From adb212bdfc678af04fa438b42ec06047a13a8f2c Mon Sep 17 00:00:00 2001
From: Giuseppe Bilotta <giuseppe.bilotta@gmail.com>
Date: Tue, 18 Sep 2007 17:31:24 +0000
Subject: first_html_par: build lists 'manually' when using Hpricot

Hpricot selectors (like doc/"css path") don't return elements in their natural (depth-first) order. Use custom searches from the root of the document to achieve this.
---
 lib/rbot/core/utils/utils.rb | 32 ++++++++++++++++----------------
 1 file changed, 16 insertions(+), 16 deletions(-)

(limited to 'lib/rbot/core/utils/utils.rb')

diff --git a/lib/rbot/core/utils/utils.rb b/lib/rbot/core/utils/utils.rb
index 32b05700..9b678def 100644
--- a/lib/rbot/core/utils/utils.rb
+++ b/lib/rbot/core/utils/utils.rb
@@ -498,25 +498,26 @@ module ::Irc
 
       txt = String.new
 
-      h = %w{h1 h2 h3 h4 h5 h6}
-      p = %w{p}
-      ar = []
-      h.each { |hx|
-        p.each { |px|
-          ar << "#{hx}~#{px}"
-        }
-      }
-      h_p_css = ar.join("|")
-      debug "css search: #{h_p_css}"
-
       pre_h = pars = by_span = nil
 
       while true
         debug "Minimum number of spaces: #{min_spaces}"
 
         # Initial attempt: <p> that follows <h\d>
-        pre_h = doc/h_p_css if pre_h.nil?
-        debug "Hx: found: #{pre_h.pretty_inspect}"
+        if pre_h.nil?
+          pre_h = Hpricot::Elements[]
+          found_h = false
+          doc.root.search("*") { |e|
+            case e.pathname
+            when /^h\d/
+              found_h = true
+            when 'p'
+              pre_h << e if found_h
+            end
+          }
+          debug "Hx: found: #{pre_h.pretty_inspect}"
+        end
+
         pre_h.each { |p|
           debug p
           txt = p.to_html.ircify_html
@@ -551,9 +552,8 @@ module ::Irc
         # we don't need
         if by_span.nil?
           by_span = Hpricot::Elements[]
-          pre_pars = doc/"div|span|td|tr|tbody|table"
-          pre_pars.each { |el|
-            by_span.push el if el[:class] =~ /body|message|text/i
+          doc.root.each("*") { |el|
+            by_span.push el if el.pathname =~ /^(?:div|span|td|tr|tbody|table)$/ and el[:class] =~ /body|message|text/i
           }
           debug "other \#1: found: #{by_span.pretty_inspect}"
         end
-- 
cgit v1.2.3