aasmith / feed-normalizer Goto Github PK

Current gem version is two years old, and there were some improvements for this time.

Maybe, it's possible to release it?

any chance you can update the current gem on rubygems.org? The latest gem only has changes up until January 2010.

Hello,

I've aggregating two feeds and sorting them by date. Here's the code:

feeds = [
    'http://blog.woralelandia.com/tag/mi-musica/feed/',
    'http://mexicoindie.net/feed/'
]

stories = []
feeds.each do |f|
    feed = FeedNormalizer::FeedNormalizer.parse f
    stories.push( *feed.entries )
end

stories.sort_by! { |k| k[ :date_published ] }

SystemStackError: stack level too deep
from /home/renich/.gem/ruby/2.2.0/gems/feed-normalizer-1.5.2/lib/structures.rb:14:in `rescue in method_missing'

I am getting 20 stories so it shouldn't be much. Could you help me out with this and maybe add a suggestion at the readme?

I ran into this while processing a twitter feed where someone had managed to inject new-line characters into the tweet.

Here's the content in question:
"AppWireless: The following stores will have the Motorola Milestone by this afternoon: \nGordon’s\n Photo, Hazard, Harlan,... http://fb.me/HmcxV9X8"

When I run clean on that, I get:
>> FeedNormalizer::HtmlCleaner.clean str
NoMethodError: undefined method traverse_element' for nil:NilClass from /usr/lib/ruby/gems/1.8/gems/activesupport-2.3.8/lib/active_support/whiny_nil.rb:52:inmethod_missing'
from /usr/lib/ruby/gems/1.8/gems/hpricot-0.8.2/lib/hpricot/traverse.rb:313:in /' from /usr/lib/ruby/gems/1.8/gems/hpricot-0.8.2/lib/hpricot/traverse.rb:310:ineach'
from /usr/lib/ruby/gems/1.8/gems/hpricot-0.8.2/lib/hpricot/traverse.rb:310:in /' from /usr/lib/ruby/gems/1.8/gems/feed-normalizer-1.5.2/lib/html-cleaner.rb:164:inremove_tags!'
from /usr/lib/ruby/gems/1.8/gems/feed-normalizer-1.5.2/lib/html-cleaner.rb:72:in `clean'
from (irb):82

From what I can tell, the problem stems from this regular expression in #clean:
# get all the tags in the document
# Somewhere near hpricot 0.4.92 "" starting to return all elements,
# including text nodes instead of just tagged elements.
tags = (doc/"").inject([]) { |m,e| m << e.name if(e.respond_to?(:name) && e.name =~ /^\w+$/) ; m }.uniq

Which is matching against the full text node, and matches "Gordon's" since it is wrapped in newlines. I'm guessing the regular expression is there to try to filter out text nodes? If so, I'd suggest changing the pattern to
/\A\w+\Z/
To force the match against to whole string. Unless there's a cleaner way to filter out text nodes?