[HTTPS-Everywhere] Python interface for the rulesets?

[Ondrej Mikle]

Hi,

On 01/09/2013 11:23 PM, Peter Eckersley wrote:
> By the way, MB just encountered this bug in the JS ruleset implementation; you
> probably didn't implement the same bug in https-everywhere-checker but just
> in case:
> 
> https://gitweb.torproject.org/https-everywhere.git/commitdiff/6ca405d010062d2b2cb91b2024d4ebf7d405dee7

I've implemented the feature for (leftmost) wildcard '*' to expand to
arbitrary level of subdomains, if it's at least 3rd-level domain:

https://github.com/hiviah/https-everywhere-checker/commits/804500fb1da64679c68c9a1a258cbd7b31d90893/rule_trie.py

E.g. it would expand for *.google.com and *.google.co.uk, but not for
*.com (according to this footnote:
https://www.eff.org/https-everywhere/rulesets#footnote1_a9j7ffp ).

Few other new features that might be useful:

- it's possible to specify list of URLs for testing with "url_list"
section in [http] section, so that they are not guessed from 'target'
ruleset elements
- I was getting lost in the ruleset tree, so I added a debug feature
that exports the ruleset tree in graphviz dot format (option
"dump_graphviz_trie" under [debug]), it looks like this:
  - http://i.imgur.com/s0burYU.png
  - http://i.imgur.com/98R2Fcl.png
  - http://i.imgur.com/Y52WWfG.png

Ondrej