From d8fce1ceebb26a2e8cd70c83aa66cade443723cc Mon Sep 17 00:00:00 2001 From: Bastian Kleineidam Date: Sun, 2 Sep 2012 22:32:14 +0200 Subject: [PATCH] Do not sort URL queue anymore. --- linkcheck/cache/urlqueue.py | 23 ----------------------- 1 file changed, 23 deletions(-) diff --git a/linkcheck/cache/urlqueue.py b/linkcheck/cache/urlqueue.py index 01485c60..60ed89f4 100644 --- a/linkcheck/cache/urlqueue.py +++ b/linkcheck/cache/urlqueue.py @@ -56,7 +56,6 @@ class UrlQueue (object): self.in_progress = {} self.checked = LFUCache(size=100000) self.shutdown = False - self.unsorted = 0 def qsize (self): """Return the approximate size of the queue (not reliable!).""" @@ -146,10 +145,6 @@ class UrlQueue (object): self.queue.appendleft(url_data) else: self.queue.append(url_data) - self.unsorted += 1 - if self.unsorted > 2000: - self._sort() - self.unsorted = 0 self.unfinished_tasks += 1 def task_done (self, url_data): @@ -199,24 +194,6 @@ class UrlQueue (object): log.debug(LOG_CACHE, "Caching alias %r", key) self.checked[key] = data - def _sort (self): - """Sort URL queue by putting all cached URLs at the beginning.""" - newqueue = collections.deque() - while self.queue: - url_data = self.queue.popleft() - key = url_data.cache_url_key - if url_data.has_result: - # Already checked and copied from cache. - newqueue.appendleft(url_data) - elif key in self.checked: - # Already checked; copy result. And even ignore - # the case where url happens to be in_progress. - url_data.copy_from_cache(self.checked[key]) - newqueue.appendleft(url_data) - else: - newqueue.append(url_data) - self.queue = newqueue - def join (self, timeout=None): """Blocks until all items in the Queue have been gotten and processed.