On Tuesday, 17 July 2012 at 17:19:31 UTC, Jonathan M Davis wrote:
On Tuesday, July 17, 2012 10:47:50 Andrei Alexandrescu wrote:
On 7/17/12 4:41 AM, monarch_dodra wrote:
> On Monday, 16 July 2012 at 22:42:47 UTC, Andrei Alexandrescu
>
> wrote:
>> Wow, this is awesome. Did you discover that by inspection
>> or by
>> testing? I think a "malicious input range" would be a great
>> tool for
>> assessing which algorithms fail on input ranges.
>>
>> Andrei
>
> The first I discovered testing with a "ConsumableRange",
> actually. The second, I found by inspection.
>
> I'll correct those two issues myself, but I don't feel
> comfortable with the other issues.
You may want to submit them as bug requests. Thanks!
Yes. Please do. It's on my todo list to improve std.algorithm
and std.range's
tests (particularly for reference type ranges), and I've gotten
started on it,
but it could take a while to get it all done, and anything that
you find will
be valuable in not only figuring out what needs fixing but also
in figuring out
what needs better testing.
bugzilla: http://d.puremagic.com/issues
- Jonathan M Davis
Hi Jonathan,
I've made changes to algorithm to the best of my abilities. If it
does not meet requirements, please tell me what is wrong, and all
work on it as I can. I've put an in-depth explanation of the
changes in the pull request description.
Slightly on topic, did you read my post about "Definition of
"OutputRange" insuficient""? Would it be OK to add "hasLength" to
range.d? This would be the first step to making outputRanges more
useable, without directly changing the definition of an output
range quite yet.
