Compilation fix on my Linux box

[Kiyokazu SUTO]

Citation (with leading "> " of each line) from article:
  <20000507232520O.kondou at inn.do.mms.mt.nec.co.jp>
    by Katsuhiro Kondou <kondou at nec.co.jp> :
> In article <20000507-211424-618c7.suto at ks-and-ks.ne.jp>,
> 	Kiyokazu SUTO <suto at ks-and-ks.ne.jp> wrote;
>
> } The manpage select(2) in my box says that the line "#include
> } <sys/time.h>" is required when one uses the select system call, and
> } that this is comforming to 4.4BSD.
>
> Aha, I've got it.  Just fixed.

Thank you very much.

> } I want to distinguish duplicate articles posted at another gateway
> } host, from other error situnations caused by my gateway program.

Humm..., I did't mention what I meant in my mind :-<, sorry.

My gateway program receives a message from MTA, make some necessary
change on the header of the message, and feed it to INN in streaming
mode.  (From some experiences, this seems to be the fastest way of
gatewaying.)

The program enters into no-CHECK mode soon, because most messages
accepted.

Since all messages do not arrive at two gateway hosts at the same
time, and since these days news feeding is much faster than mail
transfering, some messages arrive at my host as news articles before
my gateway program feeds them to INN, and when the program offers such
articles (with "takethis"), it gets "439" reply code.

I want to distinguish such case from other situation generating "439"
reply code mostly caused by mistake in my header modification code,
and for the latter case, I want to fix the mistake as soon as
possible.  Although I can do such works (for example) by scanning
files such as ~news/log/news{,.notice} and can automate such scan,
I'll be happy if I only need to check reply codes.

But I don't speak of my 2nd patch any more if it may cause problems.
Sorry again for my noisy and long explanation.

-- 
SUTO, Kiyokazu <suto at ks-and-ks.ne.jp>
http://pub.ks-and-ks.ne.jp/pgp-public-key.html