home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
Columbia Kermit
/
kermit.zip
/
newsgroups
/
misc.20020314-20021006
/
000088_dkcombs@panix.com_Wed May 22 10:56:41 EDT 2002.msg
< prev
next >
Wrap
Text File
|
2020-01-01
|
4KB
|
115 lines
Article: 13382 of comp.protocols.kermit.misc
Path: newsmaster.cc.columbia.edu!panix!not-for-mail
From: dkcombs@panix.com (David Combs)
Newsgroups: comp.protocols.kermit.misc
Subject: Re: REGET not as clever as RESEND?
Date: Wed, 22 May 2002 07:25:42 +0000 (UTC)
Organization: Public Access Networks Corp.
Lines: 96
Message-ID: <acfh5m$ead$1@reader1.panix.com>
References: <acbqst$bj2$1@panix3.panix.com> <acbrts$1c7$1@watsol.cc.columbia.edu>
NNTP-Posting-Host: panix2.panix.com
X-Trace: reader1.panix.com 1022052342 14669 166.84.1.2 (22 May 2002 07:25:42 GMT)
X-Complaints-To: abuse@panix.com
NNTP-Posting-Date: Wed, 22 May 2002 07:25:42 +0000 (UTC)
X-Newsreader: trn 4.0-test74 (May 26, 2000)
Originator: dkcombs@panix.com (David Combs)
Xref: newsmaster.cc.columbia.edu comp.protocols.kermit.misc:13382
In article <acbrts$1c7$1@watsol.cc.columbia.edu>,
Frank da Cruz <fdc@columbia.edu> wrote:
>In article <acbqst$bj2$1@panix3.panix.com>,
>David Combs <dkcombs@panix.com> wrote:
>: RESEND does a pretty good job of almost instantly
>: recovering up to the %done that it had been
>: at when the prior transfer crapped out.
>:
>: REGET seems to me so far to do a much POORER
>: job of doing that, like recovering up to only
>: maybe 1/3 of how far the prior get had gotten,
>: and then rereading a whole lot of stuff that
>: had (presumably) already been read in the prior
>: kermit-run.
>:
>: Am I the only one who has seen this?
>:
>It doesn't ring a bell. Are you saying it corrupts
>the file, or just takes longer to resynchronize before
>resuming the download?
>
>: (NOTE: I was using -i)
>:
>: QUESTION: what is the SIZE of the blocks
>: that kermit sends? (by default)
>:
>The sender sends blocks (packets) of whatever length
>the receiver tells it to. SHOW PROTOCOL at the
>receiver, see the receive packet length.
>
>: Now, kermit registers 100%, but starts getting
>: errors (two per second) trying to get the last
>: several bytes, it seems.
>:
>That doesn't have anything to do with RESEND or REGET.
>If it happens, it's because of transmission errors,
>which could happen with any kind of file transfer.
>
>Is this a straight dialup connection or a Telnet
>connection or what?
>
>If straight dialup, what kind of modem? What serial
>speed? Are you using RTS/CTS (hardware) flow control?
usr faxmodem
>
>If Telnet, the same questions about the underlying PPP
>driver.
>
>What version of Kermit do you have on your end? (Panix
>has the latest, 8.0.201, on its end, at least if we're
>talking about the same Panix).
>
>Just now I tried downloading a 3MB file from Panix,
>interrupting the download, REGET'ing the file, and
>repeating the process several times until it was fully
>downloaded. There were no delays, no errors, no
>failures. This was using Kermit 95 1.1.21 as a
>client.
>
>- Frank
Thanks, Frank, for the long answer, and suggestions
for finding out some more info.
I'm the same guy that has *always* have had difficulty
using kermit on panix (worked *fine* with netcom!).
I just got a blade-100, and it also has the sun
PCI co-processor (besides the sparc), so
I can load up win2k on that, Solaris8 on the
other.
Once I get that all done, perhaps I could grab that
*large* pizza-box, drive it in to NYC, and somehow
get it up to your office?
Then, once and for all, we'll find out what the
problem is. (Maybe also get kermit95 working
on the other cpu, too?)
As long as you have a screen that can plug into
a sun (and I have cables that connect one kind
of connector to the smaller one that the Blade uses.)
---
No emergency -- I just do downloads at 3am if I want
them to work (and if they are eg > 5mb or so).
David