treval@tauros.UUCP (Trevor Luker) (04/14/88)
g'day has anyone got any idea as to whether it is possible to run uucp (uucico) over the Sun X25 link software. Currently, with our installed software, this is not possible. The ideal setup would be a pseudo-tty device that looks like a tty to uucico, along with a pseudo-acu that accepts the NUA to call. Is this possible? I could probably write a tty-like driver that interfaces to the X25 packet level interface, and an ACU-device that does tha calling but I am sure that someone must have done it already done it. Any help / suggestions? Thanks, treval -------------------------------------------------------------------------------- Trevor Luker Commission Des Communautes Europeennes, treval@tauros.UUCP Batiment Jean Monnet, C2/60 ...!mcvax!prlb2!tauros!treval Plateau du Kirchberg (+352) 4301 4620 L-2920 Luxembourg --------------------------------------------------------------------------------
csg@pyramid.pyramid.com (Carl S. Gutekunst) (04/16/88)
In article <287@tauros.UUCP> treval@tauros.UUCP (Trevor Luker) writes: > has anyone got any idea as to whether it is possible to run uucp > (uucico) over the Sun X25 link software. The biggest problem you'll have in Luxembourg is that Sun's UUCP does not have the 'f' protocol. This is a seven-bit, printable-ASCII-only, streaming, non packetized uucp protocol that is a must over international X.25 links. If you have a source license, you could run 4.3BSD (which includes 'f') or HoneyDan- Ber (to which 'f' can be trivially added). UUCP source would also solve the other half of your problem: you could write a C-code dialer for the SunLink X.25 interface. But you're out of luck without a source license, or until GNUUCP is ready. I am amazed that Sun does not yet support UUCP/X.25. While there are only a few sites in the USA that use UUCP/X.25 extensively (uunet, pyramid, hplabs, maybe two or three more), it is almost _de_rigeur_ in Europe and Canada. Maybe if enough of us rattle the bars on Sun's cage, we can convince the powers that be to provide Bill Shannon with the extra body he needs to port and support a new UUCP version.... Uh, anyone know if Ultrix supports UUCP/X.25? <csg>
egisin@watmath.waterloo.edu (Eric Gisin) (04/18/88)
In article <19772@pyramid.pyramid.com>, csg@pyramid.pyramid.com (Carl S. Gutekunst) writes: > The biggest problem you'll have in Luxembourg is that Sun's UUCP does not have > the 'f' protocol. This is a seven-bit, printable-ASCII-only, streaming, non > packetized uucp protocol that is a must over international X.25 links. If you I think there is also an 'x' protocol, which uses X.25 directly and can use all 8 bits. 'f' was designed for use over X.2[89], which has the 7 bit restriction. > I am amazed that Sun does not yet support UUCP/X.25. While there are only a > few sites in the USA that use UUCP/X.25 extensively (uunet, pyramid, hplabs, > maybe two or three more), it is almost _de_rigeur_ in Europe and Canada. Maybe Not in Canada. Trailblazers over phone lines are much cheaper than Datapac, the national X.25 carrier. And Bell has evening discounts.
csg@pyramid.pyramid.com (Carl S. Gutekunst) (04/19/88)
In article <18341@watmath.waterloo.edu> egisin@watmath.waterloo.edu (Eric Gisin) writes: >I think there is also an 'x' protocol, which uses X.25 directly >and can use all 8 bits. 'f' was designed for use over X.2[89], >which has the 7 bit restriction. 'x' provides no error correction at all, and does not permit in-band flow control (XON/XOFF). It is suitable only for simple point-to-point links using dedicated internal X.25 hardware. The 7-bit-printable-ASCII restriction comes from international X.25 gateways, many of which insist on swiping the eigth bit for parity or somesuch. A few also do funny mappings of control characters, like munging tabs. If I set up a raw X.25 virtual circuit between here and West Germany, it will be 7 bits and there is nothing I can do about it. The X.29 and X.28-1980 standards require 8-bit transparency. Not until 1984 is there any mention of DTE-to-PAD parity generation (X.3 parameter 21), and it is disabled by default. Note that everyone who connects to uunet for UUCP via dialup Tymnet is using the 8-bit 'g' protocol over X.29; similarly, we run 'g' over X.28 PADs all the time. (Of course, we have to disable flow control on parameters 5 and 12. But calling into a Pyramid, it does that automatically.) Finally, many sites (e.g., munnari) run 'f' right on top of X.25, and don't bother with the hassle of X.29 at all. <csg>
egisin@watmath.waterloo.edu (Eric Gisin) (04/23/88)
In article <20060@pyramid.pyramid.com>, csg@pyramid.pyramid.com (Carl S. Gutekunst) writes: > [...] > The 7-bit-printable-ASCII restriction comes from international X.25 gateways, > many of which insist on swiping the eigth bit for parity or somesuch. A few > also do funny mappings of control characters, like munging tabs. If I set up a > raw X.25 virtual circuit between here and West Germany, it will be 7 bits and > there is nothing I can do about it. It's difficult to believe CCITT is so stupid to allow this in X.25 VCs. Maybe I'll have one last look at the red book to verify it. What happens if one wants to run IP, DECNET, or OSI across such a gateway? I guess you don't.
james@sunne.Sun.COM (04/25/88)
Sun's Consulting Group has recently completed implementation of the uucp-over-X25 protocol. Larry Kluger at Sun Germany (lkluger@sun.com (?)) has info. If you can't find him, let me know and I can get a better path ----james James Triplett Sun Microsystems, Lexington, Mass jtriplett@sun.com ...sun!sunne!james
jh@tut.fi (Juha Hein{nen) (04/26/88)
I haven't seen the original article but only a reply that wonders about 7 bit characters in X.25. Seven bits/char is not a problem if you running uucp f-proto which was specially designed for X.25, But what is the conclusion about the subject line: has anybody been able to run UUCP over SUnLink X.25? We tried it about a year ago with no luck. The setup was such that we created a ptty with getty on it and then specified that ptty in L-devices. When uucp was started, it first logged on in the same machine using this line and then issued a pad command to the other uucp host (or actually pad command was the login shell). We were able to start uucico in the other end but the transfer always failed because of bad checksum. We never got further in analysing the problem except that it looked like the other end was receiving at least parts of the file twice. This may have been caused by improper setting of some of the X.28/29 parameters or stty stuff. We are still waiting for SunLink X.25 version 5 which should include a profiling cabability and will then again start playing with it. So, has someone else had better luck? I know that people are running uucp over the EAN X.25 code. -- Juha Heinanen, Tampere Univ. of Technology, Finland jh@tut.fi (Internet), tut!jh (UUCP), jh@tut (Bitnet)