@SmokeDetector This is the user's 3rd spam post across the Stack Exchange Network. A) posts asking users to take off-site surveys are considered spam on Stack Exchange. B) hop.clickbank.net is an affiliate marketing program that pays the OP for people visiting the destination and buying.
@user12986714 This post [MS] (regex101) triggered "post is mostly images". It seems that accuracy of this detection might be better if code blocks are stripped, or at least the code not considered wrt. being an image.
BTW: Since the SSD migration completed, we have had a few posts which I expected to receive autoflags, but which have not, even after an hour. The flag_logs pages are 1, 2, 3, 4,
@ThomasWard I'd modified SD so that it would not stop trying to connect the the WebSocket. Thus, unless a !!/ms-down-force had been done each time SD rebooted, it would come all the way back up. I was watching for it, and it did. Given the reboots for watchlists, etc., the !!/ms-down-force you'd done earlier had been forgotten by SD.
The ones I linked to were reported by SD normally well after MS was back up.
@ThomasWard It shouldn't be. I'm not aware of any SE WebSocket dependency which MS has. SD has an SE WebSocket dependency, which we've seen be an issue over the last few/several days a couple of times. I also made a modification such that SD will auto-reboot much faster if it sees no scan activity. If there's no SE activity within 2 minutes, it will report it. It will then auto-reboot after 3 minutes of no scanning.
This may need to be adjusted. It may also be inconvenient during times SE is down, so we'll need to watch for that.
@user12986714 I was going to say make it a whole_post check and get the full post, which should have the code-stripped body too, but ... nope. Oh, ye gads... we freshly strip the code blocks for each detection that asks for it. That means we strip the code blocks nearly 40 times for every post. OK. I'll fix that. Oh... and strip them in a way that's going to be inaccurate some of the time. I'll fix that too.
So... in the not too distant future you can have it be a whole_post check and use both the normal body and the code-stripped body.
[ SmokeDetector | MS ] Blacklisted website in body, potentially bad ip for hostname in body, potentially bad ns for domain in body, potentially bad keyword in body, blacklisted user (173): Cách cài đặt Teatv trên iPad ✏️ by anabeo on apple.SE
@iBug I recall that you mentioned it. I'm assuming/hoping this is just getting all the paperwork in order and going through a process where you've already been approved/granted the position.
The internship lasts for the following year (until July 2021) which is my senior year (university grade 4). After that I'll start studying and working as a master student (will have completed my undergrad life by then).
@Makyen Per the general plan of my department, there's no course scheduled for the senior year, so more than half of my schoolmates have already planned for themselves, including internship, learning to drive automobiles, etc.
For a small portion who didn't complete all planned courses in the first three years, they have to complete them in the last year, of course.
The only mandatory task is the graduate project, which doesn't look like a course: No attendance required and (often) not even need to stay in school. The only time one's required to be present is the final review / answers.
I can't recall whether I've shared this before. My experience in the open-source world won me an invited speech at Microsoft Student Summer Camp last year in August (held by MSRA), where I spent 10 minutes talking about my participation here.
@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 ] Bad keyword in body, bad keyword in title, blacklisted website in body, link at beginning of body, pattern-matching product name in body, +1 more (433): What is Vital Synergy Keto? by Sophie Ross on meta.SE
> Blacklisted username, potentially bad keyword in body, potentially bad keyword in title, potentially bad keyword in username ---------- Blacklisted username - Position 0-18: Bike Rental in Goa Potentially bad keyword in body - Position 5-18: Rental in Goa Potentially bad keyword in title - Position 5-18: Rental in Goa Potentially bad keyword in username - Position 5-18: Rental in Goa
@Daniil I don't recall seeing a question about that from you. It's definitely not a question I'm shy about responding to. Clearly we failed to communicate. I'm sorry for my part of that.
@Daniil I'm not sure why you are using the (C|c). A) Please use non-capturing groups in watchlist and blacklist regexes (that would make this look like (?:C|c)). The capture information goes unused, so it's just additional CPU cycles spent tracking it. IME, it's about a 10% CPU savings to not use capture groups. B) (?:C|c) is normally written as a set, [Cc], rather than as a group. It's not required, but it is a style thing. I must admit that I have not done performance testing on it.
@Daniil Do I really need to do testing on every single different regex, or is it reasonable to generalize to a ballpark for what the effects will be based on an assortment of prior testing?
@Daniil This time, I missed that you has switched to asking about the (?:C|c) vs [Cc]difference, or at least I now think you were asking about. IME, the [Cc] tends to be faster, because it uses code that's optimized for a single character, whereas putting the same thing in a group uses code that's more general purpose, intended for an arbitrary regex, but it's also possible that the regex compiler detects the use of single characters and switches what it's using.
As to the general need for [Cc] to match both upper and lower case, most of the regexes we use have the case insensitive flag set, so [Cc] is redundant. However, there are some of our regexes where the case insensitive flag has not been set, so having something that accepts both upper and lower case would matter on those.
@SmokeDetector @Glorfindel I saw your edit suggestion on this question, but the question itself is off-topic and is now on CV review. I'm unsure to approve or reject...
