Next in Review Queue: The New std.path

[Lars T. Kyllingstad]

On Mon, 18 Jul 2011 10:04:26 -0400, Steven Schveighoffer wrote:

> On Mon, 18 Jul 2011 09:16:35 -0400, Lars T. Kyllingstad
> <public at kyllingen.nospamnet> wrote:
> 
>> On Sun, 17 Jul 2011 12:30:39 +0200, torhu wrote:
>>
>>> On 15.07.2011 02:20, dsimcha wrote:
>>>> Lars Kyllingstad's new and improved std.path module for Phobos is the
>>>> next item up in the review queue.  I've volunteered to be the review
>>>> manager.  Barring any objections, the review starts now and ends at
>>>> the ends two weeks from now on July 28.  This will be followed by a
>>>> week of voting, ending August 4th.
>>>>
>>>> Code:
>>>> https://github.com/kyllingstad/phobos/blob/std-path/std/path.d
>>>>
>>>> Docs:
>>>> http://www.kyllingen.net/code/new-std-path/phobos-prerelease/
>> std_path.html
>>>>
>>>>
>>> Looks nice and clean, both docs and code!  I like modules that are not
>>> overengineered :)
>>
>> Thanks!
>>
>>
>>> I noticed a couple of things that I don't think have been mentioned
>>> already:
>>>
>>> The docs for defaultExtension say that there's one case where doesn't
>>> allocate.  But the code says that it always does (.idup isn't
>>> conditional, it's just .dup with a different return type).
>>
>> There is no reason for idup to duplicate an array which is already
>> immutable.  If it does, I'd say it is a compiler bug.
> 
> No, it has to duplicate.  Part of idup is saying "I want an immutable
> return" and the other part is saying "I want a duplicate"
> 
> For example:
> 
> string x = "hello".idup;
> string y = x;
> 
> assert(x.capacity > 0); // would fail if idup just returned the same
> thing x ~= " world";
> 
> assert(y is x); // would fail.
> 
> The compiler cannot know whether you care or not.

Ah, ok.


> If you don't care, why are you calling idup on an immutable array?  Just
> use it.

The specific cases in question are templated functions, where I don't 
know whether the array is immutable unless I specifically test it.  I 
simply used arr.idup, assuming the compiler would optimise away the 
duplication when arr is immutable.  But it's no big deal to change it to

   static if (!is(T == immutable)) return arr.idup;

-Lars