richmon@astrovax.UUCP (Michael Richmond) (07/24/85)
Has anyone successfully used mail or uucp from his 7300 to another
system? I have set up mine so that it knows about 'astrovax', but when
it tries to call it always fails - the message in the LOGFILE includes
1 astrov!uucp ..[some stuff].. HANDSHAKE FAILED (you are unknown to me)
I think that this is happening because, as you can see, the system name
'astrovax' has been shortened to 'astrov', and when it calls that machine,
the two names, one from astrovax and one from my 7300, don't match. This
'feature' is actually documented in the help file for setting up system
names: it is mentioned that while other system names may have up to nine
letters, only the first six are significant. Is there any way to fix or
get around this? Were the 7300 designers really so careless?
--
Michael Richmond Princeton University, Astrophysics
{allegra,akgua,burl,cbosgd,decvax,ihnp4,noao,princeton,vax135}!astrovax!richmonmjs@eagle.UUCP (M.J.Shannon) (07/24/85)
> > Has anyone successfully used mail or uucp from his 7300 to another > system? I have set up mine so that it knows about 'astrovax', but when > it tries to call it always fails - the message in the LOGFILE includes > > 1 astrov!uucp ..[some stuff].. HANDSHAKE FAILED (you are unknown to me) > > I think that this is happening because, as you can see, the system name > 'astrovax' has been shortened to 'astrov', and when it calls that machine, > the two names, one from astrovax and one from my 7300, don't match. > -- > Michael Richmond Princeton University, Astrophysics No, this is happening because astrovax (the machine you're calling) has no entry for you in its Systems file (this is the honey danber equivalent of L.sys), and is refusing to talk to you. If you don't want astrovax to call you, add an entry reading "mypc7300sname Never". Otherwise add a full entry for your machine. -- Marty Shannon UUCP: ihnp4!eagle!mjs Phone: +1 201 522 6063
honey@down.FUN (Peter Honeyman) (07/25/85)
while we're on (or near, anyway) the subject, in versys.c change if ((iptr=strpbrk(line, " \t")) == NULL) continue; /* why? */ *iptr = '\0'; to if ((iptr=strpbrk(line, " \t")) != NULL) *iptr = '\0'; i.e., /* no good reason */. then the host name alone suffices. this is in honey danber. versys in 4.2bsd is hopelessly fucked up. 4.3 is improved, but, in typical bsd fashion, puts host name aliasing into uucp (as well as into a million other products, er programs, i am sure). peter
adam@npois.UUCP (Adam V. Reed) (07/27/85)
I have used all aspects of UUCP among three 7300's and several Vaxen with zero problems. Even executable binary files can be transfered (without uuencode etc) to a Vax and back, with absolutely no change. UUCP works fine on the phone line at 1200 and on the RS232 line at 9600 bps. The uucp administration is done through menus; it is a breeze, and requires no computer literacy to speak of. About the six-character limit on machine names: there is a lot of software out in the world which is limited to six-character names, so anyone who wishes to communicate with the whole UNIX (TM) world DOES NOT USE LONGER NAMES. A menu-based administration interface REQUIRES a length limit (think about it), and it makes sense for this limit to be backward-compatible with existing software. Giving your machine a name longer than the DOCUMENTED standard length is asking for what you get. Disclaimer: My opinions are solely my own (my company probably wouldn't claim them anyway). Adam Reed (npois!adam)
honey@down.FUN (Peter Honeyman) (07/27/85)
if they were smart, they would have DOCUMENTED a 1 character standard and auctioned off the host names. the samurai/astrovax problem was solved by adding LOGNAME=uubroken MACHINE=samurai MYNAME=astrov to the Permissions file on astrovax. to run honey danber on the 7300, you have to change a pair of lines in uucp.h from "if (a && b && !strcmp(a,b))" to "if (a != 0 && b != 0 && !strcmp(a,b))" (the compiler generates incorrect code.) other than that, honey danber runs just fine on the 7300 (even to the point of talking to rogue systems with illegal host names like princeton and astrovax). peter
guy@sun.uucp (Guy Harris) (07/27/85)
> it is mentioned that while other system names may have up to nine > letters, only the first six are significant. Is there any way to fix or > get around this? Were the 7300 designers really so careless? Blame where blame is due; it's not Convergent's fault. UUCP's queue is implemented as files in a spool directory (or directories). The queue entry is basically the directory entry, with the attributes of the entry encoded in the name. Except in 4.[23]BSD, file names are limited to 14 characters. One of the attributes of the entry is the name of the remote host. 7 characters are already spoken for, which leaves 7 characters for the host name. The System V UUCP grabbed an additional character which meant that there were only 6 characters for the host name. The "nine letters" comes from the PWB/UNIX system call (propagated into Systems III and V) to obtain the current host name from the operating system (but I'll bet that the *real* value is eight, not nine; the character array into which the name is returned is probably really eight characters - a nice power of two - plus a NUL terminator). Other versions of UNIX have different limitations. I don't know what the limit, if any, was in V7, where the "current" version of UUCP first appeared (UUCPs presently running are descendants of the V7 one). The reason why it says 6 in one place and 9 in another is that the two bits of code were designed by different people as parts of different versions of UNIX. The 7300 merely inherited those two bits of code; its designers are blameless. Guy Harris
honey@down.FUN (Peter Honeyman) (07/29/85)
blame where blame is due: it is convergent's fault. they are (apparently) looking at the ^PShere=astrovax line and comparing it to astrov (to which guy's other comments pertain). their version of uucp is the only one in which this comparison fails. peter
peter@kitty.UUCP (Peter DaSilva) (08/01/85)
Before SV (and maybe SIII) the documented standard length was 7 characters.
guy@sun.uucp (Guy Harris) (08/02/85)
> blame where blame is due: it is convergent's fault. they are (apparently) > looking at the ^PShere=astrovax line and comparing it to astrov (to which > guy's other comments pertain). their version of uucp is the only one in > which this comparison fails. The failing comparison may be their fault, but the 6-character limitation on site names isn't, whcih is what my comments pertained to. Guy Harris
honey@down.FUN (Peter Honeyman) (08/03/85)
peter dasilva asserts:
Before SV (and maybe SIII) the documented standard length was 7 characters.
try telling that to people on research, where uucp was born.
12345678
peterkre@ucbvax.ARPA (Robert Elz) (08/07/85)
In article <553@down.FUN>, honey@down.FUN (Peter Honeyman) writes: > > try telling that to people on research, where uucp was born. > 12345678 > I suppose that's why V7 mail (and all derived versions thereafter, though its ifdef'd out on 4.2) contain (approx) .. 1234567 if (strncmp(name, "researc", 7)) { strcpy(rsys, "research"); } Robert Elz ucbvax!kre