Re: [pflocal] SO_RCVTIMEO returns ENOPROTOOPT; please implement on the AF_UNIX socket-option path

Hi Samuel,

> Which entails adding timeouts on libpipe, contribution welcome!

Right. I had not traced as far as libpipe; the original report
framed the fix as a pflocal dispatch-table extension, but if
libpipe has no timeout primitive then the real work lives in the
lower layer first (libpipe gains a timeout-bearing read path),
with the pflocal SO_RCVTIMEO dispatch as the thin wrapper on top.

I am not in a position to take the libpipe side of this on a near
horizon, so I will not block the slot. If someone else has
cycles for it I am happy to be the second reviewer / tester on
canonical Debian GNU/Hurd 0.9.

For the immediate emacsclient noise the gnu-emacs-side patch is
already in review (bug#81160, "[PATCH] emacsclient: suppress
ENOPROTOOPT noise from SO_RCVTIMEO on GNU/Hurd"), so the user-
visible warning stops at the client even before pflocal grows the
option.

Thanks,
Borja Tarraso
[email protected]

El dom, 31 may 2026 a las 12:32, Samuel Thibault
(<[email protected]>) escribió:
>
> Hello,
>
> Borja Tarraso, le sam. 30 mai 2026 22:11:48 +0300, a ecrit:
> > This is a request to add SO_RCVTIMEO (and SO_SNDTIMEO by symmetry) to
> > pflocal's socket-option dispatch.
>
> Which entails adding timeouts on libpipe, contribution welcome!
>
> Samuel