Path: utzoo!attcan!uunet!cs.utexas.edu!tut.cis.ohio-state.edu!ucbvax!bloom-beacon!athena.mit.edu!watcher
From: watcher@athena.mit.edu (chris ross)
Newsgroups: comp.unix.questions
Subject: need help with sed "pattern space"
Message-ID: <12583@bloom-beacon.MIT.EDU>
Date: 11 Jul 89 05:19:10 GMT
Sender: daemon@bloom-beacon.MIT.EDU
Reply-To: watcher@athena.mit.edu (chris ross)
Distribution: na
Organization: Massachusetts Institute of Technology
Lines: 30

hi.  Got a few questions for you sed sharks...

sed claims that it "cyclically copies a line of input into a pattern space
(unless there is something left after a D command), applies in sequence all
commands whose addresses select that pattern space, and at the end of the
script copies the pattern space to the standard output (except under -n)
and deletes the pattern space"

1.  Given the above description, and reading from stdin, why do simple
scripts like "sed s/a/b/g" not produce output until the _second_ input
line has been entered?

2.  It seems to me that a script like "sed '/foo/,/bar/ d'" needs to retain
the pattern space for an arbitrary number of input lines.  Does the phrase
"start the next cycle" in the d command documentation mean "start the next
cycle NOW, without checking other addresses, doing output, or clearing the
pattern space?"  If not, why doesn't a p command after the d command print
the growing pattern space as many times as there are lines retained by
/foo/,/bar/ ?

3. Why doesn't "sed '/^$/,/^$/ D'" (note UPPERcase D) convert multiple
consecutive blank lines to a single blank line?

Obviously, I'm thinking about the pattern space in the wrong way.
Can anyone point me in the right direction?

thanx

_____________________________________
chris ross <0> watcher@athena.mit.edu