score definition doesn't match, bug?

Hi,

Have 0.9.9p1 and the following snippet in ~/News/Score:

#v+
[*]
Score:: 426 % carbon copy to me
X-Comment-To: Vadim Goncharov
To: Vadim Goncharov
#To: vadim_?nuclight@
#Cc: vadim_?nuclight@
To: nuclight@
Cc: nuclight@

Score:: 102 % hack for mail lists where previous carbon copy don't work
	References: nuclight
	In-Reply-To: nuclight
#v-

The intention is to high score of replies to me in newsgroups and gmane mail
lists. But it doesn't work: I have to add the second fragment which tries to
match my Message-Ids in References: header. But why this regexp doesn't work?
Is it my fault or bug?

Also, I have another snippet:

#v+
# dummy subjects
Score:: -99
%Expires: 
	{:
	  Subject: [Tt]est$
	  ~Subject: tester
	  ~Subject: testing
	  ~Subject: [Pp]atch
	}
	Subject: ^.?none.?$
	Subject: ^.?no subject.?$
#v-

This also doesn't work as expected - every message, even with "testing", gets
low score.

Instead, the following works correct:

#v+
[fido7.ru.unix.*]
Score:: =-100 % regular rules/faq postings
	{:
	  Subject: RULES
	  From: "Moderator of RU\.UNIX.*" <netch@segfault\.kiev\.ua>
	}
	From: "RU\.UNIX.* FAQ .*" <netch@segfault\.kiev\.ua>
	From: FAQSERVER
#v-

So, is all of this code bug, score/regexp documentation bug, or my own
misunderstanding?

-- 
WBR, Vadim Goncharov. ICQ#166852181       mailto:[email protected]
[Moderator of RU.ANTI-ECOLOGY][FreeBSD][http://antigreen.org][LJ:/nuclight]


------------------------------------------------------------------------------
What Every C/C++ and Fortran developer Should Know!
Read this article and learn how Intel has extended the reach of its 
next-generation tools to help Windows* and Linux* C/C++ and Fortran 
developers boost performance applications - including clusters. 
http://p.sf.net/sfu/intel-dev2devmay