Taylor Blau <me@xxxxxxxxxxxx> writes:
> So it appears that the old version is ever-so-slightly faster than the
> new one. But it's so noisy, and the regression is so small that it's
> hard to notice it at all.
>
> So I wouldn't strongly oppose the patch based on those numbers, but in
> principle it seems flawed.
Thanks for writing and reviewing.
As long as we were touching the function, I suspect that
the logic should become more like
if (fd #0 is not open)
open /dev/null read-only and give it to fd #0
if (fd #1 is not open)
open /dev/null write-only and give it to fd #1
if (fd #2 is not open)
open /dev/null write-only and give it to fd #2
with opening of /dev/null optimized not to happen when not needed.
