Howard Chu dixit:
>Thanks to the wonderful geniuses at Microsoft, not all the world
>uses LP64. IL32 is a thing. And this code works on those platforms.
These changes do not break LLP64 (this is the correct name
for the 64-bit Windows model).
>If you're going to submit a patch, submit a correct one. It's that
The buffer size is a separate issue from this. Feel free
to fix it, now that you found it.
The patch I submit fixed your software on systems where
time_t doesn’t fit into a long, all of them, and that
*will* include 64-bit Windows at least in 2038, except
some architectures need these fixes right now. They don’t
break anything else before 2038, and afterwards, nothing
new is broken worsely either.
>simple. If you refuse to do that, then go away.
You know what, I’ll just do that and leave you to fix
your software on your own, clearly you are oh so utterly
competent in doing that.
//mirabilos
--
“It is inappropriate to require that a time represented as
seconds since the Epoch precisely represent the number of
seconds between the referenced time and the Epoch.”
-- IEEE Std 1003.1b-1993 (POSIX) Section B.2.2.2
