Thanks to both of you. I've tried adding and removing the g modifier, but never tried without the last operator.

The sample code was a simplification of my real problem, where I'm trying to capture one specific record from one kind of table from a set of html documents, where each field has it's own line in the source.

Doing that way, I had to split the original regex in two:

1. one to identify the required record by the value of the first field
2. another to the capture of the data fields

BTW, the original regex was something like this:

  my ($k, @f) = ($h =~ m!<td.*?>(required_\d+_\d+.txt)</td>\s+(<td.*?>
+(.+?)</td>\s+)+</tr>!m);
[download]

Then, the ",+" actually meant whitespace "\s+", but I wanted to make them visible in the output. ;-)