brian@natinst.com (Brian H. Powell) (09/04/89)
We've got an AT&T 3b2/700 running System V.3.2 with HoneyDanBer UUCP.
Unfortunately, we don't have source. We're trying to set things up to call
our France office (a Xenix system) with UUCP. Problem is, it takes about 45
or 50 seconds to make the call and have the modems connect. A few seconds
before we connect, uucico invariably times out, and tries to dial again.
We've talked to AT&T, and they said the timeout value is hard-coded, and
hence cast in stone. I explained to them that there should be a way to patch
our uucico to change the hard-coded timeout value. After thinking about it
for several minutes, AT&T decided they needed to think about it some more.
Perhaps they'll be able to send us a new version of uucico with a different
hard-coded value.
Does anybody have a workaround for this? We tried adding delays in
various places, but to no avail. Anybody know the "proper" place to add
delays that might affect the beginning of the timeout period?
Thanks for any help. Please send mail.
Brian H. Powell National Instruments Corp.
brian@natinst.com 12109 Technology Blvd.
uunet!cs.utexas.edu!natinst!brian Austin, Texas 78727-6204
AppleLink:NATINST (512) 250-9119brian@natinst.com (Brian H. Powell) (09/05/89)
In article <4309@natinst.natinst.com>, I wrote: > > We've got [a system with HoneyDanBer]... > We're trying to set things up to call > our France office (a Xenix system) with UUCP. Problem is, it takes about 45 > or 50 seconds to make the call and have the modems connect. A few seconds > before we connect, uucico invariably times out, and tries to dial again. I've received about 20 answers to this, including some asking me to summarize. There were two main solutions. The one we implemented (the first one we tried worked, so it was also the last one we tried), was to change the Dialer script from this: ... ATDT\T\r ... to this ... ATDT\T\r\d\d\d\d\d\d\d\d\d\d\c Uucico doesn't start timing out until after the dial sequence completes, so this delays the end of the dialing sequence, but gets the modem started first. The other solution proposed, equally viable, is to add an extra "expect" for the CONNECT message. Basically, again in the Dialers file, change this: ... ATDT\T\r CONNECT to this: ... ATDT\T\r CONNECT \d\d\d\d\d\d\c CONNECT ^send ^expect ^send ^expect This must be a common topic, because I got so many replies. Thanks to everybody who sent mail. Next, we're working on our U.K. office... Brian H. Powell National Instruments Corp. brian@natinst.com 12109 Technology Blvd. uunet!cs.utexas.edu!natinst!brian Austin, Texas 78727-6204 AppleLink:NATINST (512) 250-9119
bdb@becker.UUCP (Bruce Becker) (09/05/89)
In article <4309@natinst.natinst.com> brian@natinst.com (Brian H. Powell) writes: | | We've got an AT&T 3b2/700 running System V.3.2 with HoneyDanBer UUCP. |Unfortunately, we don't have source. We're trying to set things up to call |our France office (a Xenix system) with UUCP. Problem is, it takes about 45 |or 50 seconds to make the call and have the modems connect. A few seconds |before we connect, uucico invariably times out, and tries to dial again. | We've talked to AT&T, and they said the timeout value is hard-coded, and |hence cast in stone. I explained to them that there should be a way to patch |our uucico to change the hard-coded timeout value. After thinking about it |for several minutes, AT&T decided they needed to think about it some more. |Perhaps they'll be able to send us a new version of uucico with a different |hard-coded value. | Does anybody have a workaround for this? We tried adding delays in |various places, but to no avail. Anybody know the "proper" place to add |delays that might affect the beginning of the timeout period? In your "Dialers" file you can put in delays after the phone number variable. Here's an example in Hayes format: hayes =W-, "" \M\d\c ATDT\T\r\c ATDT \d\d\d\d\d\m\c CONNECT System that dial to Telebit sites which are configured to respond with "PEP tones" last in the sequence have a similar timeout problem, which this approach also solves. Vary the number of "\d"'s before the "CONNECT" to suit your problem... Cheers, -- (__) Bruce Becker Toronto, Ont. w \cc/ Internet: bdb@becker.UUCP, bruce@gpu.utcs.toronto.edu `/\/-e BitNet: BECKER@HUMBER.BITNET _/ >_ Does a cow barn == a mu meson? - F. L. Wright