On 21 February 2017 at 10:48, John Colvin via D.gnu <d.gnu@puremagic.com> wrote: > On Monday, 20 February 2017 at 17:00:48 UTC, Iain Buclaw wrote: >> >> The next gcc release ought to be an interesting experiment for the >> community, as although the version is still set to 2068, gdc has now >> diverged so far from the last C++ sources that it is anything but. I'll >> probably bump it to whatever version of phobos it can compile. :-) > > > Is there a possibility that gdc will continue to diverge? I don't > necessarily think that's a bad thing, if upstream features can be > implemented in a timely manner (that's a big "if" though).
Initially, the plan is to be aligned with stable branch in terms of sharing common C++ headers, then aligning any changes that alter the code generator. The idea being that this will allow transition between C++ and D implementations to be seamless (no backend changes required). We'll keep it in this state until a time when we are ready to switch. My plan is that when the switch happens, we'll just continue to be in sync with stable, rather than a given release.
