home *** CD-ROM | disk | FTP | other *** search
- In article <2k95iv$56a@Mercury.mcs.com> karl@MCS.COM (Karl Denninger) writes:
- >Path: panix!ddsw1!not-for-mail
- >From: karl@MCS.COM (Karl Denninger)
- >Newsgroups: comp.protocols.tcp-ip.ibmpc,alt.winsock
- >Subject: Re: I'm ready to spend money - free and shareware WINSOCK not ok
- >Date: 20 Feb 1994 20:17:03 -0600
- >Organization: MCSNet Ops, Chicago, IL
- >Lines: 51
- >Message-ID: <2k95iv$56a@Mercury.mcs.com>
- >References: <2k6fag$fjh@Mercury.mcs.com> <patlee.237.000ECA2D@panix.com> <2k90jk$1dm@Mercury.mcs.com> <ashok.469.2D680B62@biochemistry.cwru.edu>
- >NNTP-Posting-Host: mercury.mcs.com
- >Xref: panix comp.protocols.tcp-ip.ibmpc:21159 alt.winsock:4537
-
-
- >In article <ashok.469.2D680B62@biochemistry.cwru.edu>,
- >Ashok Aiyar <ashok@biochemistry.cwru.edu> wrote:
- >>In article <2k90jk$1dm@Mercury.mcs.com> karl@MCS.COM (Karl Denninger) writes:
- >>
- >>>Right now there is, IMHO, NO stable Shareware/Freeware Windows selection
- >>>available with internal communications and dial scripting capability.
- >>
- >>>And, until there is something I can test and prove out to be stable, I will
- >>>continue to make this known.
- >>
- >>That statement is fine as long as you qualify it to state that the
- >>Trumpet Winsock is unstable on your test system.
-
- >It is unstable on, at least:
- >1) My IBM/Leopard DX/60 that runs everything else.
- >2) My generic no-name 486DX/33 clone. Both with 8MB RAM, 16550s, etc.
- >3) At least three customers systems here. All common Windows machines
- > with nothing special about any of them. One of those customers
- > says he is stable on A17. I have one machine which is, and one
- > which is not.
-
- >My problem, again, isn't with actual crashes. It is with lost connections
- >that "die" for long periods of time, with long being defined as 1 sec < x <
- >10 minutes. Note that ANOTHER session continues just fine. That is, I can
- >have a Telnet hang, and open a second one -- which may or may not hang. The
- >first hung one wille eventually pick back up. The second one will eventually
- >hang up too, but it will also eventually start working again. While either
- >or both of these are hung I can open an FTP, and that will also work --
- >and all three apps will then run at some point. But they will also all
- >show the "pause" problem eventually, with QVT telnets being hit the worst.
-
- >The same use pattern on a MAC with InterSLIP or MAC/PPP operates perfectly.
-
- >This is pretty clearly a stack and/or SLIP driver problem.
-
- That's not at all clear to me. I'm an extremely happy Trumpet Winsock
- user (v1.0a), and I have the very symptoms you describe, not constantly,
- but frequently enough that I had to investigate. Using the excellent
- WS_PING utility, and hop-check utility from a telnet session, I've
- been able to reliably place the blame on the net itself every time
- I bothered to check. I.e. if your modem isn't initialized to support
- the proper handshaking (SLIP is very demanding regarding your modem
- init string), if your SLIP server has connectivity problems, if
- a particular net link that is important to your SLIP server's
- Internet connectivity is down, if the remote telnet host is glitchy
- or down, any of these would produce the behavior you describe.
-
- BTW Trumpet Winsock has excellent diagnostic features that would
- show you whether your client apps are having trouble with the TCP/IP
- stack, or whether they're sending messages just fine but are stuck
- waiting on a remote response.
-
- I'm using TrumpWinsock 1.0a with a Gateway Telepath (a notoriously
- finnicky modem) and internal SLIP under both Windows v3.1 and Windows
- for Workgroups v3.11.
-
- [... stuff deleted]
-
- >--
- >--
- >Karl Denninger (karl@MCS.COM) | MCSNet - Full Internet Connectivity (shell,
- >Modem: [+1 312 248-0900] | PPP, SLIP and more) in Chicago and 'burbs.
- >Voice/FAX: [+1 312 248-8649] | Email "info@mcs.com". MCSNet is a CIX member.
-
-
-
- --
- Daniel Gross ANTI-DISCLAIMER: I take full responsibility
- FLOW Research, Inc. for my opinions, and they DO reflect the
- entropy@panix.com beliefs and policies of FLOW Research, Inc.
-
-