Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!rutgers!iuvax!bobmon
From: bobmon@iuvax.cs.indiana.edu (RAMontante)
Newsgroups: comp.binaries.ibm.pc.d
Subject: Re: Combine script (was Re: More silence in comp.binaries.ibm.pc)
Message-ID: <24211@iuvax.cs.indiana.edu>
Date: 3 Aug 89 14:56:03 GMT
References: <8472@bsu-cs.bsu.edu> <2487@astroatc.UUCP> <26918@srcsip.UUCP> <5154@umd5.umd.edu>
Reply-To: bobmon@iuvax.cs.indiana.edu (RAMontante)
Organization: malkaryotic
Lines: 23

dschuetz@umd5.umd.edu (David Schuetz) <5154@umd5.umd.edu> :
-
-As long as we're talking about combine, could somebody please tell me what
-the "sed '/^END/,/^BEGIN/d'" (or whatever) does?  I figured that it 
-deleted everything not within the BEGIN/END block of the file, but in 

Almost.  Precisely, it finds sets of lines starting with a line that
begins with `END' ("^" means "beginning of line"), and ending with a
line that begins with `BEGIN', and deletes those sets of lines.  When a
set of c.b.i.p. postings are concatenated, each one ends with an END
line, then there's some junk including headers in the next file, then
the next file's body starts with a BEGIN line.  This "internal" junk
is deleted.

However, it won't touch the stuff before the first BEGIN line (in the
first file), because there is no initial END line --- but uudecode
ignores header lines up to its own `begin 666 filename' line anyway.

Similarly, it won't scrape off garbage following the (last) END line,
but uudecode ignores trailing garbage after its own `end' line as well.

So the `sed ...' command only has an effect when multiple files are
concatenated together.