https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108487
Richard Biener changed:
What|Removed |Added
Ever confirmed|0 |1
Status|UNCONFIRMED
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108487
Richard Biener changed:
What|Removed |Added
Target Milestone|--- |10.5
Keywords|
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108487
--- Comment #9 from Jonathan Wakely ---
I think we just want to dispatch on iterator concept not iterator category.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108487
--- Comment #8 from Mark Bourgeault ---
What about something like this?
#if __cplusplus >= 201709L
template>
vector(_InputIterator __first, _InputIterator __last,
const allocator_type& __a = allocator_type())
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108487
--- Comment #7 from MARK BOURGEAULT ---
>> See PR 100070 for suggestions to deal with such iterators better.
Unless I'm missing something, there's nothing in that PR that a *user* can do
to achieve the gcc 10.3 performance w/ std::iota_view.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108487
--- Comment #6 from Jonathan Wakely ---
(In reply to Jonathan Wakely from comment #4)
> That's because iota_view's iterator is only an input iterator in gcc-10,
I meant to say gcc-10.4 there.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108487
Jonathan Wakely changed:
What|Removed |Added
See Also||https://gcc.gnu.org/bugzill
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108487
--- Comment #4 from Jonathan Wakely ---
That's because iota_view's iterator is only an input iterator in gcc-10, as a
result of r10-9796-g1cb39945993c89
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108487
Alexander Monakov changed:
What|Removed |Added
Component|tree-optimization |libstdc++
--- Comment #3 from Alexa