From: jrs@world.std.com (Rick Sladkey) Subject: Re: select() ? Date: Sat, 13 Feb 1993 01:06:57 GMT
>>>>> On 12 Feb 1993 07:12:57 +0100,
>>>>> urlichs@smurf.sub.org (Matthias Urlichs) said:
Matthias> OKOK, but then why isn't the library code right? ;-)
Well, porting mistakes do get made. I should know, I made them.
Matthias> Besides, it could be argued that when select times out it
Matthias> should leave the timeout value alone, since a timeout is
Matthias> already signalled by returning zero and the timeout value is
Matthias> likely to be used again.
I don't really like this solution. If you ever have to copy the
timeval there is no point in preserving the old one for just one
case.
Besides, since POSIX has not yet specified how select should behave,
we should just leave it like it is and follow the prudent advice
of the BSD man page.
Matthias> Anyway, here's a patch to the RPC stuff -- I didn't look for
Matthias> select() calls elsewhere in the library yet, somebody else
Matthias> can do that. And in case you're wondering, no I didn't even
Matthias> compile this, much less test. ;-)
Your patch is correct. I submitted similar (compiled and tested :-)
patches for for libc-4.3. Plus there are two other minor fixes. RPC
broadcasting is fixed and the RPC name lookup left a file handle open.
Mail me if you are interested in the patch.