@SmokeDetector v Something really weird going on here with the revision history, another person is suggesting vandalism edits and the OP is accepting them. I have flagged for mod attention
@Rob That particular watch is looking for the word "buy" relatively close to a link. Thus, IMO, the amount of information it's giving is probably OK. However, it should probably be shortened to only grab the beginning of the link, so other watches can detect the domain or URL path.
K, best to leave regex to someone other than me. :)
It seemed "fp-" on any part, but being so long didn't seem to benefit it's cause; especially if it could be split into a few parts and gain a few matches instead of one, or even just one match on one of the smaller pieces.
For example the full string is: buy NAS synology to store the firebird database.</p><p><a href="http://www.ibexpert.net/ibe/index.php?n=Doc.FirebirdPerformanceRecommendations" rel="nofollow noreferrer"> --- too bad we couldn't use the full string IF and AND the beginning and end were both listed --- another way of explaining, if "buy NAS Synology" AND ibexpert.net are not both watched then forget about it entirely. But I guess that's tough to regex.
@SmokeDetector Post has been deleted, modflag marked helpful, other user's reputation has been dramatically reduced suggesting they were systematically rep farming through suggested edits
@TetsuyaYamamoto 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.
Just a feature-request: could "potentially bad keyword" message when doing !!/test command includes line number and/or username who watched the pattern from watched_keywords.txt?
@TetsuyaYamamoto Once you know the exact pattern, you can do a search in CHQ to, usually, find who added it to the blacklist/watchlist. While you can also use git blame, or just look in the watched_keywords.txt file, I find that reasonably often there's some additional context around when it was added that's sometimes useful to see.
A "search all rooms where commands are accepted from" command might be helpful - example: !!/searchall <text> (or <regex> ?) --- especially if it opened in a new tab and didn't flood the CHQ Chat with the results (in case there were too many hits).
That would be difficult to implement - SE Chat doesn't have an API, so we would need to scrape the webpages (which is something you don't want to be doing wherever possible), and Smokey can't control the user's browser so you can't open a new tab
Even if it were easy to implement, I suspect that it wouldn't be used often enough to give us a ROI
@angussidney You can get a list of messages from one of the Chat endpoints, up to 500 messages/request from a single room, backwards from message N. It'd be a lot faster to use the chat search interface to search on all three servers (text, not regex). I actually already have code that does something close, as part of implementing an OR operator for chat search in the beta URRS. However, it relies on the searches being in the same domain. I agree it really doesn't have the ROI to be worth the time.
If you're wanting to actually do a regex search in a single room using the 500 messages/request Chat endpoint, the SOCVR Archiver does that, along with being able to specify that it's written by a specific user, or a reply to a specific user, etc. There's no UI written for it, yet, to allow arbitrary searches, but it's trivial to change the search parameters in the code. I've used it a few times that way when looking for things. It's far slower than a standard chat search, if you can use that.
@TetsuyaYamamoto That pattern looks like it's already caught by Blacklisted website in body, Blacklisted website in answer, Potentially bad keyword in body, and Potentially bad keyword in answer; append -force if you really want to do that.
[ SmokeDetector | MS ] Url in title, bad keyword in body, bad keyword in title, blacklisted website in body, blacklisted website in title, +6 more (787): topdiethub.com/brainfx/ by user132357 on graphicdesign.SE
@SmokeDetector not sure how to classify this one; the facebook link is legit, but the answer is very similar to (almost plagiarises) the existing answer
[ SmokeDetector | MS ] Bad keyword in body, bad keyword in title, bad keyword in username, blacklisted website in body, body starts with title and ends in url, +3 more (683): Mira essence cream canada by shred t3x on meta.SE
@Makyen 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.
[ SmokeDetector | MS ] Url in title, blacklisted website in body, blacklisted website in title, body starts with title and ends in url, link following arrow in body, +2 more (688): healthytalkz.com/mx-male/ by hannamontany on apple.SE
[ SmokeDetector | MS ] Bad keyword in body, bad pattern in url body, blacklisted website in body, pattern-matching product name in body, potentially bad keyword in body, +1 more (490): Makes Your Pumps Harder And Perpetual by loramartin on astronomy.SE
[ SmokeDetector | MS ] Url in title, blacklisted website in body, blacklisted website in title, body starts with title and ends in url, potentially bad keyword in body, +3 more (465): www.jbngroups.com by walker halker on graphicdesign.SE
[ SmokeDetector | MS ] Pattern-matching website in answer, potentially bad ns for domain in answer, potentially bad keyword in answer (82): PDF to HTML Mac by jones123 on stackoverflow.com
> Potentially bad ns for domain in answer, potentially bad keyword in answer ---------- Body - toolsbaer.com NS suspicious dns1.namecheaphosting.com.,dns2.namecheaphosting.com. Body - Position 0-13: toolsbaer.com
^ That NS has a note about significant FPs, but that wasn't one of them; so important to add additional confirmation. Note is: " This was briefly watched but there is a moderate number of false positives and we already have heavy measures in place against Namecheap; so we want to avoid adding so much weight that false positives will be autoflagged.".
@SmokeDetector naa- Recommends "Trial-ware" with limitations, rather than freeware or something without limitations. Not a good recommendation so not a good answer, but due to the nature of the question (software recommendation) it's not quite spam.
@doppelgreener Chris might be curious, but is more likely busy with other more important matters. If they have spare time they might see it, if Chris (the Moderator) has anything else to do they probably don't need another notification in their inbox. Why the Moderator doesn't know that the post was deleted (along with the user) on their own site is unclear; I'm neither curious nor do I need a notification. It seemed a courtesy to mention that in case something was broken on their end.
@Rob I deleted this user, and the post was marked as spam by some of the users, but not deleted as spam. The mod who deleted it simply hit the delete below the posting.
This was also the reason why I reported the URL. This time I didn't check first if there was already a report :-)
@TetsuyaYamamoto That pattern looks like it's already caught by Potentially bad keyword in answer and Potentially bad keyword in body; append -force if you really want to do that.
@Rob Speaking as a moderator myself (though I'm not Chris): if there's something worth hearing about I prefer a ping, I can work through my inbox contents. Being busy means I may entirely miss individual messages in transcripts. :P
In case it's relevant, myself I usually report spam here that I see before also deleting it.
@iBug MS seems to be running mighty slow, but working. The Flagging logs other seems to have a log of duplicate entries marked both success and fail. For example: @NobodyNada was rated both ways for this post.
@doppelgreener I'll be certain to ping you. In this case it was somehow being reported an hour after both the post and user were deleted. It wasn't clear that there was something to fix or any emergency, nor did I want to put this under the microscope. I make certain that I've been awake for an hour before making critical decisions, and do sleepy posting on The Tavern. :D
@Asuna and iBug - MetaSmoke is running fast, I don't know if Asuna's actions fixed it. The duplicated entries with opposing outcomes remain, going back 17 days.
