> ... the best way to rip a segment of text from a text file...

Interesting problem. Nobody seems to be answering yet, so 
here is my two cents: I don't think you can do it in sed, 
since it always does things a line at a time. This might 
be true for awk as well. If you don't mind getting the rest 
of the line, you can do one of these:

sed -n '/foo/,/bar/p' myfile

awk '/foo/,/bar/' myfile

I thought of a nasty way to do it involving awk and the index 
command, cut, and some others commands, but why not just 
write a short perl script? Perl seems the natural tool for 
the job, IMO, especially since you want to only grab part 
of each line. Then name the script something like "2string" 
and put it in your path. :)

#!/usr/bin/perl

my ($file,$string1,$string2) = @ARGV;

die "Usage: $0 <file> <string1> <string2>\n"
  unless defined $file and defined $string1 and defined $string2;
open(F,$file) or die "Could not open $file: $!\n";
local $/;
<F> =~ m/($string1.+$string2)/s and print "$1\n";
close(F);










Yours:
perl -e '$/="";print$&if<>=~/START.*END/s;' < file

Very nice.  I like the use of $& ... I never use that.  I would have
done a () around the match, which would have cost me several strokes.

AFAICT, your if isn't needed, which allows me to shave off one stroke:
perl -e '$/="";<>=~/START.*END/s;print$&;' < file

Usage of -0 to set $/ drops another 5:
perl -0e '<>=~/START.*END/s;print$&;' < file

Since we're using command-line args, we can lose 5 more:
perl -n0e '/START.*END/s;print$&;' file




perl -e '$/="";print$&if<>=~/START.*END/s;' < file
Non-inclusive:

perl -e '$/="";<>=~/START(.*)END/s&&print$1;' < file





perl -ne 'if(/START/){$g=1;next;}if(/END/){$g=0;next;}print if($g);' file
spreads out as:
while (<>) {
    if (/START/) {  # if the line has the start marker
        $g = 1;     # we'll start printing
        next;       # but skip this line
    }               #  (not sure if START is on a line by itself
    if (/END/) {    # once we hit the end marker
        $g = 0;     # stop printing
        next;       # but skip this line, as before
    }
    if ($g) {       # if we're supposed to be printing...
        print;      # do so
    }
}



> The second returns only a single record...

perl -e 'for(<>){$a .= $_;};$a=~s/^.*START//s;$a=~s/END.*//s;print $a;' < file
becomes:
for (<>) {
    $a .= $_;        # build a string of the whole file
}
$a =~ s/^.*START//s; # get rid of everything before START, inclusive
$a =~ s/END.*//s;    # get rid of everything after END, inclusive
print $a;            # print what's left

This is a much more "traditional" perl approach...you could also do
something with s/.*START(.*)END.*/$1/;

Both work fine here, although the first one assumes START and END are
on lines by themselves.  Without seeing the exact application, it's
hard to get a perfect solution.





One line...barely (56 strokes):
perl -ne 'if(/START/){$g=1;next;}if(/END/){$g=0;next;}print if($g);' file

Also, using regexes (59 strokes):
perl -e 'for(<>){$a .= $_;};$a=~s/^.*START//s;$a=~s/END.*//s;print $a;' < file