ragg0270@uxa.cso.uiuc.edu (Richard Alan Gerber) (03/01/90)
I have read a lot about DNET here, but I don't really have a clear picture of what I does/can do. I looked at the .doc files supplied with DNET, but they seem to assume you already know what you want to do. So my questions are these: what exactly is DNET and how is it useful? do I have to write my own bits of code to use it, (ie. do I have to have a C compiler? I have a direct line into a Sun that I can hook my Amiga up to and would like to know if DNET has a use in this situation. thanks to all, Richard Gerber reply via mail to: ragg0270@uxa.cso.uiuc.edu
cs121jj@ux1.cso.uiuc.edu (03/02/90)
I cannot get DNET to work on my A2000 either (I think both use the same port). I can get about 2 characters across the line, then DNET locks up. The rest of my Amiga functions, just DNET is gone. I also (to get rid of the 7-wire connect message at DNET's startup) must use Xon/Xoff or NO handshaking from Preferences Serial menu. RTS/CTS gives me a "7-bit" connect. My DNET startup command is: run dnet -8 -X -P0 -X0 -b19200 -m0 DNET looks neat, but I'm not sure how to get it to run. Any suggestions? (I thought the -8 would force an 8-bit connection)...
davewt@NCoast.ORG (David Wright) (03/02/90)
I have had trouble with DNET too. I was running it on an 8-bit serial line through a 2400 baud modem, and after just a few minutes, it always seems to lock up. I was also using the NET: device. ALso, I once read an article, and it claimed the DNET was about a fast as Y or Z modem when it was only sending in one direction. I found it to be MUCH slower at sending directories to the other system than should have occured if it was as fast as Y modem. I am pretty sure that it was DNET that blew it, and not the NET: device, and attempts to run the NetStat utility also locked, and the modem was no longer polling.... Any ideas? Dave
ccplumb@lion.waterloo.edu (Colin Plumb) (03/07/90)
In article <15600002@ux1.cso.uiuc.edu> cs121jj@ux1.cso.uiuc.edu writes: >run dnet -8 -X -P0 -X0 -b19200 -m0 > >DNET looks neat, but I'm not sure how to get it to run. > >Any suggestions? > (I thought the -8 would force an 8-bit connection)... DNET opens an initial window to let you talk to the serial line until it finds a DNET handshake, when it closes it and starts the protocol. The -8 flag forces 8 bits (instead of 7 bits and whatever parity) on the initial window and has nothing to do with the -m0 flag which uses 8 bits for the protocol. The -P and -X flags don't do anything unless you're using the default -m1 seven-bit protocol. Assuming you're talking to a Unix machine, be sure to specify -m0 to the Unix end as well. Otherwise the Unix end doesn't put the terminal into raw mode and the first funny character (^Z, ^Y, ^C/^? whichever you use) that comes along will kill the server. I've added a local hack to clear all the tty structures so a noise character on the line won't ^Z my DNET driver or something annoying. -- -Colin
ecarroll@vax1.tcd.ie (03/08/90)
In article <1990Mar2.122254.4259@NCoast.ORG>, davewt@NCoast.ORG (David Wright) writes: > I have had trouble with DNET too. I was running it on an 8-bit > serial line through a 2400 baud modem, and after just a few minutes, it > always seems to lock up. I was also using the NET: device. ALso, I > once read an article, and it claimed the DNET was about a fast as Y or Z > modem when it was only sending in one direction. I found it to be MUCH > slower at sending directories to the other system than should have occured > if it was as fast as Y modem. I am pretty sure that it was DNET that blew > it, and not the NET: device, and attempts to run the NetStat utility also > locked, and the modem was no longer polling.... > Any ideas? > > Dave The delay when you are doing remote directories with NET: is because a full 512 byte fileheader is being transmitted for each file in the directory. This means that for a directory with 100 files, at least 50K needs to be transmitted across the serial link (about 4 minutes at 2400 baud). Of course, this would be much faster if NET: simply asked for the filenames themselves and not the full fileheader, but then it wouldn't be as compatible with AmigaDOS. At 2400 baud, NET: is a lot of fun to play with but don't expect to use it for anything serious... (it's much quicker to open a remote CLI and type DIR for example :-) No ideas on your Dnet lockups I'm afraid. Eddy -- Eddy Carroll ----* Genuine MUD Wizard | "You haven't lived until INTER: ecarroll@vax1.tcd.ie | you've died in MUD!" UUCP: {..uunet}!mcvax!ukc!vax1.tcd.ie!ecarroll | -- Richard Bartle
walker@sas.UUCP (Doug Walker) (03/20/90)
In article <5814.25f5c5f8@vax1.tcd.ie> ecarroll@vax1.tcd.ie writes: >The delay when you are doing remote directories with NET: is because a full >512 byte fileheader is being transmitted for each file in the directory. >This means that for a directory with 100 files, at least 50K needs to be >transmitted across the serial link (about 4 minutes at 2400 baud). Of course, >this would be much faster if NET: simply asked for the filenames themselves >and not the full fileheader, but then it wouldn't be as compatible with >AmigaDOS. It's impossible for NET: to just get the filenames. NET: has no idea that the DIR command is the one asking for information about files; it just sees the request come over for a FileInfoBlock, and fills the request. However, the FIB is only 220 bytes, not 512 bytes. Add another 20 bytes or so of NET: overhead per packet and you get less than 250 bytes/file, or about 25k. Still 2 minutes, so it's not really usable at 2400 baud. As to DNET lockups, I have seen this happen as well, but I have no idea what causes it. I use a null modem hookup, and it seems to work without lockups at 9600 baud but with lockups at 19200, so I use 9600. ***** =*|_o_o|\\=====Doug Walker, Software Distiller====== BBS: (919)382-8265 = *|. o.| || NOTE NEW BBS NUMBER!^^^^^^^^ | o |// Got mole problems? Call Avogadro: 602-1023! ====== usenet: ...mcnc!rti!sas!walker plink: dwalker bix: djwalker