Port get_index and the bitset functions
#107
Conversation
`Bitmap::position` is implmemented in terms of the new `roaring_bitmap_get_index` function. The name `position` was chosen because it mirrors the rust `Iterator::position` function, which does the same thing.
CRoaring introduced a `bitset` type, and the abilty to convert a roaring bitmap into a dense bitset. This adds wrappers around the new bitset functions
| #[inline] | ||
| #[doc(alias = "bitset_clear")] | ||
| #[doc(alias = "bitset_fill")] | ||
| pub fn fill(&mut self, value: bool) { |
There was a problem hiding this comment.
Not sure I like this. Any reason why bitset_clear and bitset_fill aren't separate?
There was a problem hiding this comment.
My thoughts were to be closer to <[bool]>::fill, and that since it's inline, .fill(constant) will be optimized to a single call without a branch anyway.
I really don't want to use the name clear (since it's too similar to Vec::clear but does something VERY different), so the alternative in my mind would be Bitset::fill_zeros and Bitset::fill_ones.
I'm not hard set on this option, but it seemed like the best option to me.
|
Really struggling to find proper time to review larger amounts of code lately, but generally it looks great to me! Thanks one more time for the contribution. On that note, @Dr-Emann, you made some great contributions over the last years to |
|
I'm currently advocating for the |
|
I'd certainly be willing to help maintain the project, and I also personally think it would make sense to end up under the RoaringBitmap org. |
roaring_bitmap_get_indexasBitmap::position(the name was chosen to match withIterator::position)bitset_*methods/types/variablesbitset_*functions, androaring_bitmap_to_bitsetasBitmap::to_bitsetBecause of the version update, this will clash with #105 and #106