Daniel Lemire
Daniel Lemire
@Oppen Sounds sensible. I have no strong opinion and I am eager to know what people think. I will setup an issue.
@qijiax The reason I am not merging this PR (which I wrote) is that we have no reason to believe that it solves the problem. That is we don't know...
> For this PR, I believe that is same as 32bit roaring bitmap, when allocator return null, we can execute free function and return null. All the calling stack do...
@qijiax > Any updates on this discussion? Please read the comments on this PR (see above comments). It does not seem to get the massive support I'd want. So I...
@IvanNardi One issue here is our capacity to test this internally. E.g., it is a fall fine to have checks, but how do we know that we are doing the...
> I am not good at c++ Your code looks fine to me. See my comments on specific details.
@pauldreik There might now be a slight conflict.
@WojciechMula > How about adding a template, that would be enabled if the argument has methods size() and data()? ```c++ template concept spannable = requires(T v) { { v.size() }...
I think that this is absolutely fantastic work. It will be fun to document it. I sort of imagined something of the sort, but what you are doing is so...
@pauldreik I am marking this PR as 'approved' because my objection can be handled as a separate step.