Re: [PATCH] Fix blocking read timeouts at a small probability

[Chet Ramey]

On 2/8/21 9:37 AM, Koichi Murase wrote:

>    Instead, I believe, it is more natural to use `select(2)', which is
>    already used to implement `read -t 0'.  In the attached patch
>    `0001-Use-select-2-for-the-read-timeout.patch', I used `select(2)'
>    to implement the timeout of `read(2)'.  When `select(2)' is
>    unavailable (i.e., `HAVE_SELECT' is defined in `config.h'), it still
>    falls back to the old strategy with `SIGALRM', but I believe most of
>    modern systems support `select(2)'.  I tested the behavior with the
>    above test case, and also tested the behavior by hand.  Could you
>    take a look at the patch?

Thanks for the analysis and patch. I'll take a look. At first glance, I
would like to find a simpler way to do it.

Chet
--
``The lyf so short, the craft so long to lerne.'' - Chaucer
                 ``Ars longa, vita brevis'' - Hippocrates
Chet Ramey, UTech, CWRU    chet@case.edu    http://tiswww.cwru.edu/~chet/