Excluding all-numeric fields

[Göran Uddeborg]

Hello,

I'm trying to exclude completely numeric fields from some pages I
watch.  It seems to work as expected for unsigned numbers, but fails
for negative numbers.  I've tried to investigate a bit, and can
reproduce the problem with webdiff.  But I don't really know what I do
wrong.

I use this ignore pattern:

    [HeltNumeriskt]
    ^-?[0-9.]+$

Then I run

    webdiff --ignore=HeltNumeriskt --out=\# --archive=a+science.old.html 
--current=a+science.html
    grep 66ccff \#

This matches tree negative numbers.  If I remove the --ignore flag, I
also get a fourth, positive, number.

I've put copies of the input files at
ftp://ftp.uddeborg.se/pub/webdiff  The original source is
http://market.millistream.com/alternativa/index.php/listor/details/6321

At first I suspected it had to do with my attempt to match the
beginning and end of the field.  I saw in the webdiff code that the
ignore pattern I give is embedded into some extra regexp code.  But
since this DOES match the pattern which is positive, that seems not to
be the case.

Is there anyone who can explain what I'm doing wrong?

This is on a Fedora system with websec 1.9.0 and perl 5.14.2.