Niall Douglas
Niall Douglas
TBH I'm surprised LLFIO compiles with mingw. I certainly never test it. You'll probably need to help me here, is it the case that mingw's headers simply don't define `inaddr6_loopback`...
Yes to both please! Just to clarify: which mingw do you refer to? I remember the 64 bit one had higher quality than the original one, but that was many...
Cool, rock on with that at your convenience.
This _used_ to work on ReFS, but it hasn't been tested in a long, long time. And Windows may have changed what it publishes. I can certainly see here it...
The test you identified is for testing that function. Contribution of better tests would be welcome.
I can tell you that function copies Terabytes of data every day in 99.999% uptime cloud systems. It doesn't mean you wouldn't find a bug when you use it, but...
To my best knowledge the hole punching extents cloning implementation is correct. Implementation of something correct in all corner cases whilst retaining performance and is semantically equivalent across all platforms...
You are completely right that for `io_handle::read()`, one can no longer detect EOF because we now tolerate partial buffer fills more than before R2. This is because an `io_handle` implementation...
> This is the part I wanted to be clarified and I think it is the only sane way to specify it. Which brings me to my actual point: The...
I did some more research into this, and what happens if you read at or beyond EOF varies between systems: - Windows: error >= ERROR_HANDLE_EOF - Linux & BSD: EINVAL...