KYY%NIHCUDEC.BITNET@cunyvm.cuny.edu (M. DOBSON) (10/20/88)
In a previous message I mentioned a minor problem I am having upgrading to 1.3. That is nothing compared to the major problems BITNET is causing for me. It seems virtually everything but uuencoded stuff is munged in ways ranging from annoying to disasterous. The problems begin right with the shell archives. BITNET just doesn't seem to be able to handle the control characters that are in the sed/gres command strings, somehow a ctrl-E is being inserted. If unchanged, when this archive is fed to sh, all the leading X's are left on the lines :-( The only way I've been able to handle this is to pass the files through edlin under DOS and to a global replace of the ctrl-EX string with X. Mined won't read the files because they contain non-ascii characters. Once I've unshar'd them, I still have problems because all the tabs are converted to spaces. This causes the character counts to be wrong so I never know if the files have been damaged in other ways. This conversion of tabs to spaces also causes even the latest verion of patch to fail regularly when more than one tab on the origianl line is involved. It seems patch just can't deal with that large of a white space difference forcing me to hand apply many of the cdifs. This white space problem also means I will never be able to tell if my patched files match ast's since the sizes and crcs are all going to be off. Any suggestions for dealing with this (other than getting on uesnet :-)) would be appreciated. Mike Dobson (kyy@nihcudec.BITNET)
ncoverby@ndsuvax.UUCP (Glen Overby) (10/20/88)
In article <4951@louie.udel.EDU> KYY%NIHCUDEC.BITNET@cunyvm.cuny.edu (M. DOBSON) writes: > [Wonderful Bitnet horror story deleted] Bitnet sure has a reputation. If you're not doing something IBMish on there, it's an up-hill battle. I've always had major problems with sources sent over Bitnet have always been a problem for me. There is no coherent translation between ASCII and EBCDIC, nor is there apparently any coherency within the EBCDIC world itself! Back when the 1.1-1.2 stuff was released, I convinced our VM system programmer here at NDSU to give me some space on his LISTSERV so I could make Minix sources available there for archival purposes. One thing I have tried to do with that archive is make all the sources usable by people on bitnet; for example, the tabs are NOT expanded and will hopefully get to you that way. I suggest getting the updates from here and see if they work out better for you (info on this server is in the Minix Info Sheet). I wouldn't mind hearing from a few people about how this is actually working out. I don't know where your ^E's came from, but you can probably solve your patch spaces problem by running patch with the "-l" option (maybe you are; it wasn't clear to me). >Any suggestions for dealing with this (other than getting on uesnet :-)) >would be appreciated. Get on Usenet :-) Or, more generally, get on a Non-Bitnet machine (the internet would work fine). There are a lot of public-access Unix systems around which will let you on for little or no charge (aside from calling into them). If you consider your time worth money, it might be cheaper to wait and get the upgrade package from Prentice-Hall whenever it comes out. To steal a line from John Gilmore, Good Luck, and may the source be with you! -- Glen Overby ncoverby@plains.nodak.edu uunet!ndsuvax!ncoverby ncoverby@ndsuvax (Bitnet)
ast@cs.vu.nl (Andy Tanenbaum) (10/21/88)
In article <4951@louie.udel.EDU> KYY%NIHCUDEC.BITNET@cunyvm.cuny.edu (M. DOBSON) writes: >In a previous message I mentioned a minor problem I am having upgrading to >1.3. That is nothing compared to the major problems BITNET is causing for >me. I can assure you, you're not alone. Even I have great difficulty with BITNET, and I don't even use it. My problem is that lots of people are in your situation and they are constantly asking for help. The only thing I can think of offhand, is maybe somebody, probably a BITNET user, could write a program that combined the functions of compress and uuencode in one simple, easy to use program. If it were possible to run a file through a simple filter and get a resulting file that was smaller than the original and used only characters that BITNET accepts, this might be acceptable, although there is very much the issue of how much nuisance USENET people are willing to put up with for the benefit of BITNET people. If this filter reduced text size, by first doing some sort of compression, we could save some bandwidth, which saves everyone money. I suspect that the character set BITNET allows is larger than that used by uuencode, so a cleverly designed program should be able to avoid most of the expansion produced by uuencode. Andy Tanenbaum (ast@cs.vu.nl)