Better fix for qpopper TOP problem.

svn path=/trunk/; revision=1790

3 files changed, 22 insertions, 8 deletions

diff --git a/NEWS b/NEWS
index 2228f079..dd2da574 100644
--- a/NEWS
+++ b/NEWS
@@ -7,6 +7,11 @@
 				Release Notes:
 
 ------------------------------------------------------------------------------
+fetchmail-4.4.6 ()
+* Better workaround for TOP bounds-checking in qpopper 2.3+.
+
+There are 270 people on fetchmail-friends and 212 on fetchmail-announce.
+
 fetchmail-4.4.6 (Sat May 16 14:16:49 EDT 1998)
 * Fix a bug introduced by my 4.4.5 bug-fix release :-(
 * Make the internal line buffer eight times longer to cope with idiot
diff --git a/fetchmail-FAQ.html b/fetchmail-FAQ.html
index 5f120d16..010275c1 100644
--- a/fetchmail-FAQ.html
+++ b/fetchmail-FAQ.html
@@ -10,7 +10,7 @@
 <table width="100%" cellpadding=0><tr>
 <td width="30%">Back to <a href="index.html">Fetchmail Home Page</a>
 <td width="30%" align=center>To <a href="/~esr/sitemap.html">Site Map</a>
-<td width="30%" align=right>$Date: 1998/05/14 01:35:48 $
+<td width="30%" align=right>$Date: 1998/05/16 19:17:42 $
 </table>
 <HR>
 <H1>Frequently Asked Questions About Fetchmail</H1>
@@ -1774,11 +1774,16 @@ If your POP3 server fails to sanity-check TOP requests and clip them at
 the next end-of-message, it will fetch the entire tail of the mailbox
 starting with the current message.  The qpopper 2.41 beta 1 server is 
 known to have this bug (though the qpopper 2.2 found on many Unixes
-does not).<P>
+does not).  It's been reported to me that the logic changed in 2.3.<P>
 
 To work around it, set the <code>fetchall</code> option.  Under POP3
 only, this now has the side effect of forcing RETR use.<P>
 
+(It is possible this tip is no longer necessary.  At least one tester
+has claimed that the bounds check works but was fooled by an overflow
+condition in the TOP argument.  Decrementing the argument may have
+fixed this.)<P>
+
 <hr>
 <h2><a name="O2">O2. Every time I get a POP or IMAP message the header
 is dumped to all my terminal sessions.</a></h2>
@@ -1910,7 +1915,7 @@ Re-ordering messages is a user-agent function, anyway.<P>
 <table width="100%" cellpadding=0><tr>
 <td width="30%">Back to <a href="index.html">Fetchmail Home Page</a>
 <td width="30%" align=center>To <a href="/~esr/sitemap.html">Site Map</a>
-<td width="30%" align=right>$Date: 1998/05/14 01:35:48 $
+<td width="30%" align=right>$Date: 1998/05/16 19:17:42 $
 </table>
 
 <P><ADDRESS>Eric S. Raymond <A HREF="mailto:esr@thyrsus.com">&lt;esr@snark.thyrsus.com&gt;</A></ADDRESS>
diff --git a/pop3.c b/pop3.c
index 09565c33..7dce9264 100644
--- a/pop3.c
+++ b/pop3.c
@@ -531,7 +531,7 @@ static int pop3_fetch(int sock, struct query *ctl, int number, int *lenp)
      * the server.
      *
      * The line count passed is the maximum value of a twos-complement
-     * signed integer (we take advantage of the fact that, according
+     * signed integer minus 1 (we take advantage of the fact that, according
      * to all the POP RFCs, "if the number of lines requested by the
      * POP3 client is greater than than the number of lines in the
      * body, then the POP3 server sends the entire message.").
@@ -542,14 +542,18 @@ static int pop3_fetch(int sock, struct query *ctl, int number, int *lenp)
      *
      * Also suppress TOP if fetchall is on.  This is a kludge to get
      * around a really obnoxious bug in qpopper 2.41 beta 1 that must
-     * have been introduced sometime after the 2.2 I tested with.  The
-     * newer version doesn't bounds-check TOP requests to clip them
-     * at the next end of message.  Grrrrr...
+     * have been introduced sometime after the 2.2 I tested with (Erwan
+     * Mas claims it was in 2.3).  The newer version doesn't bounds-check TOP
+     * requests to clip them at the next end of message.  Grrrrr...
+     *
+     * (Erwan MAS <mas@nic.fr> thinks this was actually due to the TOP
+     * argument overflowing and recommended I decrement it.  Hmmm...
+     * maybe we can take out the fetchall kluge?)
      */
     if (ctl->keep || ctl->fetchall)
 	gen_send(sock, "RETR %d", number);
     else
-	gen_send(sock, "TOP %d 2147483647", number);
+	gen_send(sock, "TOP %d 2147483646", number);
     if ((ok = pop3_ok(sock, buf)) != 0)
 	return(ok);