raylau@dasys1.UUCP (Raymond Lau) (05/18/88)
In response to all the complaints about -39 unexpected EOF errors.... I'll admit. The situation confounded me. But from what Chuq mentioned, I see now the cause of all problems. Everyone complained that 1.31 didn't skip the news headers. The suggestion offered was that the line: (This file... should be checked for as it always appeared at the beginning. It was also noted that a : never starts a line in the news/mail header... Instead of checking for (This file..., what 1.40A does is check for a colon right after a CR or a LF. It will start decoding at that point. This was intended to make it easier to decode postings as you need not strip the header. Thus, this bug is really a feature. (Yea yea - that's the usual joke about bugs....but what am I to say?) It appears that everyone likes overstripping the headers. Hey, if everyone says that it's OK to check for (This file..., it should be even more permissible to check for a CR or LF! Which is why I opted for that route. -Ray (raylau@dasys1.UUCP)
dtw@f.gp.cs.cmu.edu (Duane Williams) (05/19/88)
In <4481@dasys1.UUCP>, Raymond Lau writes: | It was also noted that a : never starts a line in the news/mail header... | | Instead of checking for (This file..., what 1.40A does is check for a colon | right after a CR or a LF. It will start decoding at that point. You may complain that some people are overzealous in stripping the header, but BinHex4 can handle such files just fine. Why not look for the first line that starts with a ":"? Duane Williams -- uucp: ...!seismo!cmucspt!me.ri.cmu.edu!dtw arpa: dtw@cs.cmu.edu
geb@cadre.dsl.PITTSBURGH.EDU (Gordon E. Banks) (05/19/88)
In article <1714@pt.cs.cmu.edu> dtw@f.gp.cs.cmu.edu (Duane Williams) writes: > >You may complain that some people are overzealous in stripping the >header, but BinHex4 can handle such files just fine. > I have not found this to be the case. Binhex 4.0 often chokes when the header is too long, especially when there is a lot of explanatory text preceding the colon. This should be a trivial fix, but I don't have binhex 4.0 source code, so I am forced to strip headers. The error message given is usually "CRC error".
lih@cunixc.columbia.edu (Andrew Lih) (05/26/88)
In article <1714@pt.cs.cmu.edu> dtw@f.gp.cs.cmu.edu (Duane Williams) writes: >In <4481@dasys1.UUCP>, Raymond Lau writes: >| It was also noted that a : never starts a line in the news/mail header... >| >| Instead of checking for (This file..., what 1.40A does is check for a colon >| right after a CR or a LF. It will start decoding at that point. > >You may complain that some people are overzealous in stripping the >header, but BinHex4 can handle such files just fine. However, it gags on files with big headers like files you grab off of USENET. >Why not look for the first line that starts with a ":"? If you will remember, there was a big flurry over the fact that Ray put: (This file must be converted with BinHex or Stuffit) rather than the standard: (This file must be converted with BinHex) this was because several other programs that did dehexification of BinHex 4.0 files (UNIX, I believe) were expecting to see this EXACT header when attempting to de-hexify. If the net community demanded that this header be there in this exact form, then Ray should only expect that you would do the same for HIS program. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= """"""" Andrew "Fuz" Lih lih@cunixc.cc.columbia.edu | @ @ | Computer Consultant UI.LIH@CU20B.COLUMBIA.EDU < ^ > \ - / Columbia University Center for --- Computing Activities ...rutgers!columbia!cunixc!lih =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=