PMMail/2 and Broken Telephone Connectinons

[Agents]

Thank you Rodney -- now if we only could get Southsoft to address this issue (and one should not have to do this, the 
program should automatically time out and then restart whaterver it was doing).  And the issue of suboptimal threading....

Hakan

On Wed, 14 Jul 1999 10:08:05 -0400 (EDT), Rodney R. Korte wrote:

>On Tue, 13 Jul 1999 19:18:46 -0700, Marty Abrego wrote:
>
>>On Tue, 13 Jul 1999 16:59:52 -0700, Chris Adams wrote:
>>
>>>>Unfortunately I got a somewhat similar cavalier
>>>>response when I twice complained about this over
>>>>a year ago.  I was told that this was a limitation
>>>>of the TCP/IP stack in OS/2...  Clearly this needs
>>>>to be addressed and I quite agree with you that it
>>>>is hard to recommend a package that does not handle
>>>>this problem appropriately.
>>
>>Strange... I just tested this by emailing myself a file with a
>>sizable attachment (so I have time to break the connection).  After I
>>started retrieving it, I let it reach 10% and then pulled the plug on
>>the DSL modem.  After 30 seconds, or so, I plugged it back in.  After
>>about 30 more seconds, PM Mail picked up where it left off -- no
>>problem at all.
>
>C'mon people, get a clue.  Learn about TCP/IP and your connections.
>It's not that difficult.
>
>Sure, if you have a static IP address or happen to get the same
>IP address when you dial in the second time, PMMail will pick up
>where it left off, assuming the server didn't get a client-closed
>timeout from it's tcp stack.
>
>On the other hand, if you get a different IP address when you 
>dial in the second time (ever heard of dynamic addressing with
>SL/IP and PPP?), you have a snowball's chance in hell of PMMail
>continuing its download.
>
>What PMMail needs is a "Stop the current connection(s), and try
>again."  IT DOES NOT DO THIS RELIABLY IN THE CURRENT VERSION 
>(OS/2 2.00.1500).  (For the technical, close all current sockets
>and restart- don't wait for recv() to timeout.)
>
>This is all true, regardless of the comments of those who
>claim "Hey, you're wrong, it works."  NO IT DOESN'T.
>--
>Rodney R. Korte                   OS/2. Operate at a higher level.
>rkorte@psu.edu                 ---> MIME, PGP (at URL below) welcome.
>Phone: 814-863-0817  Fax: 814-863-6239
>http://www.personal.psu.edu/rrk102
>
>
>
>
>