[09:31:24] AttributeError: 'NoneType' object has no attribute 'json'
2017-12-12 09:31:24.182423 UTC
File "/home/ubuntu/charcoal/SmokeDetector/excepthook.py", line 54, in run_with_except_hook
run_old(*args, **kw)
File "/usr/lib/python3.5/threading.py", line 862, in run
self._target(*self._args, **self._kwargs)
File "/home/ubuntu/charcoal/SmokeDetector/chatcommunicate.py", line 145, in send_messages
response = room.room._client._do_action_despite_throttling(("send", room.room.id, msg)).json()
[ SmokeDetector | MS ] Bad NS for domain in body, bad keyword in body, blacklisted website in body, link at end of body, repeated URL at end of long post: of the quirk of the decision by Say ou on drupal.SE
[ SmokeDetector | MS ] Bad keyword in body, bad keyword in title, blacklisted website in body, pattern-matching website in body: /trembolex-ultra/ by BernieMorton on astronomy.SE
[ SmokeDetector | MS ] Bad keyword in body, blacklisted website in body, link at end of body, pattern-matching website in body, repeated URL at end of long post: the all over but the shouting by Mord ou on drupal.SE
CI on f753c22 succeeded. Message contains 'autopull', pulling...
Restart: API quota is 19082.
[ SmokeDetector | MS ] Bad keyword in answer, bad keyword with email in answer, pattern-matching email in answer, potentially bad keyword in answer, blacklisted user: In-Memory Debugger for .NET? by Scott Witcher on softwareengineering.SE
Note for moderators: Some sites, including Stack Overflow, have more permissive, overriding guidance on the handling of these flags.
What makes something spam and when should I flag it?
A post should be marked as spam only if it advertises a product, service, or similar and is unsolicited or la...
> Abuse of the system or community is everything that is created with the intention to harm them. This includes posts that contain no useful content at all – i.e. gibberish posts along the lines of:
@quartata While you might solve this by not permitting .py to be a TLD, the more general case is to not trigger on displayed text that matches the filename being linked. Eliminating such would prevent a significant percentage of the the FPs we've seen from the filter (at least those seen on SO). People shortening the link text down to just the base filename is common. Such links should not be caught by the filter when the filename happens to be something that also has the format of a domain.
@SmokeDetector It seems like that reason gets a lot of fp. Mostly because of people linking a file. Would it be a good idea to disable that for common file extensions?
@Riker Yes, it is, currently. However, a large percentage of the FPs are from a relatively low number of causes: A) link text that matches the filename being linked; B) Links to known domains (e.g. stack.imgur.com, with a valid image; etc.). Yes, if the FP rate can't be lowered, then the filter should be removed. However, when something like this is added, I'd expect a period of adjusting the filter. OTOH, those two don't cover all of the FPs, so adding them won't reduce FP to 0.
@quartata Hmmm... It looks like a more generalized rejection than just matching the file name would be desirable. Perhaps a more general rejection would be: does the link text match the filename or #fragment. In addition, another rejection should be links to URLs which are not valid (e.g. http://here and http://link).
