dg@lakart.UUCP (David Goodenough) (06/12/89)
A couple of questions: 1. I have implemented a "g" protocol program on my PC at home (don't ask why I didn't use UUPC - just look up pallio.uucp in the maps). I have one problem with it - since I can't do parallel disk I/O and serial I/O, I "shut down" serial I/O when going to the disk. Very simply, I stop transmitting, and then wait for a second, buffering any stuff that arrives for later consumption. I then go and read / write the disk. I would have expected the transmission to start up again just fine after the break (it's only 3 to 5 seconds, but for some reason the UNIX end gets in a tangle, and it takes several retries to get re-started. It always _DOES_ get going again, but it takes a while and a few wrong packets. This is a timing issue, because on the last read, I don't always get a full buffer (I read and write in 8K chunks), and so it goes faster, and sometimes the protocol starts right up again. Is there something funny in the timing out of uucico that I don't know about. ..... And ..... 2. The UNIX system I talk to responds with a "Ptg" in the uucico startup handshake. Does anyone know anything about the "t" protocol. As it is, my uucico (on the PC) is under 10K (about 20K when you add the BSS), so I have all sorts of space to add another protocol. Thanks in advance, -- dg@lakart.UUCP - David Goodenough +---+ IHS | +-+-+ ....... !harvard!xait!lakart!dg +-+-+ | AKA: dg%lakart.uucp@xait.xerox.com +---+
mboen@nixpbe.UUCP (Martin Boening) (06/16/89)
dg@lakart.UUCP (David Goodenough) writes: >A couple of questions: >2. The UNIX system I talk to responds with a > "Ptg" >in the uucico startup handshake. Does anyone know anything about the >"t" protocol. As it is, my uucico (on the PC) is under 10K (about 20K >when you add the BSS), so I have all sorts of space to add another >protocol. I don't know about the t-protocol that system means BUT in BSD uucp there's a t protocol for TCP/IP connections. In README.TCP for BSD uucp it says that the t protocol was added because the g protocol maxes out at 9000 Baud re- gardless of the physical medium. It also says that the t protocol is essen- tially the g protocol without checksumming and packetizing since it assumes an error free channel (TCP/IP, you know). It also doesn't do ioctl's on the sockets used. I hope this helps and doesn't violate any copyrights. Martin -- Email: in the USA -> ...!uunet!philabs!linus!nixbur!mboening.pad outside USA -> {...!mcvax}!unido!nixpbe!mboening.pad Paper Mail: Martin Boening, Nixdorf Computer AG, DS2, Pontanusstr. 55, 4790 Paderborn, W.-Germany