On Thu, Dec 04, 2025 at 05:20:06 -0700, Stan Marsh wrote:
> From: Eduardo Bustamante
> > The obvious interpretation is 3 seconds and 100,000 microseconds.
> > It's fixed-point arithmetic.
>
> Oh, I think it is clear that that is what it means. It is just that
> that is not what it says.
>
> I don't think it was wise of Chet to introduce this distraction.
>
> It is the sort of thing that can only makes sense if you are familiar
> with the function signature of select(2).
You're seriously overthinking this. It's extremely clear and obvious
what this means, at least in locales where "." is the decimal point:
y. The `-t' option to the `read' builtin now supports fractional timeout
values.
The fact that TMOUT also accepts fractional values was not documented
as far as I know, but now that we *know* it does, it's equally clear
and obvious how to supply those fractional values.
The only issues are the handling (or non-handling) of "," or other
locale-specific decimal points, and the documentation.
Nobody needs to understand the internals of select(2) to understand
how TMOUT=3.1 works.
