Artem Pavlenko

Results 235 comments of Artem Pavlenko

``` c++ #include #include "variant.hpp" #include struct big { double val[10]; // 8 * 10 = 80 }; i int main() { { std::cerr

planning to take a look at using `uint8_t` for `type_index` /cc @springmeyer @joto @lightmare

`std::size_t` for storing internal index was too generous and also was resulting in larger `sizeof` in some use cases comparing to `boost::variant` and `std::variant`. After 9eec1fd ```c++ #include #include #include...

> One other thing to keep in mind: The implementation of recursive_wrapper was basically taken wholesale from boost, just ported to C++11. Keeping compatibility with boost is one of our...

Just reading https://github.com/mapbox/variant/issues/50 - perhaps this should go into 1.1.0 it we have a consensus ? /cc @joto @tomilov @lightmare

@joto - I referred to implementing recursive_wrapper as an alias to or based on `std::unique_ptr` in 1.1.0 (https://github.com/mapbox/variant/milestones/1.1.0).

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=80593

@SergeyIvanov87 - It looks like you're correct but I vaguely recall having some reasons for not having a check. I'll need to dig through history to see why implementation ended...

@SergeyIvanov87 @mgambrell @springmeyer - I don't remember why there's no check for the self assignment. I can only think that at some point operaror= signature was : ```c++ VARIANT_INLINE variant&...