Re: io_select EPIPE error case? [Was: [PATCH] Prepare pflocal/io.c for a revised hurdselect.c: Pipe case (for review)]

[Samuel Thibault]

Svante Signell, le Fri 23 Nov 2012 08:00:26 +0100, a écrit :
> I'm not yet fluent in
> the terminology and how to write things properly the GNU Hurd way.

Well, it's not about any "Hurd way", but simply collaborative project :)

> Yes, no errors in the select case only for poll (according to POSIX).
> The remaining thing to check is if the proposed change affects the
> select case in the current hurdselect (until a new one is installed).
> For poll it doesn't, I've checked.

AIU hurdselect.c, it shouldn't change it too much:

              default:
                /* No other error should happen.  Callers of select
                   don't expect to see errors, so we simulate
                   readiness of the erring object and the next call
                   hopefully will get the error again.  */
                type = SELECT_ALL;
                /* FALLTHROUGH */

              case 0:

I.e. it forces the fd as readable, writable, and exception, which
doesn't seem a bad thing to me.

Samuel