This is the mail archive of the
libc-hacker@cygnus.com
mailing list for the glibc project.
Re: ENOTSUP vs EOPNOTSUPP
- To: Roland McGrath <roland@frob.com>
- Subject: Re: ENOTSUP vs EOPNOTSUPP
- From: Scott Bambrough <scottb@corelcomputer.com>
- Date: Sat, 23 Jan 1999 23:53:12 -0500
- CC: Ulrich Drepper <drepper@gnu.org>, GNU libc hacker <libc-hacker@gnu.org>
- Organization: Corel Computer Corporation
- References: <199901232020.PAA31312@baalperazim.frob.com>
Roland McGrath wrote:
>
> Please explain the following change:
>
> 1999-01-22 Ulrich Drepper <drepper@cygnus.com>
>
> * sysdeps/gnu/errlist.awk: Mark ENOTSUP as alias for EOPNOTSUPP.
>
> I do not think it is correct. Moreover, such a change should never be made
> without discussion, because it affects multiple platforms, and affects the
> universal glibc API.
>
> I added ENOTSUP because it is a 1003.1-specified error and we did not have
> it before. I considered for some time making it and EOPNOTSUPP two names
> for the same thing. 1003.1 does not specify EOPNOTSUPP, so it says nothing
> on the subject. I do not have the draft for the 1003.1[a-z] sockets spec,
> so I don't know what it says.
>
> I did look at the Unix98 spec, which specifies both ENOTSUP and EOPNOTSUPP.
> It furthermore specifies that all the error codes it specifies must be
> unique values, with the specific exception that EAGAIN might == EWOULDBLOCK.
> For that reason, I concluded ENOTSUP must be distinct from EOPNOTSUPP.
I ran into a problem building tcl with these two. They were case entries in a
switch statement.
Just my two cents worth.
Scott