kls@ditka.Chicago.COM (Karl Swartz) (06/15/91)
Bdale Garbee <bdale@col.hp.com> writes: > - which uucp do folks (Karl?) have running? I've played with the one > on plains, and while I've got it working fine when my other system > calls the pc532, outbound "dials" die in mid-session with protocol > errors. I've got both hdb and bsd sources around, am tempted to > port the bsd4.3reno uucp, but seems like real work. And Sverre Froyen <sverre@seri.gov> writes: > I use Karl's (?) uucp port but have not tried dialing out. The original > comments state that dialouts would not work because of missing hooks > in the tty driver. The uucp on plains has some major bugs (like it gets the wrong packet number in NAKs so they don't work!) which I fixed long ago, along with adding quite a bit of new functionality (sending multiple files and allowing comments and blank lines in L.sys) and making it look more like a reasonably contemporary uucp (uuname behaves as one expects and log files are somewhat more useful). I posted the modified files (not just patches but you still need the plains stuff for docs and all that) sometime last year, though I would be happy to help get the whole thing onto some pc532 archive sites. Regarding dialout, there was quite a bit of missing functionality in the 1.3 tty driver, and unfortunately the libc interface to ioctl() was only written for one special case of arguments. Lacking sources for that side of the equation there wasn't much I could do, and while I begged to get those sources I'm not sure I ever did. Months later I did get some stuff, which may have had what I needed, but by then many other things were going on and I didn't have time to look at it. Now there's 1.5h, which may fix the problems, but after looking at the patches and realizing that I needed to figure out a way to read the original PC disks rather than whatever I already have from Bruce, I never even bothered to ship the patches home from the office. > - is it possible to get smail2.5, or something else, to do the minimal > stuff necessary to treat all addresses in '@' notation, but recognize > that *anything* not local should be uucp'ed to a neighboring machine? > We do this inside HP, and it'll cause me much less grief if I don't > introduce '!' addresses any more than necessary. Smail 2.5 worked just fine on 1.3 and I imagine the same would be true for 1.5h though I haven't tried it myself. Attached is the tail end of some instructions I posted a couple of months ago for getting smail working. Subject: Re: Smail Date: Sat, 20 Apr 91 13:15:56 PDT ... BTW, you will need to fiddle with a few pieces of smail 2.5 to get it working on Minix: Makefile $< doesn't work in the standard make, so the rule $(OBJECTS): defs.h cc $(CFLAGS) -c $< fails. Either use GNU make or expand the rule into separate rules for each object file. misc.c The include of <sys/utsname.h> should be wrapped within an #ifdef UNAME. Also, the libraries do not offer tzname[] which is used by the arpadate() function so I compiled in my own value. defs.h Make sure #define UNAME is commented out and #define HOSTNAME with your explicit hostname. Happy mailing! -- Karl Swartz |INet kls@ditka.chicago.com 1-408/223-1308 |UUCP {uunet,decwrl}!daver!ditka!kls |Snail 1738 Deer Creek Ct., San Jose CA 95148 "The more people I meet the more I like my dog."