On Jul 27, 10 13:20, Andrei Alexandrescu wrote:
Philippe Sigaud wrote:
Andrei Alexandrescu wrote:
[snip]
Yah... that is all cool. My only reservations are that with what we have
right now this looks like a lot of manual special casing. By the way,
chain() with ranges sorted with the same predicate is also sorted...
No
chain([4,5,6,7], [1,2]) // <-- not sorted
unless you can also assert input[i-1].back <= input[i].front.
Then, such an effort would be much better motivated if sorted ranges had
some really interesting properties. Aside from those discussed - there's
not a lot of them!
[snip]
Andrei