@Rob For feedback, not comments, you can also use the shortcut commands, which you might find easier on mobile (i.e. you don't have to actually click on the reply arrow).
@Makyen - Is it preferable to use the alias: k, spam, rude abuse, etc. as applicable instead of using tpu- to be specific, or does SD simply change all those to tpu- and the extra effort is thrown away?
@Rob For feedback, use what you feel works for you. Ultimately, the aliases get translated to the tp or tpu- form, but the aliases may be more descriptive for yourself or others. For actual flags it makes some minor difference between rude/abusive and spam, but they are usually considered interchangeable (on most sites).
@StephenKennedy You have different feedback on different versions of this post. The versions don't appear to differ sufficiently to justify different feedback (you have 2 FP and 1 TP). It's possible that one or or more of your feedbacks isn't what you want it to be.
@SmokeDetector This user posted 3 answers suggesting/linking to the same extension in about 7 minutes (all their existing answers). Of those, this one is actually spam in it's 2nd paragraph. The first paragraph might be an answer. I've custom mod-flagged that one. This one, however, appears to potentially give the question OP what they are looking for.
@Makyen - I made a similar complaint about user10175922 on this Q&A yesterday on SO. I explained in a custom Mod message that they wrote one answer with some URLs and copied the same answer to whatever relevant questions they could find. I explained that I didn't simply flag it spam or downvote it because it might be disputed (if people didn't
bother to check the user's posting history). When I checked back shortly thereafter the post was deleted and the account spam-nuked. I explained that my accusation was self-plagiarism and apparently the complaint was well received. I simply suggested a warning but they weren't feeling charitable. I gave links since I noticed you are 20K there - for lower Rep users there's nothing left to see.
@Rob Thanks for the link. IMO, that was a good use of a custom mod-flag. I recall seeing that post when going through the backlog of MS posts which need reviewing. The moderator deleted that post and the others as spam and destroyed the user, probably due to the 6 posts that day that were spamming for that site.
I guess it's thought by some that it's a good idea to make a short soft-of helpful answer that might score a downvote (and maybe an upvote, which more than makes up for it) and use Search or Tags to find as many questions as you can possibly copy it to. Viewed individually each is just no all that great an answer, and being kind we don't downvote the new user - being on-topic and seemingly attempting to help we don't post a comment. --- He lost his account with one downvote and (AFAIK)
one flag.
* just not all
Maybe SD could/should look for (80%) duplicate posts from very low Rep users posted within the hour. IF we only cached posts after they failed one other test here it wouldn't be too memory hogging (except on particularly busy days). Caching everything for a couple of hours (what we really would like to do) probably would eat way to much memory.
@SmokeDetector - The post is probably the best example of an edge case for spam. The Doctor wants to know how to add autoplay videos to his website and gives his (seemingly legitimate) website address. IF he continues posting his website into fair questions it would be suspicious.
@Makyen - Obsessively monitoring the Status page shows that almost always there's a ping in less than 5 seconds, rarely 1 minute, most rarely it will go up to 2 minutes. According to the Docsif it can make it to 3 minutes: "... failover standby instances from the status page. Normally metasmoke handles this automatically if an instance goes down for more than 3 minutes, but ...".
@Rob Yes. However, the instance that is newly coming online should post a "failover" message. For those times I'm concerned with the instances coming out of standby, there was no such message.
@Rob The pings are independent wrt. each instance. Each instance should ping MS every minute. If MS doesn't get a ping from the active instance within some period of time (I'd have to check the code, but probably 2 or 3 minutes), then MS will tell the next instance that pings to come online (i.e. failover to that instance).
They seem to usually ping almost constantly. Sometimes it creeps past two minutes while the others happily ping away. It's like some are very busy sometimes yet all presumably get the same load (though I suspect that the CPUs and memory are different). It didn't seem to do this a couple of days ago ... A recent code modification doesn't work as well everywhere?. But that was the long and short of what little I was able to offer. Annoying to check 5 logs to see if they match, should be
relatively silent until they kick out, and all sharing the same code they should fail-alike (subject to CPU power / disk full / memory swapping, etc.).
Good luck - maybe someone will come online and recognize this problem.
Is MS's 'ping input' getting a DOS Attack which deafends it, causing it to throw the next one into action ...
@Rob In general, only one of them should be active at any one time. The others should be in standby, doing almost nothing other than pinging MS on a regular basis.
@Rob That's a Stack Exchange thing. Chat messages that have been flagged as spam/rude/abusive are shown to all users on that chat server who have more than 10k reputation. The user's are asked to either validate or invalidate the flag.
OK, but I m only a few hundred on Mathematics and I don't know the context - I do trust that if he's out of line with his firends that someone at math.SE will give him an hour off.
@Rob The notification is shown to all users on the chat server who have more than 10k reputation. So, "not sure" just means you do nothing and are not notified about that one again.
@Rob If you click on the "Mathematics" link in that popup it will open a transcript page showing the context.
My notification is gone now. I recall a Manager ALL CAPSing someone and it was flagged - well the review could only be handled by a limited number of people and the person called for the flaggers head on a plate - booting them and later apologizing. So it's good to be there, know the people, read the context. etc. First time that's ever come to me. I'll be better prepared next time and head out quicker.
The conversation still comes up in The Tavern a month latter.
@Rob There are flagged messages that are obvious and there are ones that require knowing the context. The general advice on meta is that such messages are often evaluated without context and that they are often marked as spam/R/A if they sound that way out of context. So, people should take a look at what they write and make sure it doesn't look bad when taken out of context.
@Rob Yes, I'd noticed when trying to track down your user ID numbers (prior to asking). I'd considered asking you if you were shown these notifications, because I was interested in finding out if it used cumulative rep, or required 10k on one site. Now we know. :-)
I'd be careful flagging spammy links as NAA (when they're not spam) on SO. I've had 2 flags declined in the last week because a mod mass declined all the flags instead of the spam flags. And I had a 3rd flag declined in the last week on a "thanks" answer. Just fyi.
@K.Dᴀᴠɪs - I drew attention to roboreviewing earlier today over there. A comment thread where coincidentally Makyen had visited. --- I got rejected for splitting a wall of text into paragraphs. Things are sometimes a little different on different sites. I'm on a few dozen sites regularly, got to keep the rules straight when I go visit someone's house.
It would be great if "declined" was removed if the post is removed. It's not only fairer but it's tough to remember to go back and check all your flags and dispute what you can not see.
I don't mind if the flag was a legitimate decline, but when you get 3 in a week because someone is going too fast it kinda stings a bit. The issue is that declines can sometimes carry real penalties to users who dedicate their time to keep SO clean.
Mod delete and comment on 1st and 2nd To the flaggers here, my apologies - I intended to decline one particular flag, not all. – Rob♦ yesterday @K.Dᴀᴠɪs
@Rob I left a couple of comments after yours on that meta. Basically, declining that edit was the right thing, from the site's point of view, even though it would have improved the post. Approving your edit at that time would have put the question in the reopen queue without making the question on-topic. A question only gets one entry-by edit into the reopen queue. Ideally that entry-by-edit would be saved for edits by the OP, but it's not.
Any edit to the post body that's made by a user that hasn't either voted to close or flagged the question puts the question in the reopen queue. A custom edit-reject message would have been more appropriate there, to explain, briefly, why it was being rejected.
I believe that the system offers the warning unless you have very low numbers and the hysteresis range would be too thin to squeeze one in. For example if I answer a question and it starts to accumulate downvotes I can delete it and get my points back - but I answer far more than I ask and are generally useful around here (so I have some leeway). A new user whom deletes Q's or A's gets a time-out.
[ SmokeDetector | MS ] Bad keyword in body, blacklisted website in body, link at end of body, potentially bad ns for domain in body, potentially bad keyword in body, +1 more (397): If like many of us you by user434809 on meta.SE
[ SmokeDetector | MS ] Bad keyword in body, bad keyword in title, body starts with title and ends in url, pattern-matching website in body, potentially bad keyword in body, +1 more (392): Teal Farms Keto This popular by disharose on graphicdesign.SE
[ SmokeDetector | MS ] Bad ns for domain in body, bad keyword in body, bad keyword in title, blacklisted website in body, body starts with title and ends in url, +6 more (887): Forskolin Keto Cycle by Vghyu Oiuyj on askubuntu.com
Does using "why?" to SD gets a response only in some specific rooms. I have just tried it in another chatroom (where SD is still displayed among users) but I got no response.
[ SmokeDetector | MS ] Bad keyword in body, blacklisted website in body, body starts with title and ends in url, pattern-matching website in body, potentially bad asn for hostname in body (392): We could not find out any by user155576 on wordpress.SE
!!/unwatch ^\s*<p>\s*?<a href="(?!(?:[a-z]+:)?//(?:[^/.]+\.)*(?:(?:example|apple|google|github|i\.stack\.imgur|stackexchange|stackoverflow|serverfault|superuser|askubuntu|stackapps)\.com|mathoverflow\.net|github\.io|edu|wikipedia\.org)[/"])\W*(?#Post begins with link; test for potential new detection; at watch 12649/12409/175/73)
@Makyen No such item ^\s*<p>\s*?<a href="(?!(?:[a-z]+:)?//(?:[^/.]+\.)*(?:(?:example|apple|google|github|i\.stack\.imgur|stackexchange|stackoverflow|serverfault|superuser|askubuntu|stackapps)\.com|mathoverflow\.net|github\.io|edu|wikipedia\.org)[/"])\W*(?#Post begins with link; test for potential new detection; at watch 12649/12409/175/73) in watchlist.
@Makyen No such item ^\s*<p>\s*?<a href="(?!(?:[a-z]+:)?//(?:[^/.]+\.)*(?:(?:example|apple|google|github|i\.stack\.imgur|stackexchange|stackoverflow|serverfault|superuser|askubuntu|stackapps)\.com|mathoverflow\.net|github\.io|edu|wikipedia\.org)[/"])\W*(?#Post begins with link; test for potential new detection; at watch 12649/12409/175/73) in watchlist.
!!/unwatch ^\s*<p>\s*?<a href="(?!(?:[a-z]+:)?//(?:[^/.]+\.)*(?:(?:example|apple|google|github|i\.stack\.imgur|stackexchange|stackoverflow|serverfault|superuser|askubuntu|stackapps)\.com|mathoverflow\.net|github\.io|edu|wikipedia\.org)[/"])\W*(?#Post begins with link; test for potential new detection; at watch 12649/12409/175/73)
@Makyen No such item ^\s*<p>\s*?<a href="(?!(?:[a-z]+:)?//(?:[^/.]+\.)*(?:(?:example|apple|google|github|i\.stack\.imgur|stackexchange|stackoverflow|serverfault|superuser|askubuntu|stackapps)\.com|mathoverflow\.net|github\.io|edu|wikipedia\.org)[/"])\W*(?#Post begins with link; test for potential new detection; at watch 12649/12409/175/73) in watchlist.
!!/watch ^\s*<p>\s*(?:</?\w+/?>\s*)*<a href="(?:(?:[a-z]+:)?//(?:[^" >/.]*\.)*(?:youtube\.com|youtu\.be)[/"])\W*(?![\W\w]*?</(?:code|blockquote)>)(?#Separate start with youtube watch; lower TP rate, but most TP are not otherwise detected)
Based on what I saw in Charcoal Test, I suspect that it was erroring-out when attempting to post a started message or a response to !!/rev when commit b348201 was the most current. This is potentially the result of the watch text being too long.
> Pattern-matching website in body, pattern-matching website in title, pattern-matching website in username, potentially bad ns for domain in body, potentially bad ns for domain in title ---------- Title - luntrimdiet.com NS suspicious dns2.namecheaphosting.com.,dns1.namecheaphosting.com. Title - Position 7-15: diet.com Username - Position 7-15: diet.com Body - luntrimdiet.com NS suspicious dns2.namecheaphosting.com.,dns1.namecheaphosting.com. Body - Position 7-15: diet.com
@TetsuyaYamamoto 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.
@TetsuyaYamamoto That pattern looks like it's already caught by Bad keyword in answer and Bad keyword in body; append -force if you really want to do that.
> Blacklisted website in body, blacklisted website in title ---------- Title - Position 0-21: flashcomindonesia.com Body - Position 0-21: flashcomindonesia.com
When possible, I'd appreciate someone with write access to SD's GitHub repo (e.g. @iBug @ArtOfCode) merging PR #2566. The PR (and already added replacement !!/watch) should significantly reduce the number of FP reports we're getting on one of my recent watches. For some reason, SD claims that it doesn't exist when I tried to !!/unwatch it. Thus, this PR to remove it.
vote: give Mak push access to repo? although he doesn't code Python, Mak's knowledge in regex is reliable. I promote this so we have another helping hand with our blacklists and relevant PRs (cc @Undo @ArtOfCode)
In case anyone needs to know this later, there's a CI check for blacklist file format. To be brief, blacklist items should not duplicate within the same file or across files (PCRE comment stripped), and blacklist files should be in Unix format (LF newlines and one trailing newline at end of file). (cc @Makyen)
Thanks. I wasn't aware of the line-ending check, but would have just continued with what already exists. I'm glad to hear that the PCRE comments are stripped for the integrity testing. It's good to know the other fields in the watchlists aren't actually used. I'd gotten that impression, but hadn't actually exhaustively checked. Personally, I'm glad they are there for information, as it's information I've used on several occasions.
