I see. gcc 4.9 doesn't support all of the C++14 standard, so if we definitely need to keep supporting gcc 4.9 for another 6 months, then we may have to wait. > 5 years is a long time to be frozen on gcc 4.x (4.8 was the minimum version when we started the project).
On Mon, Sep 28, 2020 at 4:10 PM Neal Richardson <neal.p.richard...@gmail.com> wrote: > > R is looking much more favorable for C++14 or even C++17 since the 4.0 > release in April: > https://cran.r-project.org/doc/manuals/r-release/R-exts.html#Using-C_002b_002b14-and-later > > It's unclear how well the old Rtools Windows toolchain would support > C++14--we can try and see, at least it's gcc 4.9. But as far as CRAN is > concerned, I believe we can stop supporting that in April when R 4.1 is > released (they test on current and previous x.y versions). > > Neal > > On Mon, Sep 28, 2020 at 1:06 PM Wes McKinney <wesmck...@gmail.com> wrote: > > > hi folks, > > > > I think this has come up before, but I'm curious what would be the > > roadblocks (if any) for upgrading our minimum C++ standard from 11 to > > 14. One of the last hangers-on is probably our Python manylinux1 build > > which uses gcc 4.8, but is there anything else? The minimum supported > > MSVC for C++14 is VS 2017 -- would that cause any issues? > > > > Thanks, > > Wes > >
