gensub RE problem?

[Jim Hart]

GNU Awk 3.1.0, compiled from source on Darwin.

I believe I've found a problem in gensub's handling of regular 
expression matching. The string to be matched against  is:

<div class="bodytext">                                   <a href="/2
/hi/science/nature/2212629.stm"><img height="120" hspace="5" vspace="0" 
width="1
00" border="0" src="/media/images/38213000/jpg/_38213850_websmall.jpg" 
align="le
ft"></a>                                                                
<a href=
"/2/hi/science/nature/2212629.stm"><span class="h1">Global body needed 
to fight
poverty</span></a><br>                  A new global body for economic 
developme
nt is needed, says The Lancet in the lead-up to the Sustainable 
Development Summ
it.<br clear="ALL">     </div>

The gensub command:

gensub(/.*<br>([^<]*)</br.*|.*/,"\\1",1,itemString)

returns:

                      A new global body for economic developme
nt is needed, says The Lancet in the lead-up to the Sustainable 
Development Summ
it.

as one would expect. Whereas:

gensub(/.*<br>(.*)</br.*|.*/,"\\1",1,itemString)

returns null. And:

gensub(/.*<br>(  *)([^<]*)</br.*|.*/,"\\1",1,itemString)

returns only one space, not the many that follow the <br>. And:

gensub(/.*<br>(  *)([^<]*)</br.*|.*/,"\\2",1,itemString)

returns the same thing as the first example, with all the leading spaces 
included.


The man page re_format says:

"In  the  event  that  an RE could match more than one sub-
       string of a given string, the RE matches the one  starting
       earliest  in  the string.  If the RE could match more than
       one substring starting  at  that  point,  it  matches  the
       longest.   Subexpressions  also match the longest possible
       substrings, subject to the constraint that the whole match
       be  as long as possible, with subexpressions starting ear-
       lier in the RE taking priority over ones  starting  later."

Note that last part. The earlier subexpression should be extending to 
maximum length. Yet, it appears that gensub is returning the shortest 
possible match for (  *) and (.*), not the longest.

Comments? Opinions? Did I miss something? Is gawk just calling OS 
routines so the problem is actually in Darwin?