diff options
author | Ralph Amissah <ralph@amissah.com> | 2008-09-16 00:36:14 -0400 |
---|---|---|
committer | Ralph Amissah <ralph@amissah.com> | 2008-09-16 00:36:14 -0400 |
commit | bc9228c42269bfb4d451ca2d2d92a6a12afb094f (patch) | |
tree | 42b666a81ef350d0ff6153a49d159b3a1c348959 /lib/sisu/v0/spell.rb | |
parent | Updated sisu-0.68.0 (diff) | |
parent | fixes: alphabet list (concordance, dal_idx), and file types (dal_expand_inser... (diff) |
Merge branch 'upstream' into debian/sid
Diffstat (limited to 'lib/sisu/v0/spell.rb')
-rw-r--r-- | lib/sisu/v0/spell.rb | 3 |
1 files changed, 2 insertions, 1 deletions
diff --git a/lib/sisu/v0/spell.rb b/lib/sisu/v0/spell.rb index 477a472f..8f89c470 100644 --- a/lib/sisu/v0/spell.rb +++ b/lib/sisu/v0/spell.rb @@ -72,7 +72,8 @@ module Utility end def check @input.each do |data| - data.gsub!(/(https?|www|ftp|gopher|png|jpg|gif|html|htm| )\S+/i,' ') + data.gsub!(/(https?|www|ftp|gopher|png|jpg|gif|html|htm)\S+/i,' ') + data.gsub!(/( |#{Mx[:nbsp]})/i,' ') data.gsub!(/<\/?(table|tr|td|b|p|href).*?>/i,' ') data.gsub!(/(<==.+|<:\S+>|<!.+?!>|^0~.+|\{\{\{|~)/,' ') data.gsub!(/(["|<>)(\n'`'.;&_-]|\=)/,' ') |