Path: utzoo!attcan!uunet!cs.utexas.edu!wuarchive!mit-eddie!bloom-beacon!eru!hagbard!sunic!nuug!sigyn.idt.unit.no!eik!usenet From: tl@cmi.no (tom lislegaard) Newsgroups: comp.sys.mac.comm Subject: MacTCP problem Message-ID: <1990Dec10.122937.25016@eik.ii.uib.no> Date: 10 Dec 90 12:29:37 GMT Sender: usenet@eik.ii.uib.no (Usenet posting account) Reply-To: tl@cmi.no (tom lislegaard) Organization: Chr. Michelsen Institute, Bergen, Norway Lines: 50 I have tried a few MacTCP based applications lately, and experienced strange problems with certain programs and certain hosts. The characteristic behaviour of a failing program is that the connection is made and recognized as 'established' by both parties, but the called port does not respond with any data. Each packet from the client is ACK'ed by the server, but thats all. A few examples: - Using the example hypercard stack of macTCP-toolkit I can connect without problems to ftp, nntp, or smtp ports of a sun (any OS), but telnet does not work. Same with a mips running Umips 4.30. A pyramid with OSx 5.0 is dead silent on all ports with this toolkit. Trying a vax with some ancient version of excelan tcp/ip I'm able to login as normal. - NCSA telnet 2.3 works with any of the systems above. - Bill Cramers newsreader 'TheNews' has no trouble with either the pyramid or sunOS nntp. - The mail client 'eudora' works with any sun, but not with pyramid. (by the way both eudora and TheNews are quite nice, much better than the hypercard stuff like Mews, netnews, popmail....). The mac is directly on the ethernet with a excelan etherport II card. I've tried a few different things on the mac side: - system versions 6.03 and 6.05 - macTCP 1.0, 1.0.1, 1.0.1a2 - pulling out any startup documents and the like from the system folder and booting only finder. Traceing with etherfind reveals only differences in the advertised window size and max segment size ( why does macTCP use different mss for different programs ?). I hope some macTCP or tcp/ip experts can explain whats going on here ( or even better - propose a solution ). Tom Lislegaard CMI