@EliahKagan The regex contains an unescaped "."; in most cases, it should be "\."; append -force if you really want to do that.
@EliahKagan That pattern looks like it's already caught by Potentially bad keyword in body and Potentially bad keyword in answer; append -force if you really want to do that.
@EliahKagan That pattern looks like it's already caught by Potentially bad keyword in body and Potentially bad keyword in answer; append -force if you really want to do that.
Yeah the regex is good. The only potential improvement I could see would be capturing all the repetition with one pattern; this regex would catch each pair as a separate instance. Which isn't a bad thing, but it's a little more verbose
@Das_Geek Actually, you're right. Then also, if they keep at it, and iwe blacklist it, there's no need for huge scores to be generated. I'll replace the regex. So far most of the repeated-pillow posts have a trailing space... but not the most recent one. I'll write a slightly less elegant regex that matches even if they don't.
One of their troll posts with the word only has one occurrence. But matching that would have a lot of false positives, even if it had to be the misspelling.
@Das_Geek That pattern looks like it's already caught by Pattern-matching email in answer and Pattern-matching email in body; append -force if you really want to do that.
@SmokeDetector fp- User has two posts with debatably sealion-y wording. Gonna fp this though, as it's not quite enough for me to go tp, and the mods have already addressed the user.
It looks like Osiris disappeared for about 15 minutes, but it's back now. @ThomasWard I don't expect you to do anything about it at this point. This is just a note to mark the occurrence in case it repeats.
@ThomasWard I only saw it at the tail end of the time Osiris was not pinging, so didn't get a chance to take a detailed look, but it looked like MS was OK during that time. The SmokeDetector Status userscript showed Osiris as red (not responding for 14 minutes), while showing the other instances as good (status updated within last 2 minutes). However, it's possible that was merely a coincidence of me seeing it after MS got better and the only SD instanace which hadn't updated yet was Osiris.
@Das_Geek That pattern looks like it's already caught by Pattern-matching website in body and Pattern-matching website in answer; append -force if you really want to do that.
I'll try reworking Sandy in the next day or two. Meanwhile, Coral will remain available (believed to be stable) for an extended period. I'm not expecting any outage for Coral in the near future.
@SmokeDetector This individual post has been caught so many times now
@Machavity What do you figure for unwatching the domain in that last report? The user learned their mistake and hasn't done anything bad since, it's the only thing Smokey is catching on the post, and that individual post has been caught soo many time now and fp-ed
I would probably do it, but it'd be good to have a second opinion
When I click the MS link on the chat message it goes to the disambiguation page. Might it be considered expected behavior to show feedback for everything linked from that page, i.e., all versions of the post in cases where the MS link goes to the disambiguation page?
@Das_Geek But if this is isolated or not adequately explained, I think there's something else that might have contributed. I didn't misclic, but& I did do *something weird. Although I'm sure the only feedback that of mine that ever went onto that post was fp, I did also end up viewing the original MS entry for the post (not due to a bug, but due to clicking the top entry instead of the bottom one).
That one, which captured it before the mod edit, was tp, and I clicked tp. Doing so didn't add any new feedback to it--I assume because it's too old to accept any more non-conflicting feedback--and my feedback history on MS never showed anything for that. But maybe it was momentarily accessible from the MS API?
@EliahKagan That is, I didn't do anything that could give tp feedback on the recent report, but I did do something that would ordinarily have given tp feedback on the oldest report. I was going to say that's what I thought caused the weird effect... except that feedback was never actually applied to the old report.
@EliahKagan @Das_Geek When there are multiple MS posts for an SE post, AIM makes no attempt to accurately determine which MS post report corresponds to the SD chat report which you are viewing. In general, it will use feedback from all of them, with only one feedback counting from each individual. It's a known issue.
@Das_Geek np. It is something I've worked on, mostly in FIRE, with the intent to copy/share with AIM, but haven't gotten it quite working yet. It's mostly working in testing. :-)
@Das_Geek Historically, that regex has about 27.5% TP. The regex I removed has in the ballpark of an order of magnitude more hits on SE, which would drastically reduce the %TP number.
@Makyen Sure, I wasn't arguing for its reinstatement nor was I trying to be combative. I asked mainly because I used that regex's existence on the watchlist as justification for adding the one that got removed. I 100% agree with your reasoning; I should have been more careful and looked at the SE results before I added that pattern
@Das_Geek I'd seen your comment in SOCVR and was looking into it. I agree that it shouldn't be on the blacklist. At most, it should be a watch. Feel free to move it.
@ThomasWard It has been smooth, which is a testament to the good work of everyone involved. I'm sure Undo is happy about it too. I know it's easy to focus on the problems which crop up/still exist, but there's a whole bunch of stuff that could have been issues. Sure, there are some things which were problems, and a few that still should be resolved, but that's to be expected. Personally, I feel it went/is going well.
@Makyen I don't disagree. The big part of things was getting outside communication working without issue, and it's been humming along well with minimal interference. And the certs autorenew without any type of intervention on our part, which is also a plus. AND we have the future capacity to load-balance or failover between MS instances if we have to
@Makyen Just making sure. I am neither quick-witted nor quick-fingered enough to sustain any form of disagreement online, so I always try to avoid any negativity where possible
