in reply to Re^2: Defining Characters in Word Boundary?
in thread Defining Characters in Word Boundary?

is there a meaningful difference between (?![a-z]) and (?=[^a-z])? is the former recommended?

Yes, they're different regular expression patterns that match different things. (?![a-z]) asserts "not followed by any of the characters from 'a' through 'z', which includes not being followed by any character." (?=[^a-z]) asserts "followed by a single character that is not any of the characters from 'a' through 'z'." The former is a negative assertion; the latter is a positive assertion.

In your case, (?![a-z]) is what you would want to use.

[PerlMonks posting tip: Enclose Perl code in <code></code> tags, even code within paragraphs.]

UPDATE: Removed color.

Replies are listed 'Best First'.
Re^4: Defining Characters in Word Boundary?
by AnomalousMonk (Archbishop) on Jan 20, 2011 at 23:28 UTC
    In your case,  (?![a-z]) is what you would want to use.

    One behavioral difference between these regexes and, in the case of the OP, the reason iaw4 would (probably) want to use this regex is that it can match at the end of a string and thus emulates the behavior of the  \b assertion. (Note:  \b can also match at the start of a string.) 

    >perl -wMstrict -le
"my $str = 'abcd';
 for my $rx (qr{(?=[^a-z])}, qr{(?![a-z])}, qr{\b}) {
   my @offsets;
   push @offsets, $-[1] while $str =~ m{ ($rx) }xmsg;
   if (@offsets) {
     print qq{$rx matches '$str' at offset(s) @offsets};
     }
   else {
     print qq{$rx does not match '$str'};
     }
   }
"
(?-xism:(?=[^a-z])) does not match 'abcd'
(?-xism:(?![a-z])) matches 'abcd' at offset(s) 4
(?-xism:\b) matches 'abcd' at offset(s) 0 4

    [download]
[reply]
[d/l]
[select]

      (?<![a-zA-Z]) emulates \b at the beginning of an alphabetic string and (?![a-zA-Z]) emulates \b at the end of an alphabetic string.

      But as I explained before, it doesn't seem from iaw4's problem description that asserting a boundary match is even needed.

      UPDATE: Removed color.

[reply]
[d/l]
[select]
        Needed? Not likely. But conceptually, (?![a-zA-Z]) is simpler. With it, we're defining the macro without requiring the presence of some external context.
[reply]
[d/l]
Re^4: Defining Characters in Word Boundary?
by Anonymous Monk on Jan 20, 2011 at 18:08 UTC
[reply]

      I used color here very purposefully to set apart Perl code from its surrounding explanatory text. I did this for clarity. I didn't use color "casually."

      Please avoid using themes that don't "play well" with the colors in my PerlMonks posts. :)

      UPDATE: Removed color and rescinded my remarks.

[reply]
        I used color here very purposefully to set apart Perl code from its surrounding explanatory text. I did this for clarity. I didn't use color "casually."

        CODE tags already stand out from the surrounding text, its part of their purpose.

        What is casual about your usage is: font tags instead of span/div, no background color.

        Please avoid using themes that don't "play well" with the colors in my PerlMonks posts. :)

        Like Bones McCoy said Dammit, Jim you're not on the standards committee :)

        Themes don't assume a background and text color, they specify each explicitly. Your non-standard usage ignores a themes text and background color specification. When we customize css, we follow the built-in themes like RFCs.

        Signatures and Node Templates

        Also, keep in mind that many monks surf with custom CSS and non-standard color schemes. If you use explicit colors in your signature, it may look groovy-cool with your settings but be an obnoxious mess with theirs.

        Hard-coded colors breaking custom CSS settings.
[reply]

In Section Seekers of Perl Wisdom