www.digitalmars.com         C & C++   DMDScript  

digitalmars.D - Chainable template mixin and opCat()

reply "Per =?UTF-8?B?Tm9yZGzDtnci?= <per.nordlow gmail.com> writes:
I just stumbled upon the following very interested idea:

http://forum.dlang.org/thread/bug-6043-3 http.d.puremagic.com%2Fissues%2F

1. Add the following (where?)

template Chainable() {
        Chain!(typeof(this), Range) opCat(Range)(Range r) {
            return chain(this, r);
        }
}

2. Then in the body of a lazy range both Phobos ones like map, 
filter, zip, iota, and so on, and user-defined ones, you may add:

mixin Chainable;

This allows to write more natural and readable code like:

     iota(5) ~ filter!q{a}([1,0,3])
     map!abs([1,2,-5]) ~ [10, 20]

instead of more noisy:

     chain(iota(5), filter!q{a}([1,0,3]))
     chain(map!abs([1,2,-5]), [10, 20])

See also bug 5638 for a useful optimization.

I'd be glad to work on Phobos PR for this :)
Jun 01 2015
parent reply "Per =?UTF-8?B?Tm9yZGzDtnci?= <per.nordlow gmail.com> writes:
On Monday, 1 June 2015 at 11:20:52 UTC, Per Nordlöw wrote:

 template Chainable() {
        Chain!(typeof(this), Range) opCat(Range)(Range r) {
            return chain(this, r);
        }
 }


Would it suffice to restrict `opCat` as

     Chain!(typeof(this), Range) opCat(Range)(Range r)
     if (isInputRange!Range && is(CommonElementType!(typeof(this), 
Range)))

?

given that

template CommonElementType(Rs...)
{
     alias CommonElementType = CommonType!(staticMap!(ElementType, 
Rs));
}

In general, should we let `Chain` do the error checking or should 
we copy its retrictions into `opCat`s restrictions?
Jun 01 2015
parent reply "w0rp" <devw0rp gmail.com> writes:
I think my only complaint for this is that it wouldn't work for 
all ranges, so you'd be able to use the operator some times, but 
not others. This is because you have to define the operators 
inside of the structs or classes, and you can't write operators 
as free functions. (Which is a good feature in general.)
Jun 01 2015
next sibling parent reply "Per =?UTF-8?B?Tm9yZGzDtnci?= <per.nordlow gmail.com> writes:
On Monday, 1 June 2015 at 11:47:43 UTC, w0rp wrote:

 I think my only complaint for this is that it wouldn't work for 
 all ranges, so you'd be able to use the operator some times,


I still think the simplicity of

     mixin Chainable;

and it's effect on usability is short enough to be worth the 
effort :)
Jun 01 2015
parent "Per =?UTF-8?B?Tm9yZGzDtnci?= <per.nordlow gmail.com> writes:
On Monday, 1 June 2015 at 12:13:51 UTC, Per Nordlöw wrote:

 I still think the simplicity of

     mixin Chainable;

 and it's effect on usability is short enough to be worth the 
 effort :)


https://github.com/D-Programming-Language/phobos/pull/3353

Destroy!
Jun 01 2015
prev sibling parent Timon Gehr <timon.gehr gmx.ch> writes:
On 06/01/2015 01:47 PM, w0rp wrote:

 you can't write operators as free functions. (Which is a good feature in
 general.)


No, it is a pointless non-uniformity in function call syntax.
Jun 01 2015