RALPH@UHHEPG.BITNET (12/12/87)
Date: 12-DEC-1987 00:43:27.18 From: Ralph Becker-Szendy RALPH AT UHHEPG To: GATEWAY::"info-cpm@simtel20.arpa",RALPH Subj: Problems with the Pascal UUDECODER Hi I have a little problem: I suspect the standard cp/m UUDECODEr had a bug. Here is the story: When i request a binary file (LBR, squeezed or suchlike) from the BITNET server at RPI i have to request it as UUENCODED (because our mail handler munches every binary file), and even then all the (up-arrow) are converted into ^(twiddle). So, i transfer the UUENCODED file to the cp/m micro, and run it through the regular UUDECODER (made from PD:<CPM.STARTER-KIT>UUDECODE.HEX or .PAS, from SIMTEL20 or the server at RPI). Unfortunately, the UUDECODER doesn't like the file: he complains about the second to last line being too short (in 80% of the cases). Here an example: M56<')',8<%.%1QT% "X$]#0&\C$#$#0#D#296#4&0'D5%%64$-?T#D#(#-/M MVP@!H) "(0 V+!H:&AH:&AH:&AH:&AH:&AH:&AH:&AH:&AH:&AH:&AH:&AH:M 3&AH:&AH:&AH:&AH:&AH:&AH:&@M <----- this line is (supposedly) too short M end And, even worse, in the rare case that UUDECODE doesn't complain about such a line being too short, sometimes the output file is wrong (like several characters missing). But, miracle, some files get through without any problem. On the other hand, i am sure that the uuencoded files did NOT get munched on the way to the micro (test: transmit it back, and compare). I have tested this with more than 30 files, and after a while it is no fun any more. So, the question is: does this UUDECODEr have a known bug, am i making a mistake, or what ? Maybe the UUENCODER used by the server is wrong (but that sounds VERY improbable). Any help is appreciated. As usual: if i get enough answers, i'll summarize then to the net. Ralph Becker-Szendy RALPH@UHHEPG.BITNET University of Hawaii / High Energy Physics Group (808)948-7391 Watanabe Hall #203, 2505 Correa Road, Honolulu, HI 96822 "Hawaii - it's not just for tourists. People actually live and work there."
FISHER@CICGE.RPI.EDU ("John S. Fisher") (12/14/87)
The Pascal UUDECODE appears to be in complete agreement with all of the other UUDECODE programs I've been able to compare it to. The problem is being generated by the uuencoder used by the Bitnet file server. Apparently, the (correct) uuencoders always construct encoded data that is a multiple of 4 bytes (which becomes 3 bytes when decoded). The encoder used by my server encodes only the exact number needed. Some, if not all, of the standard decoders depend on finding a multiple of 4 bytes, thus the error. At any rate, I'll repair my uuencoder as soon as possible. However, if you take any of the files you have that are a problem and replace the trailing "M" in the problem line with " M" (3 spaces and "M") you should be able to circumvent the error. Regards, JSFisher