[nhc-bugs] Building nhc98 on Windows 2000

[Simon Peyton-Jones]

| > > /tmp/... directory *does* exist, but the \tmp\... directory does=20
| > > not. (And indeed, why should it, since the invocation=20
| didn't ask for=20
| > > it?)
| >=20
| > Notice that both GHC and the mingw system calls (the MS provided C=20
| > runtime functions, really) are capable of understanding paths that=20
| > contain either "/" or "\" as directory separators, so=20
| you're reading=20
| > too much into the slashes being reversed in that output message.
|=20
| So am I right in thinking that if GHC did no internal=20
| filepath mangling whatsoever, everything file-related would just work?

GHC does mangle '/' into '\' on Windows.  That is jolly convenient, and
we certainly rely on it when using GHC (e.g to build GHC).
I don't know why doing so should be a problem; after all, it's Windows.

Simon