[bugzilla, github, reddit, youtube] Remove handlers on shutdown

This solves #615. When reloading these modules, a new compiled regex
object was created and inserted into the url handler set. Because it was
a new object, it didn't replace the old one (even though the pattern was
the same). This meant that every time the module was reloaded, another
version of the function was added, meaning the output would be produced
an additional time.

willie/modules/bugzilla.py

@@ -14,6 +14,8 @@
 from willie.module import rule
 
 
+regex = None
+
 def configure(config):
     """
 
@@ -30,6 +32,7 @@ def configure(config):
 
 
 def setup(bot):
+    global regex
     regexes = []
     if not (bot.config.has_option('bugzilla', 'domains')
             and bot.config.bugzilla.get_list('domains')):
@@ -45,6 +48,10 @@ def setup(bot):
     bot.memory['url_callbacks'][regex] = show_bug
 
 
+def shutdown(bot):
+    del bot.memory['url_callbacks'][regex]
+
+
 @rule(r'.*https?://(\S+?)'
        '(/show_bug.cgi\?\S*?)'
        '(id=\d+).*')

willie/modules/github.py

@@ -24,6 +24,7 @@
              '([A-z0-9\-]+/[A-z0-9\-]+)/'
              '(?:issues|pull)/'
              '([\d]+)')
+regex = re.compile(issueURL)
 
 def checkConfig(bot):
     if not bot.config.has_option('github', 'oauth_token') or not bot.config.has_option('github', 'repo'):
@@ -45,12 +46,17 @@ def configure(config):
         config.interactive_add('github', 'repo', 'Github repository', 'embolalia/willie')
     return chunk
 
+
 def setup(bot):
-    regex = re.compile(issueURL)
     if not bot.memory.contains('url_callbacks'):
         bot.memory['url_callbacks'] = tools.WillieMemory()
     bot.memory['url_callbacks'][regex] = issue_info
 
+
+def shutdown(bot):
+    del bot.memory['url_callbacks'][regex]
+
+
 @commands('makeissue', 'makebug')
 def issue(bot, trigger):
     """Create a GitHub issue, also known as a bug report. Syntax: .makeissue Title of the bug report"""

willie/modules/reddit-info.py → willie/modules/reddit.py

@@ -16,17 +16,22 @@
 domain = r'https?://(?:www\.|np\.)?reddit\.com'
 post_url = '(%s/r/.*?/comments/[\w-]+)' % domain
 user_url = '%s/u(ser)?/([\w-]+)' % domain
+post_regex = re.compile(post_url)
+user_regex = re.compile(user_url)
 
 
 def setup(bot):
-    post_regex = re.compile(post_url)
-    user_regex = re.compile(user_url)
     if not bot.memory.contains('url_callbacks'):
         bot.memory['url_callbacks'] = tools.WillieMemory()
     bot.memory['url_callbacks'][post_regex] = rpost_info
     bot.memory['url_callbacks'][user_regex] = redditor_info
 
 
+def shutdown(bot):
+    del bot.memory['url_callbacks'][post_regex]
+    del bot.memory['url_callbacks'][user_regex]
+
+
 @rule('.*%s.*' % post_url)
 def rpost_info(bot, trigger, match=None):
     r = praw.Reddit(user_agent='Willie IRC bot - see dft.ba/-williesource for more')

willie/modules/youtube.py

@@ -22,13 +22,19 @@
 else:
     from html.parser import HTMLParser
 
+regex = re.compile('(youtube.com/watch\S*v=|youtu.be/)([\w-]+)')
+
+
 def setup(bot):
-    regex = re.compile('(youtube.com/watch\S*v=|youtu.be/)([\w-]+)')
     if not bot.memory.contains('url_callbacks'):
         bot.memory['url_callbacks'] = tools.WillieMemory()
     bot.memory['url_callbacks'][regex] = ytinfo
 
 
+def shutdown(bot):
+    del bot.memory['url_callbacks'][regex]
+
+
 def ytget(bot, trigger, uri):
     bytes = web.get(uri)
     result = json.loads(bytes)