Preserved Topic: regexp fun with sed!

for file in `grep -l \(\# *`
do sed -e "s/(\(#[^\)]*\))/\<a href=\"?profiles\&do_view=1\&num_profile=\\1\"\>\(#\\1\)\<\/a\>/g" $file > $file.tmp;
done

that.. should change something like (#454) to <a href="?profiles&do_view=1&num_profile=454">(#454)</a>
instead it changes it to <a href="?profiles&do_view=1&num_profile=#454">(##454)</a>
any suggestions?

"s/(\(#[^\)]*\))/\<a href=\"?profiles\&do_view=1\&num_profile=\\1\"\>\(#\\1\)\<\/a\>/g"

You grouped the # sign with the number when you put \(#[^)]*\). Take the # out of the group and the expression should work.

use this regex:

s/(#\([^)]*\))/<a href=\"?profiles/\&do_view=1\&num_profile=\\1\">(#\\1)</a>/g

You have a lot of extrat back slants (\) in the string that I don't think are necessary. They probably don't hurt any thing but the do make the line hard to read. I dropped them out. If you need them to make the line work, put them back, but I would try the line without all the extra back slants first. The quotes (") you have around the expression should make most of them unnecessary.


-- not necessarily stoned... just beautiful.





[This message has been edited by hyperbole (edited 02-27-2004).]