On Fri, Apr 27, 2001 at 08:32:21AM -0400, Timothy Kimball ([EMAIL PROTECTED]) spew-ed
forth:
>
> grep() will do it easily:
>
> @lines = grep { ! /\.txt$|\.scc$/ } @lines;
>
> or do it when you read the file:
>
> @lines = grep { ! /\.txt$|\.scc$/ } <INFO>;
You are making the RE engine do a log of work there.
[root@fluffhead /root]# perl -Mre=debug -wle '$_=qq{foo.txt}; /\.(txt|foo)$/';
Compiling REx `\.(txt|foo)$'
size 14 first at 1
1: EXACT <.>(3)
3: OPEN1(5)
5: BRANCH(8)
6: EXACT <txt>(11)
8: BRANCH(11)
9: EXACT <foo>(11)
11: CLOSE1(13)
13: EOL(14)
14: END(0)
anchored `.' at 0 (checking anchored) minlen 4
Guessing start of match, REx `\.(txt|foo)$' against `foo.txt'...
Found anchored substr `.' at offset 3...
Guessed: match at offset 3
Matching REx `\.(txt|foo)$' against `.txt'
Setting an EVAL scope, savestack=3
3 <foo> <.txt> | 1: EXACT <.>
4 <foo.> <txt> | 3: OPEN1
4 <foo.> <txt> | 5: BRANCH
Setting an EVAL scope, savestack=3
4 <foo.> <txt> | 6: EXACT <txt>
7 <foo.txt> <> | 11: CLOSE1
7 <foo.txt> <> | 13: EOL
7 <foo.txt> <> | 14: END
Match successful!
Freeing REx: `\.(txt|foo)$'
versus:
[root@fluffhead /root]# perl -Mre=debug -wle '$_=qq{foo.txt}; /\.txt$|\.foo$/'
Compiling REx `\.txt$|\.foo$'
size 9
1: BRANCH(5)
2: EXACT <.txt>(4)
4: EOL(9)
5: BRANCH(9)
6: EXACT <.foo>(8)
8: EOL(9)
9: END(0)
minlen 4
Matching REx `\.txt$|\.foo$' against `foo.txt'
Setting an EVAL scope, savestack=3
0 <> <foo.txt> | 1: BRANCH
Setting an EVAL scope, savestack=3
0 <> <foo.txt> | 2: EXACT <.txt>
failed...
0 <> <foo.txt> | 6: EXACT <.foo>
failed...
failed...
Setting an EVAL scope, savestack=3
1 <f> <oo.txt> | 1: BRANCH
Setting an EVAL scope, savestack=3
1 <f> <oo.txt> | 2: EXACT <.txt>
failed...
1 <f> <oo.txt> | 6: EXACT <.foo>
failed...
failed...
Setting an EVAL scope, savestack=3
2 <fo> <o.txt> | 1: BRANCH
Setting an EVAL scope, savestack=3
2 <fo> <o.txt> | 2: EXACT <.txt>
failed...
2 <fo> <o.txt> | 6: EXACT <.foo>
failed...
failed...
Setting an EVAL scope, savestack=3
3 <foo> <.txt> | 1: BRANCH
Setting an EVAL scope, savestack=3
3 <foo> <.txt> | 2: EXACT <.txt>
7 <foo.txt> <> | 4: EOL
7 <foo.txt> <> | 9: END
Match successful!
Freeing REx: `\.txt$|\.foo$'
When you look at /\.txt$|\.foo$/ you see what the RE engine is doing (much
easier to see if you use -Mre=debugcolor). It tries
to match in portions; foo.txt -> oo.txt -> o.txt -> .txt -> (if all failed)
foo.foo -> oo.foo -> o.foo -> .foo
The first one immediately sees the anchor, and tries to match less patterns;
txt -> (if failed) foo
A really good book to read is Mastering Regular Expressions, if you are new to
regular expressions (of course, read perlre), and a great article on how the RE
engine works is at http://perl.plover.com/Regex/article.html
Cheers,
Kevin
--
[Writing CGI Applications with Perl - http://perlcgi-book.com]
It would be easier to pay off the national debt overnight than to neutralize
the long-range effects of OUR NATIONAL STUPIDITY.
-- Frank Zappa
