ipfw status report - 7

[Tatsiana Elavaya]

Last week I've spent mostly trying to rewrite rule dump the right way.
I had at least 3 working implementations, but they all turned to be
incorrect. Original idea of dumping rules in one pass now seems plain
wrong to me. I'm going to produce complete in memory ruleset and dump
it afterwards. That would help adding skipto-rules before big chunk of
rules and perform some minimal optimization at preprocessing stage. My
intent is still to keep resulting ruleset as close to hll variant as
possible.
Never the less, weekly progress is following:
* Add regression tests
* Use call operator @ to refer to other condition sets/rulesets
* Implement shell-like variables
* Support escaped strings
* Add id and str tokens (also containing line number)
* Use cmd list to store rule actions