More Regexp madness. Fixes, optimizations, and a new method to parse channel lists

diff --git a/ChangeLog b/ChangeLog
index 4655e5ed669dbce9543e9bb0c2a382467be19099..5680f0dfa39ce5e7e32623888535f471f8b6d5ba 100644 (file)
--- a/ChangeLog
+++ b/ChangeLog
@@ -2,6 +2,9 @@
 
        * Plugin: new unreplied() method to handle PRIVMSGs which have not
        been replied to.
+       * Regexp: more regexp madness. BasicUserMessage and derivate classes
+       now have a new method parse_channel_list() to parse list of channel
+       names.
 
 2007-02-12  Giuseppe Bilotta <giuseppe.bilotta@gmail.com>
 

diff --git a/lib/rbot/core/utils/extends.rb b/lib/rbot/core/utils/extends.rb
index 1f33026962c1de64639bff59efbe79fa0f3f5d7f..b2b4f5951b69c869e27a4b4280edc1aa8fbcdbd5 100644 (file)
--- a/lib/rbot/core/utils/extends.rb
+++ b/lib/rbot/core/utils/extends.rb
@@ -66,17 +66,30 @@ class ::String
   end
 end
 
-# Extensions to the Regexp class, with some commonly used regular expressions.
+# Extensions to the Regexp class, with some common and/or complex regular
+# expressions.
 #
-module ::Rx
+class ::Regexp
+
+  # A method to build a regexp that matches a list of something separated by
+  # optional commas and/or the word "and", an optionally repeated prefix,
+  # and whitespace.
+  def Regexp.new_list(reg, pfx = "")
+    if pfx.kind_of?(String) and pfx.empty?
+      return %r(#{reg}(?:,?(?:\s+and)?\s+#{reg})*)
+    else
+      return %r(#{reg}(?:,?(?:\s+and)?(?:\s+#{pfx})?\s+#{reg})*)
+    end
+  end
 
+  IN_ON = /in|on/
 
   # We start with some IRC related regular expressions, used to match
   # Irc::User nicks and Irc::Channel names
   #
   # For each of them we define three versions of the regular expression:
-  #  * a generic one, which should match for any server but may turn out to match
-  #    more than a specific server would accept
+  #  * a generic one, which should match for any server but may turn out to
+  #    match more than a specific server would accept
   #  * an RFC-compliant matcher
   #  * TODO a server-specific one that uses the Irc::Server#supports method to build
   #    a matcher valid for a particular server.
@@ -85,27 +98,79 @@ module ::Rx
     CHAN_FIRST = /[#&+]/
     CHAN_SAFE = /![A-Z0-9]{5}/
     CHAN_ANY = /[^\x00\x07\x0A\x0D ,:]/
-    GENERIC_CHAN = /(?:#{CHAN_FIRST}|#{CHAN_SAFE})#{CHAN_ANY}+/
+    GEN_CHAN = /(?:#{CHAN_FIRST}|#{CHAN_SAFE})#{CHAN_ANY}+/
     RFC_CHAN = /#{CHAN_FIRST}#{CHAN_ANY}{1,49}|#{CHAN_SAFE}#{CHAN_ANY}{1,44}/
 
+    CHAN_LIST = Regexp.new_list(GEN_CHAN)
+
+    # Match "in #channel" or "on #channel" and/or "in private" (optionally
+    # shortened to "in pvt"), returning the channel name or the word 'private'
+    # or 'pvt' as capture
+    IN_CHAN = /#{IN_ON}\s+(#{GEN_CHAN})|(here)|/
+    IN_CHAN_PVT = /#{IN_CHAN}|in\s+(private|pvt)/
+
+    # As above, but with channel lists
+    IN_CHAN_LIST_SFX = Regexp.new_list(/#{GEN_CHAN}|here/, IN_ON)
+    IN_CHAN_LIST = /#{IN_ON}\s+#{IN_CHAN_LIST_SFX}|anywhere|everywhere/
+    IN_CHAN_LIST_PVT_SFX = Regexp.new_list(/#{GEN_CHAN}|here|private|pvt/, IN_ON)
+    IN_CHAN_LIST_PVT = /#{IN_ON}\s+#{IN_CHAN_LIST_PVT_SFX}|anywhere|everywhere/
+
     SPECIAL_CHAR = /[\x5b-\x60\x7b-\x7d]/
     NICK_FIRST = /#{SPECIAL_CHAR}|[[:alpha:]]/
     NICK_ANY = /#{SPECIAL_CHAR}|[[:alnum:]]|-/
-    GENERIC_NICK = /#{NICK_FIRST}#{NICK_ANY}+/
+    GEN_NICK = /#{NICK_FIRST}#{NICK_ANY}+/
     RFC_NICK = /#{NICK_FIRST}#{NICK_ANY}{0,8}/
-  end
 
+    # Match a list of nicknames separated by optional commas, whitespace and
+    # optionally the word "and"
+    NICK_LIST = Regexp.new_list(GEN_CHAN)
+
+  end
 
   # Next, some general purpose ones
-  DIGITS = /[0-9]+/
-  HEX_DIGIT = /[0-9A-Za-z]/
+  DIGITS = /\d+/
+  HEX_DIGIT = /[0-9A-Fa-f]/
   HEX_DIGITS = /#{HEX_DIGIT}+/
   HEX_OCTET = /#{HEX_DIGIT}#{HEX_DIGIT}?/
-  DEC_OCTET = /[01]?[0-9]?[0-9]|2[0-4][0-9]|25[0-5]/
+  DEC_OCTET = /[01]?\d?\d|2[0-4]\d|25[0-5]/
   DEC_IP = /#{DEC_OCTET}.#{DEC_OCTET}.#{DEC_OCTET}/
   HEX_IP = /#{HEX_OCTET}.#{HEX_OCTET}.#{HEX_OCTET}/
   IP = /#{DEC_IP}|#{HEX_IP}/
 
 end
 
-
+module ::Irc
+
+
+  class BasicUserMessage
+
+    # We extend the BasicUserMessage class with a method that parses a string
+    # which is a channel list as matched by IN_CHAN(_LIST) and co. The method
+    # returns an array of channel names, where 'private' or 'pvt' is replaced
+    # by the Symbol :"?", 'here' is replaced by the channel of the message or
+    # by :"?" (depending on whether the message target is the bot or a
+    # Channel), and 'anywhere' and 'everywhere' are replaced by Symbol :*
+    #
+    def parse_channel_list(string)
+      return [:*] if [:anywhere, :everywhere].include? string.to_sym
+      string.scan(
+      /(?:^|,?(?:\s+and)?\s+)(?:in|on\s+)?(#{Regexp::Irc::GEN_CHAN}|here|private|pvt)/
+                 ).map { |chan_ar|
+        chan = chan_ar.first
+        case chan.to_sym
+        when :private, :pvt
+          :"?"
+        when :here
+          case self.target
+          when Channel
+            self.target.name
+          else
+            :"?"
+          end
+        else
+          chan
+        end
+      }.uniq
+    end
+  end
+end