Path: utzoo!attcan!uunet!van-bc!rthurlow From: rthurlow@van-bc.UUCP (Rob Thurlow) Newsgroups: comp.sys.atari.st Subject: Binaries news group postings Message-ID: <1815@van-bc.UUCP> Date: 19 Jun 88 00:50:25 GMT Reply-To: rthurlow@van-bc.UUCP (Rob Thurlow) Organization: Wimsey Associates, Vancouver, BC. Lines: 70 Well, time for one of my infrequent postings again. I just had a look at Jwahar Bammi's ZMDM binaries posted to the comp.binaries.st group, and I'm afraid it didn't make it here in very good shape. The sequence of events for this stuff is as follows: 1) The binaries came into the UUCP node I connect to, and got batched to me as if I were just another UUCP site. The articles get put in files not much larger than 100K as a rule, with all newsgroups mixed randomly and the sequence within a group not guaranteed. 2) I stuck my UUPC boot disk into my machine one day and let all the stuff come in that wanted to. Because of some (timing?) problems with the current UUPC software, I had to restart the transmission several times. This goes on for quite some time, and the days files took up a sizable part of an 800K disk, say 450K. That's quite a bit at 1200 baud with uucp g-protocol. 3) I file the disk for a few days until I get time to work on the stuff. In the interim, the binaries of course expire on the Unix machine. 4) When I finally have some spare time, I sit down with Gulam and start being an article processor. Anything interesting gets saved to RAM disk with Gulam's cut and paste. I find that the binaries are in parts too large to do easily, since Gulam can only handle 32K in its paste buffer. But it gets done. 5) I run out of RAM disk space twice while decoding, but eventually the decode works fine. Then I try to get ARC to tell me the contents of the package, and the program shows me lots of funny numbers and characters after it tells me the archive is bad. So now I have no way to fix the thing. There was no sequence checking for any of the parts. The uuencoded binary was just chopped up by hand into 40K pieces. There is no way to detect a missing line so that I can ask for just one piece, and there is no hope of getting them all again from my Unix connection. I can try cut-and-paste again, but without the I-didn't-think-they-were-that-useful multi-part lines, I don't think it will work any better. We're pretty remote from the rest of the world up here in Vancouver in terms of net topology, and it is more of a pain than usual for me to post to the net, so I'm really reluctant to ask people to send replacements by mail. So I lose. I pay for connect time, too; usually my luck is *much* better. My plea: if you are going to post to comp.binaries.st, PLEASE get hold of Dumas' multi-part uuencoder, and let it chop your postings into nice tame ~30K pieces that will go through most mailers, and let it do the sequence checking and add the chaining lines so that people like me have a better chance at uudecoding correctly when it gets here properly and have a better idea of what to fix or when to give up when the group loses or mangles things. I think it is a little too much to ask the moderator to rearrange it, but I'm sure he would make happy noises if you asked him for the sources/binaries to the Dumas uu?code. And if that doesn't work, I think I can get it to you, too. The source was supposed to run great on Unix or the ST. BTW: Jwahar, I *liked* the earlier ZMDM a lot, and used it quite regularly before I went over to getting my stuff over UUCP. Thanks from all of us for the posting, even if it didn't quite make it for me. And thanks to Howard Chu for a new ARC: I can't beleive what a handicap it was before to have an ARC that bombed when its output was redirected. -- ------------------------------------------------------------------------ "There was something fishy about the butler. I think he was a Pisces, | probably working for scale." - Nick Danger | uunet-----\ | Robert Thurlow !van-bc!rthurlow | ubc-cs----/ | ------------------------------------------------------------------------