Comment # 5 on bug 116 from
(In reply to Bruno Medeiros from comment #4)
> (In reply to Iain Buclaw from comment #3)
> > Those patches aren't essential. The recent mingw builds are evident of that.
> > However I am aware of some essential runtime fixes in those patches (some
> > TLS related) but almost all are unsuitable for inclusion.
> Ian, I didn't understand your point, it seems self contradictory. You say
> the patches aren't essential, and yet at the same time that there are "some
> essential runtime fixes in those patches" ?..

First point was in response to patches needed to apply to build it (compiler
patches, build patches).  These are not required.

The second point was runtime support, which needs some love.

For instance, I have implemented EmuTLS in the runtime backend, but it's not
yet working with the GC (attempts to hook the data regions into the D GC have
so far caused deadlocks).  If this were to be fixed, you can build GDC on MinGW
with --disable-tls

You are receiving this mail because: