pidgin2adium 2.0.2 → 3.0.0

data/.autotest

@@ -0,0 +1,9 @@
+Autotest.add_hook :initialize do |at|
+  at.add_mapping(/ext\/.*\/(.*)\.[ch]/) do |_, m|
+        ["test/test_#{m[1]}_extn.rb"]
+      end
+end
+
+Autotest.add_hook :run_command do |at|
+  system "rake compile"
+end

data/History.txt

@@ -1,3 +1,12 @@
+=== 3.0.0 / 2010-01-31
+* balance_tags.rb is now a C extension (Pidgin2Adium.balance_tags_c)
+  - the pure-ruby mixin balance_tags (without the trailing "_c") is gone
+* Better handling of command-line arguments
+* Format time zones offsets correctly (e.g. "+0500", not "+-0500")
+* Write Yahoo! and Jabber logs to correct directories (#27710)
+* Better matching of regexes against time strings
+* Better documentation
+
 === 2.0.2 / 2009-12-18
 * Much better documentation (more of it, and higher quality too!)
 * Allow user-provided output dir at commandline

data/Manifest.txt

@@ -1,10 +1,18 @@
+.autotest
 History.txt
 Manifest.txt
 README.rdoc
 Rakefile.rb
 bin/pidgin2adium
+ext/balance_tags_c/balance_tags_c.c
+ext/balance_tags_c/extconf.rb
 lib/pidgin2adium.rb
-lib/pidgin2adium/balance_tags.rb
 lib/pidgin2adium/log_converter.rb
 lib/pidgin2adium/log_file.rb
 lib/pidgin2adium/log_parser.rb
+script/console
+script/destroy
+script/generate
+tasks/extconf.rake
+tasks/extconf/balance_tags_c.rake
+test/test_balance_tags_c_extn.rb

data/README.rdoc

@@ -37,10 +37,7 @@ Or:
  $ pidgin2adium -i pidgin-logs -a gabe,gabeb-w,gbw
 
 ===Example (using library)
-The library style allows you to parse a log file and get back a
-LogFile[link:classes/Pidgin2Adium/LogFile.html] instance for easy reading, manipulation, etc.
-You can also create log files yourself using Pidgin2Adium.parse_and_generate[link:classes/Pidgin2Adium.html#M000006].
-For batch processing, use LogConverter[link:classes/Pidgin2Adium/LogConverter.html].
+The library style allows you to parse a log file and get back a LogFile[link:classes/Pidgin2Adium/LogFile.html] instance for easy reading, manipulation, etc. If you don't need to do anything with the individual messages, use Pidgin2Adium.parse[link:classes/Pidgin2Adium.html#M000002].
  
  require 'pidgin2adium'
  logfile = Pidgin2Adium.parse("/path/to/log/file.html", "gabe,gbw,gabeb-w")
@@ -81,15 +78,17 @@ For batch processing, use LogConverter[link:classes/Pidgin2Adium/LogConverter.ht
 	puts "Successfully wrote out log file!"
 	puts "Path to output file: #{success}"
     end
-    # This deletes search indexes to Adium re-indexes the new chat logs.
-    # Not automatically called after log_file.write_out()
-    # If you're converting >1 log, call it after converting all of them, since
-    # it takes up a tiny bit of processing power.
+    # This deletes search indexes so Adium re-indexes the new chat logs.
+    # It is not automatically called after log_file.write_out()
+    # Call it after converting all the logs, since it takes up a bit of
+    # processing power.
     Pidgin2Adium.delete_search_indexes()
  end
 
 ===Example 2 (using library)
-If you want to output the logfile to an output dir instead of just parsing it, use Pidgin2Adium.parse_and_generate:
+If you want to parse the file and write it out instead of just parsing it, use Pidgin2Adium.parse_and_generate.
+
+Note: For batch processing, use LogConverter[link:classes/Pidgin2Adium/LogConverter.html].
  
  require 'pidgin2adium'
  # Both options are optional; without :output_dir, writes to Adium log dir

data/Rakefile.rb

@@ -18,12 +18,23 @@ $hoe = Hoe.spec 'pidgin2adium' do
   #self.post_install_message = 'PostInstall.txt' # TODO remove if post-install message not required
   self.rubyforge_name       = self.name # this is default value
   # self.extra_deps         = [['activesupport','>= 2.0.2']]
+  
+  self.spec_extras[:extensions]  = "ext/balance_tags_c/extconf.rb"
+end
+
+$hoe.spec.rdoc_options = %w{--main README.rdoc}
+
+# Use hanna RDoc template, if available
+begin
+    gem "hanna"
+    $hoe.spec.rdoc_options << '-T hanna'
+rescue GEM::LoadError
+    # hanna not installed, continue
 end
-# Use hanna RDoc template
-$hoe.spec.rdoc_options = %w{--main README.rdoc -T hanna}
 
 require 'newgem/tasks'
 Dir['tasks/**/*.rake'].each { |t| load t }
+task :postrelease => [:publish_docs, :announce]
 
 # TODO - want other tests/tasks run by default? Add them to the list
 # remove_task :default

data/bin/pidgin2adium

@@ -32,7 +32,7 @@ oparser = OptionParser.new do |opts|
 	    "is you. Whitespace and case do not matter.") do |aliases|
 	options[:aliases] = aliases
     end
-    opts.on('-o', '--out [OUT_DIR]',
+    opts.on('-o', '--out OUT_DIR',
 	    'The top-level directory under which to',
 	    'store the logs (each in its own folder',
 	    'by screen name).',

data/ext/balance_tags_c/balance_tags_c.c

@@ -0,0 +1,198 @@
+/*
+ * Balances tags of string using a modified stack. Returns a balanced string.
+ * 
+ * From Wordpress's formatting.php and rewritten in C by
+ * Gabe Berke-Williams, 2010.
+ * 
+ * Original Author:: Leonard Lin <leonard@acm.org>
+ * License:: GPL v2.0
+ * Copyright:: November 4, 2001
+ */
+
+#include <ruby.h>
+
+#ifndef RARRAY_PTR
+#define RARRAY_PTR(arr)  RARRAY(arr)->ptr
+#endif
+
+#ifndef RSTRING_LEN
+#define RSTRING_LEN(str) RSTRING(str)->len
+#endif
+
+VALUE balance_tags_c(VALUE, VALUE);
+static VALUE mP2A;
+
+/*
+ * call-seq: Pidgin2Adium.balance_tags_c(text) => text with balanced HTML tags
+ *
+ * Balances tags of _text_. Returns modified text.
+ */
+VALUE balance_tags_c(VALUE mod, VALUE text){
+    if( TYPE(text) != T_STRING ){
+	rb_raise(rb_eArgError, "bad argument to balance_tags_c, String only please.");
+    }
+    VALUE tagstack = rb_ary_new2(1);
+    int stacksize = 0;
+    VALUE tagqueue = rb_str_new2("");
+    VALUE ZERO = INT2FIX(0),
+	  ONE = INT2FIX(1);
+    VALUE newtext = rb_str_new2("");
+    // Known single-entity/self-closing tags
+    VALUE single_tags = rb_ary_new3(5,
+	    rb_str_new2("br"),
+	    rb_str_new2("hr"),
+	    rb_str_new2("img"),
+	    rb_str_new2("input"),
+	    rb_str_new2("meta"));
+    // Tags that can be immediately nested within themselves
+    VALUE nestable_tags = rb_ary_new3(4,
+	    rb_str_new2("blockquote"),
+	    rb_str_new2("div"),
+	    rb_str_new2("span"),
+	    rb_str_new2("font"));
+    // 1: tagname, with possible leading "/"
+    // 2: attributes
+    VALUE tag_regex = rb_reg_regcomp(rb_str_new2("<(\\/?\\w*)\\s*([^>]*)>"));
+    VALUE pos;  // position in text
+    VALUE match;
+    VALUE tag;
+    VALUE attributes;
+    VALUE t; // loop variable when iterating over tagstack at end of while loop
+    int matchlen;
+    int done = 0;
+    int j, k, i; // loop counters
+
+    // WP bug fix for comments - in case you REALLY meant to type '< !--'
+    rb_funcall(text, rb_intern("gsub!"), 2,
+	    rb_str_new2("< !--"),
+	    rb_str_new2("<    !--"));
+
+    // WP bug fix for LOVE <3 (and other situations with '<' before a number)
+    rb_funcall(text, rb_intern("gsub!"), 2,
+	    rb_reg_regcomp(rb_str_new2("<([0-9]{1})")),
+	    rb_str_new2("&lt;\\1"));
+
+    pos = rb_funcall(text, rb_intern("=~"), 1, tag_regex);
+    done = (pos == Qnil);
+    while ( ! done ){
+	rb_str_concat(newtext, tagqueue);
+	match = rb_funcall(text, rb_intern("match"), 1, tag_regex);
+	tag = rb_funcall(rb_reg_nth_match(1, match), rb_intern("downcase"), 0);
+	attributes = rb_reg_nth_match(2, match);
+
+	matchlen = NUM2INT(rb_funcall(rb_reg_nth_match(0, match), rb_intern("size"), 0));
+
+	// clear the shifter
+	tagqueue = rb_str_new2("");
+	// Pop or Push
+	if (0 == rb_str_cmp(rb_str_substr(tag, 0, 1), rb_str_new2("/"))){ // End Tag
+	    rb_funcall(tag, rb_intern("slice!"), 2, ZERO, ONE);
+	    // if too many closing tags
+	    if(stacksize <= 0){
+		tag = rb_str_new2("");
+		//or close to be safe: tag = '/' << tag
+	    } else if (0 == rb_str_cmp(RARRAY_PTR(tagstack)[stacksize - 1], tag)){
+		// found closing tag
+		// if stacktop value == tag close value then pop
+		// Close Tag
+		tag = rb_str_append(rb_str_new2("</"), tag);
+		rb_str_concat(tag, rb_str_new2(">")); 
+		// Pop
+		rb_ary_pop(tagstack);
+		stacksize--;
+		} else { // closing tag not at top, search for it
+		    for(j=stacksize-1; j>=0; j--){
+			if(0 == rb_str_cmp(RARRAY_PTR(tagstack)[j], tag) ){
+			    // add tag to tagqueue
+			    for(k = stacksize-1;k>=j;k--){
+				rb_str_concat(tagqueue, rb_str_new2("</"));
+				rb_str_concat(tagqueue, rb_ary_pop(tagstack));
+				rb_str_concat(tagqueue, rb_str_new2(">"));
+				stacksize--;
+			    }
+			    break;
+			}
+		    }
+		    tag = rb_str_new2("");
+		}
+	    } else {
+		// Begin Tag
+
+		// Tag Cleaning
+
+		if( ( RSTRING_LEN(attributes) > 0 && // test length before rb_str_substr
+		    (0 == rb_str_cmp(rb_str_substr(attributes, -1, 1), rb_str_new2("/"))) ) ||
+			(0 == rb_str_cmp(tag, rb_str_new2(""))) ){
+		    // If: self-closing or '', don't do anything.
+		} else if ( rb_ary_includes(single_tags, tag) ) {
+		    // ElseIf: it's a known single-entity tag but it doesn't close itself, do so
+		    rb_str_concat(attributes, rb_str_new2("/"));
+		} else {
+		    // Push the tag onto the stack
+		    // If the top of the stack is the same as the tag we want
+		    // to push, close previous tag
+		    if ( (stacksize > 0) &&
+			(Qfalse == rb_ary_includes(nestable_tags, tag)) &&
+			(0 == rb_str_cmp(rb_ary_entry(tagstack, stacksize - 1), tag))){
+			tagqueue = rb_str_new2("</");
+			rb_str_concat(tagqueue, rb_ary_pop(tagstack));
+			rb_str_concat(tagqueue, rb_str_new2(">"));
+			stacksize--; 
+		    }
+		    rb_ary_push(tagstack, tag);
+		    stacksize++;
+		}
+
+		// Attributes
+		if( 0 != rb_str_cmp(attributes, rb_str_new2("")) ){
+		    attributes = rb_str_plus(rb_str_new2(" "), attributes);
+		}
+		tag = rb_str_plus(rb_str_new2("<"), tag);
+		rb_str_concat(tag, attributes);
+		rb_str_concat(tag, rb_str_new2(">"));
+		//If already queuing a close tag, then put this tag on, too
+		if( RSTRING_LEN(tagqueue) > 0 ){
+		    rb_str_concat(tagqueue, tag);
+		    tag = rb_str_new2("");
+		}
+	    }
+	    rb_str_concat(newtext,
+		    rb_str_plus(rb_str_substr(text, 0, pos-1), tag));
+	    text = rb_str_substr(text,
+		    NUM2INT(pos)+matchlen,
+		    RSTRING_LEN(text) - (NUM2INT(pos)+matchlen));
+	    pos = rb_funcall(text, rb_intern("=~"), 1, tag_regex);
+	    done = (pos == Qnil);
+	}
+
+	// Clear Tag Queue
+	rb_str_concat(newtext, tagqueue);
+
+	// Add Remaining text
+	rb_str_concat(newtext, text);
+
+	i = NUM2INT(rb_funcall(tagstack, rb_intern("length"), 0)) - 1;
+	// Empty Stack
+	for(; i >= 0; i--){
+	    // Add remaining tags to close
+	    t = RARRAY_PTR(tagstack)[i];
+	    rb_str_concat(newtext, rb_str_new2("</"));
+	    rb_str_concat(newtext, t);
+	    rb_str_concat(newtext, rb_str_new2(">"));
+	}
+
+	// WP fix for the bug with HTML comments
+	rb_funcall(newtext, rb_intern("gsub!"), 2,
+		rb_str_new2("< !--"),
+		rb_str_new2("<!--"));
+	rb_funcall(newtext, rb_intern("gsub!"), 2,
+		rb_str_new2("<    !--"),
+		rb_str_new2("< !--"));
+
+	return newtext;
+}
+
+void Init_balance_tags_c(){
+    mP2A = rb_define_module("Pidgin2Adium");
+    rb_define_module_function(mP2A, "balance_tags_c", balance_tags_c, 1);
+}

data/ext/balance_tags_c/extconf.rb

@@ -0,0 +1,4 @@
+require 'mkmf'
+
+dir_config("balance_tags_c")
+create_makefile("balance_tags_c")

data/lib/pidgin2adium.rb

@@ -14,7 +14,7 @@ module Pidgin2Adium
     ADIUM_LOG_DIR = File.expand_path('~/Library/Application Support/Adium 2.0/Users/Default/Logs/') << '/' 
     # These files/directories show up in Dir.entries()
     BAD_DIRS = %w{. .. .DS_Store Thumbs.db .system}
-    VERSION = "2.0.2"
+    VERSION = "3.0.0"
     # For displaying after we finish converting
     @@oops_messages = []
     @@error_messages = []
@@ -38,6 +38,7 @@ module Pidgin2Adium
     module_function :log_msg, :oops, :error
     #######################
 
+    # Parses the provided log.
     # Returns a LogFile instance or false if an error occurred.
     def parse(logfile_path, my_aliases)
 	logfile_path = File.expand_path(logfile_path)
@@ -54,7 +55,8 @@ module Pidgin2Adium
 
 	return parser.parse()
     end
-    
+   
+    # Parses the provided log and writes out the log in Adium format.
     # Returns the path to the converted log, false if an error occurred, or
     # Pidgin2Adium::FILE_EXISTS if file already exists AND opts[:overwrite] =
     # false.

data/lib/pidgin2adium/log_file.rb

@@ -26,12 +26,12 @@ module Pidgin2Adium
 	    # key is for Pidgin, value is for Adium
 	    # Just used for <service>.<screenname> in directory structure
 	    service_name_map = {'aim' => 'AIM',
-		'jabber' =>'jabber',
+		'jabber' =>'Jabber',
 		'gtalk'=> 'GTalk',
 		'icq' => 'ICQ',
 		'qq' => 'QQ',
 		'msn' => 'MSN',
-		'yahoo' => 'Yahoo'}
+		'yahoo' => 'Yahoo!'}
 	    
 	    @service = service_name_map[service.downcase]
 	end

data/lib/pidgin2adium/log_parser.rb

@@ -7,7 +7,7 @@
 require 'parsedate'
 require 'time' # for Time.zone_offset
 
-require 'pidgin2adium/balance_tags'
+require 'balance_tags_c'
 require 'pidgin2adium/log_file'
 
 module Pidgin2Adium
@@ -42,12 +42,12 @@ module Pidgin2Adium
 	    # Time regexes must be set before pre_parse().
 	    # "4/18/2007 11:02:00 AM" => %w{4, 18, 2007, 11, 02, 00, AM}
 	    # ONLY used (if at all) in first line of chat ("Conversation with...at...")
-	    @time_regex_first_line = %r{(\d{1,2})/(\d{1,2})/(\d{4}) (\d{1,2}):(\d{2}):(\d{2}) ([AP]M)}
+	    @time_regex_first_line = %r{^(\d{1,2})/(\d{1,2})/(\d{4}) (\d{1,2}):(\d{2}):(\d{2}) ([AP]M)$}
 	    # "2007-04-17 12:33:13" => %w{2007, 04, 17, 12, 33, 13}
-	    @time_regex = /(\d{4})-(\d{2})-(\d{2}) (\d{2}):(\d{2}):(\d{2})/
+	    @time_regex = /^(\d{4})-(\d{2})-(\d{2}) (\d{2}):(\d{2}):(\d{2})$/
 	    # sometimes a line in a chat doesn't have a full timestamp
 	    # "04:22:05 AM" => %w{04 22 05 AM}
-	    @minimal_time_regex = /(\d{1,2}):(\d{2}):(\d{2})( [AP]M)?/
+	    @minimal_time_regex = /^(\d{1,2}):(\d{2}):(\d{2})( [AP]M)?$/
 
 	    # Whether or not the first line is parseable.
 	    @first_line_is_valid = true
@@ -90,10 +90,10 @@ module Pidgin2Adium
 		/(.+) is offering to send file/,
 		/^Transfer of file .+ complete$/,
 		/Error reading|writing|accessing .+: .+/,
-		/You cancelled the transfer of/,
+		/You cancell?ed the transfer of/,
 		/File transfer cancelled/,
-		/(.+) cancelled the transfer of/,
-		/(.+) cancelled the file transfer/,
+		/(.+?) cancell?ed the transfer of/,
+		/(.+?) cancelled the file transfer/,
 		# Direct IM - actual (dis)connect events are their own types
 		/^Attempting to connect to (.+) at .+ for Direct IM\./,
 		/^Asking (.+) to connect to us at .+ for Direct IM\./,
@@ -106,9 +106,6 @@ module Pidgin2Adium
 		/^Conflicting Key Received!$/,
 		/^Error in decryption- asking for resend\.\.\.$/,
 		/^Making new key pair\.\.\.$/,
-		# file transfer
-		/You canceled the transfer of/,
-		/(.+?) canceled the transfer of/, 
 		# sending errors
 		/^Last outgoing message not received properly- resetting$/,
 		/Resending\.\.\./,
@@ -178,9 +175,8 @@ module Pidgin2Adium
 	    if tz_match and tz_match[1]
 		tz_offset = tz_match[1]
 	    else
-		zone = Time.local(Time.new.year).zone
 		# "-0500" (3d rather than 2d to allow for "+")
-		tz_offset = sprintf('+%03d00', Time.zone_offset(zone) / 3600)
+		tz_offset = sprintf('%+03d00', Time.zone_offset(Time.now.zone) / 3600)
 	    end
 	    return tz_offset
 	end
@@ -255,7 +251,7 @@ module Pidgin2Adium
 		user_SN = first_line_match[3].downcase.tr(' ', '')
 		partner_SN = first_line_match[1]
 		pidgin_chat_time_start = first_line_match[2]
-		basic_time_info = case @first_line
+		basic_time_info = case pidgin_chat_time_start
 				  when @time_regex: [$1.to_i, $2.to_i, $3.to_i]
 				  when @time_regex_first_line: [$3.to_i, $1.to_i, $2.to_i]
 				  end
@@ -488,18 +484,15 @@ module Pidgin2Adium
 		
 		styleparts = style.split(/; ?/)
 		styleparts.map! do |p|
-		    if p =~ /^color/
-			# Regarding the bit with the ">", sometimes this happens:
-			# <span style="color: #000000>today;">today was busy</span>
-			# Then p = "color: #000000>today"
-			# Or it can end in ">;", with no text before the semicolon.
-			# So remove the ">" and anything following it.
-			
-			# Use regex instead of string, to account for funky ">" stuff
-			if p =~ /color: #000000/
+		    if p[0,5] == 'color'
+			if p.include?('color: #000000')
 			    next
 			elsif p =~ /(color: #[0-9a-fA-F]{6})(>.*)?/
-			    # Keep the color but remove the bit after it
+			    # Regarding the bit with the ">", sometimes this happens:
+			    # <span style="color: #000000>today;">today was busy</span>
+			    # Then p = "color: #000000>today"
+			    # Or it can end in ">;", with no text before the semicolon.
+			    # So keep the color but remove the ">" and anything following it.
 			    next($1)
 			end
 		    else
@@ -546,7 +539,6 @@ module Pidgin2Adium
     # Basic message with body text (as opposed to pure status messages, which
     # have no body).
     class XMLMessage < Message
-	include Pidgin2Adium
 	def initialize(sender, time, buddy_alias, body)
 	    super(sender, time, buddy_alias)
 	    @body = body
@@ -571,7 +563,7 @@ module Pidgin2Adium
 	    normalize_body_entities!()
 	    # Fix mismatched tags. Yes, it's faster to do it per-message
 	    # than all at once.
-	    @body = balance_tags(@body)
+	    @body = Pidgin2Adium.balance_tags_c(@body)
 	    if @buddy_alias[0,3] == '***'
 		# "***<alias>" is what pidgin sets as the alias for a /me action
 		@buddy_alias.slice!(0,3)

data/script/console

@@ -0,0 +1,10 @@
+#!/usr/bin/env ruby
+# File: script/console
+irb = RUBY_PLATFORM =~ /(:?mswin|mingw)/ ? 'irb.bat' : 'irb'
+
+libs =  " -r irb/completion"
+# Perhaps use a console_lib to store any extra methods I may want available in the cosole
+# libs << " -r #{File.dirname(__FILE__) + '/../lib/console_lib/console_logger.rb'}"
+libs <<  " -r #{File.dirname(__FILE__) + '/../lib/p2a.rb'}"
+puts "Loading p2a gem"
+exec "#{irb} #{libs} --simple-prompt"

data/script/destroy

@@ -0,0 +1,14 @@
+#!/usr/bin/env ruby
+APP_ROOT = File.expand_path(File.join(File.dirname(__FILE__), '..'))
+
+begin
+  require 'rubigen'
+rescue LoadError
+  require 'rubygems'
+  require 'rubigen'
+end
+require 'rubigen/scripts/destroy'
+
+ARGV.shift if ['--help', '-h'].include?(ARGV[0])
+RubiGen::Base.use_component_sources! [:rubygems, :newgem, :newgem_theme, :test_unit]
+RubiGen::Scripts::Destroy.new.run(ARGV)

data/script/generate

@@ -0,0 +1,14 @@
+#!/usr/bin/env ruby
+APP_ROOT = File.expand_path(File.join(File.dirname(__FILE__), '..'))
+
+begin
+  require 'rubigen'
+rescue LoadError
+  require 'rubygems'
+  require 'rubigen'
+end
+require 'rubigen/scripts/generate'
+
+ARGV.shift if ['--help', '-h'].include?(ARGV[0])
+RubiGen::Base.use_component_sources! [:rubygems, :newgem, :newgem_theme, :test_unit]
+RubiGen::Scripts::Generate.new.run(ARGV)

data/tasks/extconf.rake

@@ -0,0 +1,13 @@
+namespace :extconf do
+  desc "Compiles the Ruby extension"
+  task :compile
+end
+
+task :compile => "extconf:compile"
+
+task :test => :compile
+
+BIN = "*.{bundle,jar,so,obj,pdb,lib,def,exp}"
+$hoe.clean_globs |= ["ext/**/#{BIN}", "lib/**/#{BIN}", 'ext/**/Makefile']
+$hoe.spec.require_paths = Dir['{lib,ext/*}']
+$hoe.spec.extensions = FileList["ext/**/extconf.rb"].to_a

data/tasks/extconf/balance_tags_c.rake

@@ -0,0 +1,43 @@
+namespace :extconf do
+  extension = File.basename(__FILE__, '.rake')
+
+  ext = "ext/#{extension}"
+  ext_so = "#{ext}/#{extension}.#{Config::CONFIG['DLEXT']}"
+  ext_files = FileList[
+    "#{ext}/*.c",
+    "#{ext}/*.h",
+    "#{ext}/*.rl",
+    "#{ext}/extconf.rb",
+    "#{ext}/Makefile",
+    # "lib"
+  ]
+
+
+  task :compile => extension do
+    if Dir.glob("**/#{extension}.{o,so,dll}").length == 0
+      STDERR.puts "!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!"
+      STDERR.puts "Gem actually failed to build.  Your system is"
+      STDERR.puts "NOT configured properly to build #{GEM_NAME}."
+      STDERR.puts "!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!"
+      exit(1)
+    end
+  end
+
+  desc "Builds just the #{extension} extension"
+  task extension.to_sym => ["#{ext}/Makefile", ext_so ]
+
+  file "#{ext}/Makefile" => ["#{ext}/extconf.rb"] do
+    Dir.chdir(ext) do ruby "extconf.rb" end
+  end
+
+  file ext_so => ext_files do
+    Dir.chdir(ext) do
+      sh(PLATFORM =~ /win32/ ? 'nmake' : 'make') do |ok, res|
+        if !ok
+          require "fileutils"
+          FileUtils.rm Dir.glob('*.{so,o,dll,bundle}')
+        end
+      end
+    end
+  end
+end

data/test/test_balance_tags_c_extn.rb

@@ -0,0 +1,10 @@
+require "test/unit"
+
+$:.unshift File.dirname(__FILE__) + "/../ext/balance_tags_c"
+require "balance_tags_c.so"
+
+class TestBalanceTagsCExtn < Test::Unit::TestCase
+  def test_truth
+    assert true
+  end
+end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification 
 name: pidgin2adium
 version: !ruby/object:Gem::Version 
-  version: 2.0.2
+  version: 3.0.0
 platform: ruby
 authors: 
 - Gabe B-W
@@ -9,9 +9,29 @@ autorequire:
 bindir: bin
 cert_chain: []
 
-date: 2009-12-18 00:00:00 -08:00
+date: 2010-01-31 00:00:00 -05:00
 default_executable: 
 dependencies: 
+- !ruby/object:Gem::Dependency 
+  name: rubyforge
+  type: :development
+  version_requirement: 
+  version_requirements: !ruby/object:Gem::Requirement 
+    requirements: 
+    - - ">="
+      - !ruby/object:Gem::Version 
+        version: 2.0.3
+    version: 
+- !ruby/object:Gem::Dependency 
+  name: gemcutter
+  type: :development
+  version_requirement: 
+  version_requirements: !ruby/object:Gem::Requirement 
+    requirements: 
+    - - ">="
+      - !ruby/object:Gem::Version 
+        version: 0.3.0
+    version: 
 - !ruby/object:Gem::Dependency 
   name: hoe
   type: :development
@@ -20,7 +40,7 @@ dependencies:
     requirements: 
     - - ">="
       - !ruby/object:Gem::Version 
-        version: 2.4.0
+        version: 2.5.0
     version: 
 description: |-
   Pidgin2Adium is a fast, easy way to convert Pidgin (formerly gaim) logs to the
@@ -30,23 +50,31 @@ email:
 - pidgin2adium@brandeis.edu
 executables: 
 - pidgin2adium
-extensions: []
-
+extensions: 
+- ext/balance_tags_c/extconf.rb
 extra_rdoc_files: 
 - History.txt
 - Manifest.txt
 - README.rdoc
 files: 
+- .autotest
 - History.txt
 - Manifest.txt
 - README.rdoc
 - Rakefile.rb
 - bin/pidgin2adium
+- ext/balance_tags_c/balance_tags_c.c
+- ext/balance_tags_c/extconf.rb
 - lib/pidgin2adium.rb
-- lib/pidgin2adium/balance_tags.rb
 - lib/pidgin2adium/log_converter.rb
 - lib/pidgin2adium/log_file.rb
 - lib/pidgin2adium/log_parser.rb
+- script/console
+- script/destroy
+- script/generate
+- tasks/extconf.rake
+- tasks/extconf/balance_tags_c.rake
+- test/test_balance_tags_c_extn.rb
 has_rdoc: true
 homepage: http://rubyforge.org/projects/pidgin2adium/
 licenses: []
@@ -55,10 +83,10 @@ post_install_message:
 rdoc_options: 
 - --main
 - README.rdoc
-- -T
-- hanna
+- -T hanna
 require_paths: 
 - lib
+- ext/balance_tags_c
 required_ruby_version: !ruby/object:Gem::Requirement 
   requirements: 
   - - ">="
@@ -78,5 +106,5 @@ rubygems_version: 1.3.5
 signing_key: 
 specification_version: 3
 summary: Pidgin2Adium is a fast, easy way to convert Pidgin (formerly gaim) logs to the Adium format
-test_files: []
-
+test_files: 
+- test/test_balance_tags_c_extn.rb

data/lib/pidgin2adium/balance_tags.rb

@@ -1,118 +0,0 @@
-module Pidgin2Adium
-    # Balances tags of string using a modified stack. Returns a balanced
-    # string, but also affects the text passed into it!
-    # Use text = balance_tags(text).
-    
-    # From Wordpress's formatting.php; rewritten in Ruby by Gabe
-    # Berke-Williams, 2009.
-    # Author:: Leonard Lin <leonard@acm.org>
-    # License:: GPL v2.0
-    # Copyright:: November 4, 2001
-    def balance_tags( text )
-	tagstack = []
-	stacksize = 0
-	tagqueue = ''
-	newtext = ''
-	single_tags = %w{br hr img input meta} # Known single-entity/self-closing tags
-	#nestable_tags = %w{blockquote div span} # Tags that can be immediately nested within themselves
-	nestable_tags = %w{blockquote div span font} # Tags that can be immediately nested within themselves
-	# 1: tagname, with possible leading "/"
-	# 2: attributes
-	tag_regex = /<(\/?\w*)\s*([^>]*)>/
-
-	# WP bug fix for comments - in case you REALLY meant to type '< !--'
-	text.gsub!('< !--', '<    !--')
-
-	# WP bug fix for LOVE <3 (and other situations with '<' before a number)
-	text.gsub!(/<([0-9]{1})/, '&lt;\1')
-
-	while ( pos = (text =~ tag_regex) )
-	    newtext << tagqueue
-	    tag = $1.downcase
-	    attributes = $2
-	    matchlen = $~[0].size
-
-	    # clear the shifter
-	    tagqueue = ''
-	    # Pop or Push
-	    if (tag[0,1] == "/") # End Tag
-		tag.slice!(0,1)
-		# if too many closing tags
-		if(stacksize <= 0)
-		    tag = ''
-		    #or close to be safe: tag = '/' << tag
-		elsif (tagstack[stacksize - 1] == tag) # found closing tag
-		    # if stacktop value == tag close value then pop
-		    tag = '</' << tag << '>' # Close Tag
-		    # Pop
-		    tagstack.pop
-		    stacksize -= 1
-		else # closing tag not at top, search for it
-		    (stacksize-1).downto(0) do |j|
-			if (tagstack[j] == tag)
-			    # add tag to tagqueue
-			    ss = stacksize - 1
-			    ss.downto(j) do |k|
-				tagqueue << '</' << tagstack.pop << '>'
-				stacksize -= 1
-			    end
-			    break
-			end
-		    end
-		    tag = ''
-		end
-	    else
-		# Begin Tag
-
-		# Tag Cleaning
-		if( (attributes[-1,1] == '/') || (tag == '') )
-		    # If: self-closing or '', don't do anything.
-		elsif ( single_tags.include?(tag) )
-		    # ElseIf: it's a known single-entity tag but it doesn't close itself, do so
-		    attributes << '/'
-		else
-		    # Push the tag onto the stack
-		    # If the top of the stack is the same as the tag we want to push, close previous tag
-		    if ((stacksize > 0) &&
-			! nestable_tags.include?(tag) &&
-			(tagstack[stacksize - 1] == tag))
-			tagqueue = '</' << tagstack.pop << '>'
-			stacksize -= 1
-		    end
-		    tagstack.push(tag)
-		    stacksize += 1
-		end
-
-		# Attributes
-		if(attributes != '')
-		    attributes = ' ' << attributes
-		end
-		tag = '<' << tag << attributes << '>'
-		#If already queuing a close tag, then put this tag on, too
-		if (tagqueue)
-		    tagqueue << tag
-		    tag = ''
-		end
-	    end
-	    newtext << text[0,pos] << tag
-	    text = text[pos+matchlen, text.length - (pos+matchlen)]
-	end
-
-	# Clear Tag Queue
-	newtext << tagqueue
-
-	# Add Remaining text
-	newtext << text
-
-	# Empty Stack
-	tagstack.reverse_each do |t|
-	    newtext << '</' << t << '>' # Add remaining tags to close
-	end
-
-	# WP fix for the bug with HTML comments
-	newtext.gsub!("< !--", "<!--")
-	newtext.gsub!("<    !--", "< !--")
-
-	return newtext
-    end
-end