feel very strongly that opApply and ranges accomplish different enough goals that they should both be kept.

Nothing to be sorry about. I was replying to your statement that parallel foreach cannot be implemented in terms of ranges.

>
>opApply is good when you **just** want to define foreach syntax and nothing else, with maximum flexibility as to how the foreach syntax is implemented.  Ranges are good when you want to solve a superset of this problem and define iteration over your object more generally, giving up some flexibility as to how this iteration will be implemented.
>
>Furthermore, ranges don't allow for overloading based on the iteration type.  For example, you can't do this with ranges:
>
>foreach(char[] line; file) {}  // Recycles buffer.
>foreach(string line; file) {}  // Doesn't recycle buffer.

Ok

>
>They also don't allow iterating over more than one variable, like: foreach(var1, var2, var3; myObject) {}

I guess it is a relatively rare case. And you can always provide a range returning a tuple.

foreach(r; myObject.range) { writeln(r.front.var1, r.front.var2,
r.front.var3); }

The common case of iterating over index and value of a random access range should be supported by foreach.

>
>Contrary to popular belief, opApply doesn't even have to be slow.  Ranges can be as slow as or slower than opApply if at least one of the three functions (front, popFront, empty) is not inlined.   This actually happens in practice.  For example, based on reading disassemblies and the code to inline.c, neither front() nor popFront() in std.range.Take is ever inlined.

I think that is a problem with the compiler, not ranges.

> If the range functions are
>virtual, none of them will be inlined.

Ok

>
>Just as importantly, I've confirmed by reading the disassembly that LDC is capable of inlining the loop body of opApply at optimization levels >= O3.  If D becomes mainstream, D2 will eventually also be implemented on a compiler that's smart enough to do stuff like this.  To remove opApply for performance reasons would be to let the capabilities of DMD's current optimizer influence long-term design decisions.
>

Convinced. Let's keep opApply.

Hello Walter,

> Bill Baxter wrote:
> 
>> On Fri, Nov 20, 2009 at 4:05 PM, Walter Bright
>> <newshound1@digitalmars.com> wrote:
>>> Bill Baxter wrote:
>>> 
>>>> Right, but if you do define it (in order to do something extra upon
>>>> initialization -- validate inputs or what have you) then it no
>>>> longer works at compile time.
>>>> 
>>> Right, but the static initialization then shouldn't work, either.
>>> 
>> Why not?  It works if you use static opCall(int,int) instead of
>> this(int,int).
>> 
> Because if you need runtime execution to initialize, a back door to
> statically initialize it looks like a bug.
> 

who said anything about needing runtime? If the constructor can be evaluated as CTFE with it's only result being the struct getting set up, why not let it. I can think of serval useful thing to do with that (validation, denormalization, etc.)

BCS wrote:
> Hello Walter,
> 
>> Bill Baxter wrote:
>>
>>> On Fri, Nov 20, 2009 at 4:05 PM, Walter Bright
>>> <newshound1@digitalmars.com> wrote:
>>>> Bill Baxter wrote:
>>>>
>>>>> Right, but if you do define it (in order to do something extra upon
>>>>> initialization -- validate inputs or what have you) then it no
>>>>> longer works at compile time.
>>>>>
>>>> Right, but the static initialization then shouldn't work, either.
>>>>
>>> Why not?  It works if you use static opCall(int,int) instead of
>>> this(int,int).
>>>
>> Because if you need runtime execution to initialize, a back door to
>> statically initialize it looks like a bug.
>>
> 
> who said anything about needing runtime? If the constructor can be evaluated as CTFE with it's only result being the struct getting set up, why not let it. I can think of serval useful thing to do with that (validation, denormalization, etc.)
> 
> 
Read my post. It's just a compiler bug.

On Fri, 20 Nov 2009 00:48:28 -0500, Don <nospam@nospam.com> wrote:

> Now that we have struct literals, the old C-style struct initializers don't seem to be necessary.
> The variations with named initializers are not really implemented -- the example in the spec doesn't work, and most uses of them cause compiler segfaults or wrong code generation. EG...
>
> struct Move{
>     int D;
> }
> enum Move genMove = { D:4 };
> immutable Move b = genMove;
>
> It's not difficult to fix these compiler problems, but I'm just not sure if it's worth implementing. Maybe they should just be dropped? (The { field: value } style anyway).

Brought up in another thread, a good use of static initializers for structs: arrays of POD literals.

For example:

struct RGB
{
  ubyte red, green, blue;
}

RGB[256] PALETTE = [
{0x00, 0x00, 0x00},
{0x01, 0x01, 0x01},
...
];

can you do something like this without static initializers?  My recollection is that this is the only way to have a struct array literal.

-Steve

Steven Schveighoffer:
> can you do something like this without static initializers?  My recollection is that this is the only way to have a struct array literal.

http://codepad.org/8HnF62s2

Bye,
bearophile

On Tue, 29 Dec 2009 15:02:50 -0500, bearophile <bearophileHUGS@lycos.com> wrote:

> Steven Schveighoffer:
>> can you do something like this without static initializers?  My
>> recollection is that this is the only way to have a struct array literal.
>
> http://codepad.org/8HnF62s2

OK, that makes sense.  Last time I remember having issues with CTFE unless the static initializer was used, but that was on D1.

-Steve