while(<>) {
    tr/\r//d;  # get rid of the CR
    print if /^\@QUERY/ .. /^\@ENDQUERY/;
}
[download]

The flipflop operator, .. in scalar context, is a particular one which has built in memory of its previous state. Its initial state is false. So it'll return true from the first time the expression on its left hand side returns true, and every time you'll test it after that, it'll still remain true, up to and including the first time the expression on its right hand side becomes true, since this start. After that, it'll return false again... and you can start again from the top.

If you really want to do it your way, try

while(<>) {
    tr/\r//d;
    print if /^\@QUERY/ .. eof;
}
[download]

@lines = grep { $_ !~ /^[ \t]*\n?\r?$/ } @lines
[download]

I am guessing you meant

@lines = grep { $_ !~ /^\s*\n?\r?$/ } @lines
[download]

instead of

@lines = grep { $_ != /^\s*\n?\r?$/ } @lines
[download]

(i.e., s/!=/!~/).

Also as a nitpick, \s is an abbreviation to the character class [\ \t\r\n\f] (reference: perlreftut)so the \n and the \r are never matched as the * is greedy (i.e. they are unnecessary).

-enlil

You did say this was in a string (scalar)?

print $string;
blah



@QUERY
meaningful data
meaningful data
... etc...
@ENDQUERY

@QUERY2
more data
more data
... etc ...
@ENDQUERY2

$string =~ s[(?:\r?\n\s*)+(\@QUERY)][\n$1]g;

print $string;
blah
@QUERY
meaningful data
meaningful data
... etc...
@ENDQUERY
@QUERY2
more data
more data
... etc ...
@ENDQUERY2
[download]

Yes, it's a string, so I ought to be able to do what you suggest... however that regexp doesn't seem to match my string. I threw together the following to test it:

$string = "
blah



@QUERY
meaningful data
meaningful data
... etc...
@ENDQUERY

@QUERY2
more data
more data
... etc ...
@ENDQUERY2";

print "string1: $string\n\n";

$string =~ s[(?:\r?\n\s*)+(\@QUERY)][\n$1]g;

print "string2: $string\n\n";'
[download]

And my output is something like this. Can you help me tweak the regexp?

string1:
blah




meaningful data
meaningful data
... etc...



more data
more data
... etc ...


string2:
blah




meaningful data
meaningful data
... etc...



more data
more data
... etc ...
[download]

Thanks,

AH

P.S. bart I'm looking at ".." and it seems very powerful... I may end up using it, once get my head around it.

----------

Using perl 5.6.1 unless otherwise noted.
Apache 1.3.27 unless otherwise noted.
Redhat 7.1 unless otherwise noted.
[download]

If you turned "use strict" on, you'd see the problem.

string = "
blah



@QUERY
meaningful data
meaningful data
... etc...
@ENDQUERY

@QUERY2
more data
more data
... etc ...
@ENDQUERY2";

Possible unintended interpolation of @QUERY in string at (eval 1) line
+ 1, <> line 16.
Possible unintended interpolation of @ENDQUERY in string at (eval 1) l
+ine 1, <> line 16.
Possible unintended interpolation of @QUERY2 in string at (eval 1) lin
+e 1, <> line 16.
Possible unintended interpolation of @ENDQUERY2 in string at (eval 1) 
+line 1, <> line 16.
[download]

Replace the "s with 's when you initialise $string and you will see the correct result.

Also, as Enlil points out elsewhere in the thread, my regex was more complex than necessary. It can be replaced with s[\s+(\n\@QUERY)][$1]g;.

---

Examine what is said, not who speaks.

"Efficiency is intelligent laziness." -David Dunham
"Think for yourself!" - Abigail

$string =~ s/\A\s+@QUERY/@QUERY/
[download]

$string =
  substr($string, index($string, '@QUERY'));
[download]