Re: ENOTSOCK

[David Michael]

On Mon, Mar 13, 2017 at 12:43 PM, Samuel Thibault
<samuel.thibault@gnu.org> wrote:
> Hello,
>
> David Michael, on dim. 12 mars 2017 19:43:40 -0700, wrote:
>> A few of the socket-related functions in glibc don't seem to return
>> ENOTSOCK.  Would it be more "correct" to return this as opposed to
>> EMIG_BAD_ID?  In particular, can at least send() be made to return
>> ENOTSOCK for PulseAudio compatibility?  The following example uses
>> both MIG_BAD_ID and EOPNOTSUPP as some other functions do, although
>> I've only seen it return EMIG_BAD_ID.
>
> Oh indeed, thanks for the suggestion, now commited upstream and in
> Debian.

Thanks for taking care of that, but could it also be backported to the
tschwinge/Roger_Whittaker branch?  I'm still using that as my glibc
source for Hurd, and I imagine it will be a while before it catches up
to where upstream is now.

Thanks.

David