Ben FrantzDale

Results 34 comments of Ben FrantzDale

I landed here looking for the reverse (Related: https://stackoverflow.com/questions/72889401/can-boost-pfr-be-used-to-iterate-the-fields-of-a-type-as-member-pointers) I want ``` template using ith_member_type = decltype(boost::pfr::get(std::declval())); template auto get_member_pointer() -> ith_member_type T::* { ??? } ```

I’m not sure I follow. Are you saying I would specialize `detail::spatial_predicate< LineBoundedSpace, ...>`? That feels wrong since I’d assume anything in `detail` is an implementation detail not a customization...

OK, with that I get a compile error: https://godbolt.org/z/3qxdPM I'll take a stab at implementing the multi-level `satisfies` equivalent.

LMK if this is interesting and the degree of granularity we'd want and I'll consider doing it. I can see a few different approaches: 1. Just one big `openvdbFwd.h` that...

That all makes sense. I haven’t tried it. The “wink out” optimization feels potentially interesting but again, I haven’t tried it.

In particular, I'd be curious if people have profiled the destruction of enormous trees. I see that `~Tree()` calls `Tree::clear()` which calls `stealNodes` in serial then ``` tbb::parallel_for(tbb::blocked_range(0, leafnodes.size()), DeallocateNodes(leafnodes));...

I'm seeing the same thing. It goes on to complain about similar things: ``` [...]abseil/absl/container/internal/btree.h:2292:37: error: comparison of integers of different signs: 'int' and 'absl::container_internal::btree::(anonymous enum at [...]abseil/absl/container/internal/btree.h:1056:3)' [-Werror,-Wsign-compare] left->count()...

I think changing the c'tor to this addresses the issue: ``` template inline constexpr unit_t(const std::chrono::duration& value) noexcept : nls(units::convert(static_cast(std::chrono::duration_cast(value).count()))) { } ``` although that may open a hole for...

I came here to open a ticket like this. I'd be happy to propose some additional wording. Here's what I'm looking for: 1. A dead-simple example: A type that only...

@vitaut if you give me a bit of feedback on my proposed wording, I'm happy to PR it. Or if you'd rather I could PR a change in wording and...