ezkl / sniffles Goto Github PK

Example: http://www.movabletype.com/blog/

Example: http://www.wunderground.com/

Example: http://www.alternet.org/

Example: http://www.topix.com/

Example: http://community.invisionpower.com/

Example: http://demo-store.prestashop.com/en/

Example: http://google.blogspot.com/

Example: http://www.forum-software.org/demo/mybb

Example: http://blog.posterous.com/

Example: http://www.softpedia.com/

Example: http://www.pandoramoa.com/

Example: https://foursquare.com/

This one could be interesting. ASP is an obvious commonality between sites, but it looks like much of the CMS structure is freeform.

Example: http://sites.kiwanis.org/kiwanis/en/home.aspx

Example: http://staff.tumblr.com/

A page can only belong to one CMS, so I think it seems smart to stop checking if it's another CMS if one is detected. However with the way things are currently implemented, I'm not sure on the best way to add this.

That way the sniffer response object could be something along the lines of

:cms => {:platform=>:wordpress, :version=>false, :feed=>"http://feeds.feedburner.com/pearsonified", :theme=>"thesis_17", :pingback=>"http://www.pearsonified.com/xmlrpc.php"}

Rather than

:phpbb=>{:found=>false}, :vbulletin=>{:found=>false}, :wordpress=>{:found=>true, :version=>false, :feed=>"http://feeds.feedburner.com/pearsonified", :theme=>"thesis_17", :pingback=>"http://www.pearsonified.com/xmlrpc.php"}}

I'm not entirely sure on the best implementation either. But currently the only way to find out what CMS a site actually was would be to look through all the CMS keys, and find which were true? Unless I'm overlooking something.

Example: http://www.metafilter.com/

Example: http://www.jonessoda.com/products/

The current WordPress implementation is flawed, it simply checks for "wp-content" anywhere within the source. This will pick up things such as hotlinking to images within the wp-content folder on your domain, or any other domain for that matter.

Suggestion is to use Nokogiri to parse the document, and then check if any link tags(CSS) contain "wp-content".

Example: http://demo.oscommerce.com/

Example: http://www.beenverified.com/

I tried to create a new github organization and the name sniffles is already taken. Not really a big deal as it isn't a Ruby library, but I'm thinking it might be prudent to change the name of the library to something unique, at least in the context of github. I've also grown to dislike the repetition in the namespace.

I'm open to suggestion.

/cc @JakeAustwick, @dchuk

Example: http://www.craftown.com/

Example: https://foursquare.com/

Example: http://shop.angrybirds.com/

Example: http://demo.forum-software.org/vanilla/

Example: http://www.whitehouse.gov/ (Ha!)

Example: http://demo.cs-cart.com/community/

The current version of Sniffles doesn't work because of some crappy code I wrote.
Thanks are due to @dominikh in #ruby-lang on Freenode for pointing this out.

Example: https://www.vbulletin.com/forum/forum.php

Example: http://guistyles.com/

Not sure on the best implementation, but maybe an "depth" option could be added, which affects how much analysis is actually run. You may only want to find the platform of the site for instance, or may not be interested in that at all and only actually want to know what javascript libraries it uses.

Not sure what you think of this, or it you have ideas on the best way to implement this?

The v0.1.0 release includes some pretty dramatic changes to the API. I need to reflect this in the documentation. While I'm at it, I'd like to figure out a good way to organize the documentation now that the feature-set is becoming a bit more rich.

Example: http://tmz.com/

Example: http://store.steelcase.com/

Example: http://www.joomla.org/

Add detection support.

URL to check against: http://www.phpbb.com/styles/demo/3.0/board/?style=0

Example: http://www.downlopedia.com/

Example: http://www.squidoo.com/

Example: http://orangecrushbeads.com/

Example: http://twitter.github.com/bootstrap/

Thinking about breaking each sniffing "pattern" into individual modules; i.e. sniffles-wordpress or sniffles-jquery. The base module would provide the parsers and query methods for common HTML nodes. Thinking mostly about making it easy to extend.

Any feedback on this kind of design @JakeAustwick and @dchuk?

Example: http://www.sixflags.com/national/index.aspx

Example: http://flatpress.org/home/blog.php

Example: http://xenforo.com/community/

I should have done this from the beginning, but I'd like to start using TomDoc for library documentation. It is simple and expressive on top of being easy to remember and implement.

I'm going to make TomDoc a requirement for the v0.2.0 release.

Example: http://expressionengine.com/blog