`A M' gives consistently wrong results

[Harry Putnam <reader@newsguy.com>]

Why does:
`A M'
     List groups that match a regexp (`gnus-group-list-all-matching')

Give consistently wrong results?

In group buffer I have many groups that match this regexp:
nnml:[A-Z]

That is, nnml groups that begin with at least one capital letter.

However `A M nnml:[A-Z]' finds a conglomerate of groups
beginning with a-z and A-Z.

Apparently the alleged regexp matching is not case sensitive.

I wonder why it is necessary to spoil good regexp matching like that?
Most applications allow you to choose case insenstive with an -i flag
or similar and default to real (i.e. case sensitive regexp), or at
least have something sensible like mixed upper lower causing case
sensitivity. 

Can I make `gnus-group-list-all-matching' case sensitive without
writing 5 acres of poor trial and error lisp or crawling thru 400 meters
of hot coals and broken beer bottles?

Shouldn't case sensitive, or at least the mixed upper lower control be
the default?  Why is it considered better to make case sensisitve
difficult?  Its not that hard to write regexp so that they are case
insensitive if one wants that.  But it is impossilbe to write a regexp
that is case sensitive if case insensitive is the default.

Regexp are designed to handle that need already, why mess that up?