petera@utcsri.UUCP (Smith) (03/04/86)
How many times have we seen uuencoded stuff get screwed up somewhere on the net. It has happened with every binary I have seen on the net so far. I think it is time that we overhaul uuencode/decode to do two things. First it should split large files into chunks small enough to allow transmission by all the mailers on the net. Second it should provide some checksum information to avoid the kind of things that have happend with PC-LISP. After writing PC-LISP I don't have time to work on this but I think the entire net would benefit from such a program. As a side thought perhaps the program could run a compression algorithm first before encoding. We would cut down the net traffic for binaries significantly this way. One checksum per block would be sufficient. Then when you tried to assemble the blocks and one was bad you could request that someone resend you that block. Eg. block: 1 of: 15 of: pc-lisp.exe Mxxxxxxxxxxdddddddddddddxxxxxxxxxxxxxxx..... Mxxxxxxxxxxdddddddddddddxxxxxxxxxxxxxxx..... Mxxxxxxxxxxdddddddddddddxxxxxxxxxxxxxxx..... M...... ........ M...... Mxxxxxxxxxxdddddddddddddxxxxxxxxxxxxxxx..... Mxxxxxxxxxxdddddd sum:10234 The encoding function sould split the entire binary into blocks and place them in files called say 'pc-lisp.000,pc-lisp.001 ....'. It would be pretty simple to call uuencode as a subroutine to do this. Computing a checksum should be very carefully done to aviod machine dependencies. Of course this is just what I would like to see, perhaps we could thrash out the general specs and then let someone go on it. What do you lot think? Peter Ashwood-Smith University of Toronto.
sgt@alice.UucP (Steve Tell) (03/07/86)
This has probably been brought up before, but why not just send source code? If the botch is small, it might be hand-fixable, and would be obvious to anyone lookig at the code. After all, this is net.sources, not net.binaries. :^)
campbell@maynard.UUCP (03/07/86)
I believe atob and btoa use checksums to prevent mungage. Why not use them? -- Larry Campbell The Boston Software Works, Inc. ARPA: maynard.UUCP:campbell@harvard.ARPA 120 Fulton Street UUCP: {harvard,cbosgd}!wjh12!maynard!campbell Boston MA 02109
phil@amdcad.UUCP (Phil Ngai) (03/08/86)
In article <2270@utcsri.UUCP> petera@utcsri.UUCP (Smith) writes: > > What do you lot think? I think you should use mod.sources. -- "We must welcome the future, remembering that soon it will become the present, and respect the past, knowing that once it was all that was humanly possible." Phil Ngai +1 408 749 5720 UUCP: {ucbvax,decwrl,ihnp4,allegra}!amdcad!phil ARPA: amdcad!phil@decwrl.dec.com
danz@hplsla.UUCP (danz) (03/10/86)
Hey, everybody, lets quit evading. This chap has a good point. My experience has been that the chance of downloading, decoding, and ending up with a working binary is abou 50%, which is disgusting. I suspect that some improvements to the encode/decode pair are in order, and the changes suggested sound reasonable. As for the argument that "this is net.sources not net.binaries" -- my experience has been that downloading, compiling and ending up with a working binary is about 50% also, unless one wants to spend hours in someone else's source code. Binaries serve a good purpose -- lets insure that they can be transferred confidently. Dan hplabs!hplsla!danz