Closed mpark closed 6 years ago
Hi Zach! I'm planning to work on this next week. Last time I tried this it was having trouble with the cartesian product generation here: https://github.com/mpark/variant/blob/master/include/mpark/variant.hpp#L574-L575
So I'll have to figure out how to deal with that part.
Thanks for the recommendation!
Closed by 12b77f3325b81ba56a3c3f0b88e61b57505b9ff0.
The minimum repro of the problems were https://wandbox.org/permlink/ypmovjwyTaXnSnwy.
In short, S<alias<pack..., thing>>
drops thing
and incorrectly passes alias<pack...>
instead.
This caused the true
in all
implementation to be dropped in bool_sequence<Bs..., true>
,
and the J
was dropped in index_sequence<Is..., J>
. Since the issue was due to the fact that
bool_sequence
and index_sequence
are aliases, integer_sequence
is now used directly.
Thanks for the quick resolution!
ETA? Would it help to prod you on this if I told you I just recommended this implementation to some at-risk youths who are forced to use GCC 4.8.3?