use hardcoded cache size instead of len(blog)

It seems the blog posts are not rotated strictly chronologically,
so sometimes previously already posted things got posted again.

scrape.py

@@ -176,9 +176,7 @@ async def main():
                     pass
 
             # trim the cache
-            # len(blog) is usually bigger than the count of posts in our category,
-            # so with len(blog) instead of the latter we have some buffer
-            while len(cache) > len(blog):
+            while len(cache) > 100:
                 cache.remove(cache[0])
 
             # set new cache event