On Sunday, 23 March 2014 at 00:50:34 UTC, Walter Bright wrote:
It's become clear to me that we've underspecified what an
InputRange is.
Some thoughts:
- A while ago I realized I did not know how some of those
underspecified details should be implemented (I think my initial
doubt was if the range was expected/required to enforce(!empty)).
Assuming it was just my ignorance (and incomplete or hard to find
documentation), I asked in the IRC channel and, IIRC, the answer
I got there was that ranges are an abstract concept, so such
minutia were supposed to be implementation details. From this
forum thread and that IRC feedback, it seems that, indeed, people
have been operating under an illusion of a shared universal
assumption. I'm very happy to see this addressed.
- (Lots of text deleted because I was no longer sure about it ...)
- Is it allowed for an InputRange to become empty and then not
empty again? For a finite RandomAccessRange to increase in
length? E.g.:
// Example 1: (with "outside" control flow)
if(someQueue.empty)
addStuffToQueue();
assert(!someQueue.empty);
// Example 2: (with only range-related control flow)
if(!someQueue.empty)
{
auto n = someQueue.length;
someQueue.popFront();
assert(someQueue.length == n-1); // can this fail?
}
- Is it allowed to not call front? E.g.:
void dropAll(Range)(Range range)
{
while(!empty)
range.popFront();
}
