On Tue, May 22, 2007, Jörn Zaefferer wrote:
> Dan G. Switzer, II wrote:
> >> This is a little off-topic, but when doing a regex search and replace
> >> within a text editor, how can I replace one character within a
> >> specific pattern?
> >>
> >> I want to get rid of newlines within <td> tags. This finds them:
> >> <td>[^<]+(\r\n).+</td>
> >>
> >> How do I specify that I only want to replace the matched set?
> >>
> >
> > You group all the contents and then the replacement string are all the
> > matched sets pieced back together:
> >
> > sHtml.replace(/(<td>[^<]+)(\r\n)(.+</td>)/gi, "$1$3")
> >
> If I got that right, you could even mark the second group to be skipped by
> adding a colon:
>
> sHtml.replace(/(<td>[^<]+)(:\r\n)(.+</td>)/gi, "$1$2")
The syntax requires a question mark: (?:...)
> Or just skip the parentheses?
>
> sHtml.replace(/(<td>[^<]+)\r\n(.+</td>)/gi, "$1$2")
Yes, but this IMHO is still too weak because...
1. the ".+" in this regex is greedy and matches too much and this
way you would only remove newlines from every _second_ <td>...</td>
construct. So one has to use at least .+? to fix this.
2. Additionally, I recommend to use \r?\n to support both the Windows
CR-LF and Unix LF-only field.
3. The [^<]+ I do not understand as it would NOT allow to remove the
newlines when there is additional markup in the <td> container as in
"<td>...\n...<span>...</span>...</td>". I recommend to replace it
with just ".*?".
4. The "+" qualifier should be actually "*" as it might be fully valid
to have a "<td>\r\n</td>" container ;-)
5. The </td> has to be written escaped as in <\/td> within the regex
construct.
6. As the "." regex character in JavaScript does NOT match newline
character one has to use "(.|\r?\n)*".
So, I recommend the following stronger version:
sHtml.replace(/(<td>.*?)\r?\n((?:.|\r?\n)*?<\/td>)/gi, "$1$2")
But even this still has the problem that it is unable to remove MULTIPLE
occurences of newlines in the SAME <td> container. If this should be
also allowed one has to trick a little bit more:
sHtml = sHtml.replace(
/(<td>)(.*\r?\n(?:.|\r?\n)*)(<\/td>)/gi,
function ($0, $1, $2, $3) {
return $1 + $2.replace(/\r?\n/g, "") + $3;
}
);
This now should be a strong enough version and finally
do what was requested...
Ralf S. Engelschall
[EMAIL PROTECTED]
www.engelschall.com
