Open CaseyCarter opened 8 years ago
Why not just merge from upstream? It'll be painful and time-consuming, but merges after that will be much easier.
Why not just merge from upstream?
That's what I mean by "bringing this library current with upstream." I'll merge everything from upstream, and then upstream less-invasive workarounds. Eventually upstream will be able to support MSVC directly (I fervently hope this will happen sometime in the VS "15" release cycle) and this fork will be a small delta of the more obnoxious workarounds needed to support the VS2015-era compiler.
Which parts of upstream range-v3 are missing in this fork? Everything added since the fork from ericniebler/range-v3@6eb5c831ffe12cd5cb96390dbe917ca1b248772d. Roughly, that means:
sample
algorithm and view are absent.range
class template hasn't been renamed toiterator_range
.IteratorRange<I, S>
is nowSentinel<S, I>
upstream).adjacent_filter
view is absent.drop_exactly
view is absent.view::set_difference
,view::set_intersection
,view::set_symmetric_difference
, andview::set_union
) are absent.There are also many small differences due to bug fixes and evolution of interfaces that has occurred over the last year in upstream range-v3. I don't have an exhaustive list of those differences, and frankly I'd rather invest time into bringing this library current with upstream range-v3 than extensively investigating and documenting the differences.
Please, if you see any glaring differences while using the library that you think should be brought to the attention of others, post them in this thread and I'll add them to this post.