« first day (1645 days earlier)      last day (2185 days later) » 
00:00 - 09:0009:00 - 00:00

12:04 AM
@ArtOfCode If you could include time before, time after, and a delta, that'd be great. Bonus points if the line starts with the delta for easy reading.
 
I have a script that takes in a raw query_times.log and spits out a nicely formatted list like the one I posted earlier
so the plan is to take out the stuff from before the optimize, run it, then replace the stuff before and remove the stuff after and run again
 
If you could put the numbers first, that'd be extra sparkley. Otherwise, that sounds like it'll work.
 
that should be easy enough
 
Could you not do a a = rows.map do |r| JSON.parse(r) end; a.select { |e| # condidtions }; a.each { |e| puts "Something pretty" }
 
you also have to remove query strings and group :id based routes
and average the runtimes
 
12:08 AM
Oh, I forgot about that. I'll just let you do your thing.
 
After optimization:
32262.0655464: /api/posts/search/regex
28301.713785000025: /api/graphql
11504.304329000002: /reasons
3523.6828028000004: /sites/dash
2403.0355759999998: /graphs/af_accuracy
2173.9869645: /graphs/ttd
1915.420300414634: /spammers/site
1460.034583255814: /search
1286.8210834700442: /posts/uid/:api_param/:native_id
1174.4862132222222: /graphs/reports
986.090603: /graphs/report_counts
650.3141928125: /spammers/:id
499.49303400000025: /flagging/logs/unflagged
498.1048799450549: /
281.6319950833333: /posts.json
Before optimization:
32460.71397385904: /api/posts/search/regex
6334.821014822165: /search
3777.5436900500004: /reasons
3420.3947524083333: /flagging/conditions/:id/edit
3028.25373228: /sites/dash
2372.5605041999997: /spammers/site
2056.7701464090906: /github/pull_request_hook
1465.8562900000015: /flagging/run_ocs
1246.674546108108: /data/sql/queries/run
990.5861849166666: /flagging/conditions/:id/enable
931.7783069999998: /flagging/conditions/:id
846.3821730714285: /graphs/reports
715.9203569426752: /domains
681.3531042916667: /graphs/report_counts
 
Um...
Are you sure that's not backwards?
 
but there's just under 6x as much data for before as there is after, so the averages are a lot better on the before
yes, yes that is backwards
 
Oh good
So that actually made a difference
cc @Undo
@ArtOfCode Deltas? If it wouldn't be too much of a pain?
 
Deltas on... what? Each route between before and after?
 
12:12 AM
Yeah. Or I could just take your results and split on ": " and get the deltas myself.
 
not in the same order
stand by, shouldn't be terrible
 
@ArtOfCode rows.sort_by { |a| a[1] }
But better if you just do the magic
Because typing is a bit ouch because camlok and soca hate me. grumble grumble
 
4874.786431566351: /search
3420.3947524083333: /flagging/conditions/:id/edit
2055.8933652979795: /github/pull_request_hook
1465.8562900000015: /flagging/run_ocs
1216.109221108108: /data/sql/queries/run
990.5861849166666: /flagging/conditions/:id/enable
931.7783069999998: /flagging/conditions/:id
715.9203569426752: /domains
610.5159859999999: /flagging/conditions/preview
470.28305247368417: /admin/user_feedback
457.14020378536566: /spammers/site
357.60275928571434: /admin/invalidated
335.88466400000004: /review/history
 
The next thing I want to try is shrinking the mysql thread pool. @Undo how many non-MS applications use that mysql server?
 
did before-after because I can't math, so positive values are better
 
12:20 AM
Redunda, Sentinel
 
So with 5 connections (default rails connection pool size) you probobly want ~20 theads (5 for just-in-case)
 
One of those has an ActionCable attachment too, like metasmoke
 
That's nearly the default. Lemme check the docs...
 
Major gain on search and condition accuracy calculation, though
 
Default is 16, so that should work fine as-is.
@Undo You don't tend to be using much of your swap space do you?
So... the next thing would be looking at innodb_buffer_pool stuff, and that goes a little bit into mysql then my DBA knowledge.
Or running OPTIMIZE TABLES on more tables.
 
12:29 AM
I doubt you'll find enough stuff in most other tables for OPTIMIZE to have much of an impact
 
flag_logs maybe
Certainly can't hurt to try doing it to more tables... maybe all the tables?
 
A significant part of it is looking at the indexes, which are B trees, which don't really get deep enough to get particularly screwed until you're looking at hundreds of thousands of items
Flag logs potentially, feedbacks definitely
 
It also does compression stuff to make IO faster
 
feedbacks is approaching 400k IIRC
 
Feedbacks has already been optimized
Feedbacks and posts were the first 2 targets
 
12:33 AM
oh, fair
 
Honestly, it does no harm and only good things, so I'd just do it to every table because why not
Maybe even make it a weekly job
mysqlcheck -u <username> -p -o <database>
 
@Undo What is the InnoDB buffer pool size at the moment?
SELECT @@innodb_buffer_pool_size;
 
I'd assume it's the default?
BTW has any thought been given to migrating to postgres? It wouldn't be much faster, but it has many more sparklies.
 
ew
no
[status-declined]
We're so closely developed for MySQL that migrating would be far more effort than it's worth
and MySQL is so performance focused that it's kinda hard to beat, which is a good thing when you have a huge dataset
 
Last I read, preformace is pretty similar?
 
12:42 AM
Last time I used it it was pretty similar on small sets, absolutely no idea how it does with something this size
 
In the world of databases, I think ours is relatively small.
But regardless, it's be a pain to migrate. I was just curious if it'd be proposed in the past.
 
Once or twice maybe
but yeah, too much of a pain for not enough gain
 
In other news, tomorrow or monday I want to show y'all a thing I've been working on that relates to Andys proposal.
 
what kinda thing is this
 
A sort-of concept for a modular smokey.
I've been working on it for a while, just because I wanted to see if some ideas that I have would work, and I generally like decoupling stuff/modularification.
 
12:55 AM
aye aye
 
Also, it's ruby, and I think y'all want to stick with python. It's just an idea for how things might look.
 
I mean
I wouldn't object to Ruby... but I think @Andy might crucify us
 
1:21 AM
Maybe
 
Only a maybe? Better than going to C# then :)
 
We like our languages interpreted.
 
Ruby is pretty awesome for text processing, to be fair
and I love a good one-liner
 
And DSLs/metaprogramming
 
this is literally the last line I typed into the terminal
File.write('deltas.txt', qtb.map { |i| [i[0].to_f - (qta.select { |ai| ai[1] == i[1] }&.first&.first&.to_f || 0), i[1]] }.sort_by { |i| i[0].to_f }.reverse.map { |i| "#{i[0]}: #{i[1]}" }.join("\n"))
 
1:23 AM
Kazam
 
API quota rolled over with 11710 requests remaining. Current quota: 19999.
math: 345
stackoverflow: 162
askubuntu: 116
superuser: 68
physics: 63
english: 22
electronics: 22
apple: 22
serverfault: 21
unix: 20
graphicdesign: 20
security: 18
dba: 16
rpg: 15
mathoverflow.net: 14
codegolf: 14
webapps: 12
tex: 11
es.stackoverflow: 11
gis: 11
drupal: 10
android: 10
mathematica: 10
aviation: 9
ell: 9
writing: 8
wordpress: 8
engineering: 8
meta: 8
diy: 8
pt.stackoverflow: 7
travel: 6
dsp: 6
anime: 6
bitcoin: 6
academia: 6
gaming: 6
worldbuilding: 6
softwareengineering: 6
ethereum: 5
 
@ArtOfCode 134217728
@thesecretmaster Not on the current box. Currently 268/1023MB, and that's higher than normal.
 
RAM is 1344/2G
 
Have console access now?
 
1:31 AM
Aye
Not much time, few minutes
 
54 mins ago, by thesecretmaster
mysqlcheck -u <username> -p -o <database>
Optimizes all tables
Because why not
 
Pending
metasmoke.abuse_comments
note     : Table does not support optimize, doing recreate + analyze instead
status   : OK
metasmoke.abuse_contacts
note     : Table does not support optimize, doing recreate + analyze instead
status   : OK
metasmoke.abuse_report_statuses
note     : Table does not support optimize, doing recreate + analyze instead
status   : OK
metasmoke.abuse_reports
note     : Table does not support optimize, doing recreate + analyze instead
status   : OK
metasmoke.announcements
note     : Table does not support optimize, doing recreate + analyze instead
 
You saw the deltas Art got from query_times.log?
 
Yep
Looks like a decent improvement
Not an order of magnitude, but double digit percentages is nice.
Gotta run, ping with anything else
 
So, let's see if we can chop time off of that with these. Less likely, because these tables are orders of magnitiude smaller, but let's see in 24 hours.
 
1:36 AM
@ArtOfCode Sorry we're not doing Ruby
Leaving aside the fact that Ruby is a nightmare to optimize
 
eh
 
I think you all are over estimating how much will have to change here
 
Language migrations are big things, that wasn't exactly a serious remark
If we wanted to migrate away from Python then Ruby would be on the shortlist, but that's an if.
 
I suspect if we moved to anything it would be a compiled language. But. We're not doing that.
 
@quartata That's been my thought too. At its heart, this is really just pulling findspam out
 
1:38 AM
@Undo And bodyfetcher.
 
bodyfetcher is SE domain code, yeah
most of WS is too
 
I wouldn't object to doing a fairly serious re-architect of the internals while we're at it, though
as opposed to just ripping the files apart and adding a socket o.e. between them
 
Oh, sure. But MVP is just pulling a few things out of SmokeDetector into their own modules
 
aye
 
Thought I've had today: MS could have huge security gains if I threw together a quick external app to handle tokens.
 
1:41 AM
Like
 
metasmoke would then say "flag this with this account" and it does it, without metasmoke (and thus anyone writing code for metasmoke) remotely having access to raw tokens
 
findspam is the only thing that will be complete rewritten
and I'm already doing that a little bit anyways
The rest, and this is key to understand, is about identifying the implicit interfaces the other modules follow, and turning those into actual interfaces that our SE code is just a child class of
 
@Undo think there's a balance to be had there
 
For chatcommunicate + spamhandling which will probably be squished into one, this is really just these: init/parse_room_config, tell_channels, is_privileged, handle_report
 
Only person who could realistically get their hands on tokens is me. You could, in effect, get the same security gains just by saying "Art might have access to tokens too" on sign up.
 
1:44 AM
Bodyfetcher is more complicated. But I did a lot of the groundwork for this when thinking about Inferno. I think we can do it.
 
@ArtOfCode Actual maliciousness is only the tip of the iceberg (not terribly worried about that, you have other ways to be evil). There's unintentional stuff, too - remember when the new websocket implementation happily spewed out whatever you asked for?
We got lucky and no one tried that for three months, but crap happens.
Wouldn't need to worry about miniprofiler memory-dumping any token secrets, etc.
 
Agree, with the caveat that a situation in which metasmoke spews out plain text tokens seems pretty unlikely
 
Encrypted keys and the token are both in memory at the same time, and the implementation is conveniently public
So basically plaintext, with a step or two in between
 
I could see it spewing encrypted tokens because someone made a boo boo fairly easily, but to spew in plaintext you'd need to make a boo boo in the autoflagging functions; if you're touching those, multiple people are going to be reviewing the code
Your changes to autoflagging last night I pulled down and ran to make sure logging hadn't changed
 
Anyway, it's a thing I thought about today. Probably won't happen. But getting tokens to a happier place would definitely be a minor win.
data dumps become less scary
 
1:48 AM
Aye
Questions about how to do it best, as usual, but aye
 
heck, that's a big step to just giving you SSH keys
afk
 
@Undo Make it a Lambda.
 
well that could have been bad. Discovered the fan on my laptop wasn't working proper heh
(had to take the thing apart to fix it, but it's working now...)
 
@ThomasWard O_o
 
yep i know right?
fireball waiting to happen without a fan
 
1:54 AM
literally, considering how evil your stuff is
 
the age of the laptop is partly involved, though, this is a four year old laptop
in this case, it was excess dust in the fan ventport so
still had to take this thing apart
but in the process, I fixed something that had stopped working, the front-panel lights were not working at one point, after this got a repair, so they work now yay!
and the fan works now too, but yeah literal fireball waiting to happen basically
 
Imagine if you had wanted to do some e.g. ML training and thought "I'll let it run overnight"
 
I'm sure all kinds of fun things happen when GPUs overheat
@Andy @Undo I think we should actualy start with the output/soamhandling part first. Starting with findspam/bodyfetcher requires doing both, whereas we can leave those unmodified while changing chatcommunicate to do whatever
 
[ SmokeDetector | MS ] Few unique characters in answer, repeating characters in answer: Can I grow an apple tree by planting an apple from a nearby tree? by Roxanne Latham on gardening.SE
 
@quartata "fun things" are also known as "fire"
 
2:00 AM
tpu- by NobodyNada
 
Of course I need to do asyncio first
 
2:25 AM
[ SmokeDetector | MS ] Bad keyword in answer, bad keyword in username: Spasi yang digunakan dalam C# by Sattamatkapro.mobi on stackoverflow.com
[ SmokeDetector | MS ] Potentially bad keyword in body: Block all EU visitors due to GDPR by ndk on stackoverflow.com
tpu- by NobodyNada on Spasi yang digunakan dalam C# [MS]
 
@SmokeDetector Rolled back rude edit
@SmokeDetector tp-
!!/watch sattamatkapro\.mobi
 
@NobodyNada 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.
 
!!/test sattamatkapro.mobi
 
> Bad keyword in body, bad keyword in title, bad keyword in username
----------
Title - Position 1-11: sattamatka
Body - Position 1-11: sattamatka
Username - Position 1-11: sattamatka
 
hmm, is that NWB?
 
2:31 AM
@NobodyNada NWB?
 
@AshishAhuja It doesn't require a word boundary
 
oh, okay.
 
bad_keywords.txt requires a word boundary before and after the keyword, but there's a bad_keywords_nwb which does not
 
I see 2 cat,s.
 
@NobodyNada "satta ?matka"
 
2:34 AM
@Makyen Yep, thanks!
@FreezePhoenix hmm, @Ashish are you a cat?
I didn't think so, but I'm not sure
 
Eh... Foxy. Close enough. Hey folks!
 
@NobodyNada I'm a Fennekin. I still have no idea what it is..
It seems to be a pokemon character.
 
ah, I never played Pokemon, so neither do I
 
I'm a bipolar Phoenix
 
I'm an incompetent human :P
 
2:38 AM
@thesecretmaster Now now...
"Roadkill pasta" Good grief
 
@FreezePhoenix oh yeah...our tags are interesting
Jan 12 at 1:17, by ArtOfCode
that'll be good with no context in the sidebar
 
@SmokeDetector this needs delete votes methinks
 
@NobodyNada You can add patterns that effectively don't start/end on a word boundary in blacklist or watchlist. You just have to take into account that the pattern does get bookended by \b. Thus, you need your pattern to consume all of the \w or \W characters which are before and after it. In this instance, \w*satta ?matka\w* would produce the same results, with the minor difference of including the extra in the matched string, which would look somewhat different in the "why".
 
I find it kind of amusing to think that if their website really is as simple as they claim in the time they took raging about GDPR they probably could have updated their privacy policy
 
@Makyen That's true
 
2:44 AM
@quartata Can't delete until closed. Once closed, it will Roomba anyway (assuming no one answers & gets upvoted).
 
I know. But they're just going to keep complaining in the comments which I think is motivation to delete it sooner after closing
 
Positivity incoming
 
@FreezePhoenix ...did your picture just change, or am I crazy?
 
@NobodyNada Yes.
 
(assuming yes means the picture changed) did you change it a while ago and it just propagated to chat, or did you somehow get it to sync to chat immediately?
 
2:50 AM
Changed it about a minute ago.
 
huh, how'd it make it to chat so fast then?
 
@ArtOfCode Usually it takes hours...I guess he caught it right before a resync?
 
or at some opportune moment in the cache item's lifecycle
the foolproof trick to get it to propagate quickly is to find a friendly mod
 
Odd. It's happened every time after the first change per day. The nth change however, takes a while
 
2:55 AM
@FreezePhoenix did you perhaps just change your parent site? The profile pic you have now on IPS is the one I remember seeing
 
Yes. But even with that,it still takes a while for the nth one where n is greater than one.
 
Hey... Zoe is still around?
 
Restart: API quota is 19610.
 
She was here earlier though
 
3:02 AM
Interesting
 
@FreezePhoenix shrug it's just timezones...
 
Not talking about that. We live in the same timezone +- 1 hour
 
from your activity graphs it looks like two very distinct timezones...
 
I usually I only have access to chat in the evenings.
Despite consistent access to SE and internet.
 
sounds like a personal 'when i am online' problem :P
just means you need to be online earlier :P *shot*
 
3:14 AM
Not really... Can't help it with the Wi-Fi filter(s)
 
then you should really stop using restricted corporate/public wifis for things :P
just saying.
 
And wat am I to do about it? (I'm totally with you though)
 
i dunno, you don't have a phone with a mobile hotspot?
or VPN?
 
VPN would be disabled... And mobile hotspot caught in minuts. Yeh, I've got it bad.
 
tp- by Makyen
 
3:21 AM
@FreezePhoenix where in God's good earth do you work/use the Internet where they would detect your mobile hotspot and come find you and 'eliminate the problem'?
some government building somewhere?
Soviet Russia?
 
it's called a school I do believe
 
Area 51?
I mean, seriously.
 
they're pretty terrible about filtering
 
@ArtOfCode and how would schools neutralize a mobile hotspot
just asking
 
I can detect a mobile hotspot
 
3:22 AM
@thesecretmaster maybe a Soviet Russia one, but not any normal one.
 
@ArtOfCode I'd be surprised at a school that was that competent and strict at network administration
 
with this magical piece of technology called... a phone
 
@ArtOfCode doesn't mean that they're actively hunting them down and destroying them.
 
literally just SSID scan
 
They can, but why would they?
 
3:22 AM
A place where they dont want you doing shit th at WE don't want to do in the first place.
 
@ThomasWard wouldn't put it past a school
 
@ThomasWard that reminds me...I've got a domain now, I need to set up a VPN server on my RPi
Is OpenVPN good? That's the only name I've heard
 
@ArtOfCode I would, considering that's a full time security and IT force needing to manage and monitor the entire IT network simultaneously and then subsequently finding and torching individuals' devices which are acting as hotspots
@ArtOfCode not to mention the potential for legal hell if they confiscate the device as such
probably should be asking though:
@FreezePhoenix so they don't let you use your own device, then, and force you to use their devices in a school environment?
 
Oh good sir, you vastly overestimate the legal oversight of schools
 
or are you not at a school
 
3:24 AM
@NobodyNada not saying I'm at a schooo, but our networking admins are competent enough to set up a blocking filter.
 
@ArtOfCode actually? Not really since I know quite a few people who won the fight in court because their phone was taken for "undisclosed reasons" when it wasn't really being used in any rule-breaking way
the civil reparations alone were pretty heavy
 
All you have to do is ban phones, stick a WiFi receiver next to every access point you have to do SSID scans, and when you pick up a new hotspot you go search that classroom
 
@FreezePhoenix I wasn't saying that schools don't set up filters -- I hear about it all the time --just that they don't usually seem to do a good job at it
 
Just make having a phone against the rules, then you don't need probable cause of breaking rules to confiscate it
this ain't law enforcement
 
@FreezePhoenix that then lends credence to the "How would they find/detect/neutralize a hotspot", but then I could say "This is a problem with where you work sorry you're screwed.
 
3:25 AM
@ArtOfCode It's also possible to share internet over Bluetooth or USB
 
@ArtOfCode depends on the school, legal jurisdiction, etc.
 
@NobodyNada yup, they can't detect everything
@ThomasWard any sensible jurisdiction won't interfere with a school enforcing its rules as long as they're not way out of line
 
@ThomasWard Hotspot = phone...
 
@ArtOfCode so if the phone is in someone's pocket, but off, and the school takes it arbitrarily just because it's on their person, even if it's off, that's not out of line?
 
@ThomasWard not if having a phone is against school rules
 
3:27 AM
I think you and I need to have a discussion over 'sensible enforcement of their rules'
 
this was literally the rule at my school
 
@ArtOfCode again, depends on the school and the legal jurisdiction
 
@ThomasWard that's what their supposed to do...
 
the schools around here have a rule: "No phones", but only enforced to the point of if a phone is being used
the act of having a phone and putting it into your locker or backpack is not violation of the rules, use therein of the phone is
so again, "depends on the school and legal jursidictions"
 
that's what I'm saying
if that's the rule then you're correct
 
3:28 AM
two people I know had their phone in their backpacks, off, had the phone taken just because "they saw it", and then their parents took the school to court and won
 
if the rule is "you can't have phones", period, then confiscating it would be perfectly reasonable
 
@ThomasWard Precisely.
 
@ArtOfCode this is an argument we may as well 'put away' since it's not going to be won reasonably here ;)
the whole point is it depends, so bleh
 
@ThomasWard yeah... the US is a litigious society for whatever reason
 
∆∆ hey...
@ArtOfCode When you can set a listener for a USB plugin or a Bluetooth event, that takes those off the list.
 
3:33 AM
@FreezePhoenix assuming BYOD is also banned, then yes
 
@ThomasWard Did I say you do?
BYOD?
 
bring your own device
i.e. use your own laptop
 
that statement was directed at AoC, though I should probably not talk/argue because i'm a bit tired/irritated/pissed off for reasons.
 
@ThomasWard hey snap :) I edited, it was probably unnecessary. Should probably get some sleep, too
 
@ThomasWard on sorry. Take a walk or you need to.
 
3:35 AM
@ArtOfCode issue is "too much sleep" not "not enough sleep"
 
GOSH darn this autocorrect
 
unfortunately
slept for about 14 hours last night, so has a skewed sleep schedule currently
 
\o/ people's favorite colors?
 
4:05 AM
[ SmokeDetector | MS ] Bad keyword in body, blacklisted website in body, link at end of body, pattern-matching product name in body, pattern-matching website in body: Understandably too, (while you think about the by lagecoc on askubuntu.com
tpu- by NobodyNada
 
4:58 AM
[ SmokeDetector | MS ] Repeating characters in answer: How to solve 2 tetrated 0.5 times? by Cael TheSquid on math.SE
 
5:08 AM
hey @Art can you pop into Root Access real quicm
 
@quartata or not?
:P
 
eh
Didn't know how many diamonds were aware
 
oh, we talk a lot amongst ourselves ;)
'tis part of our insanity and what gives us our power ;)
 
@quartata sigh
 
[ SmokeDetector | MS ] URL-only title, bad NS for domain in body, bad NS for domain in title, bad keyword in body, bad keyword in title, +5 more: www.healthyorder.org/keto-tone/ by bosamaxaka on askubuntu.com
tpu- by NobodyNada
tp by Nisse Engström
tp by Nisse Engström on How to solve 2 tetrated 0.5 times? [MS]
 
5:50 AM
[ SmokeDetector | MS ] URL in title, bad NS for domain in body, bad NS for domain in title, bad keyword in body, bad keyword in title, +6 more: www.healthyorder.org/keto-tone/ by Himet1958 on drupal.SE
tpu- by NobodyNada
 
6:24 AM
[ SmokeDetector | MS ] Blacklisted website in answer: Does a link to your site with a URL parameter pass link juice? by Shamsi on webmasters.SE
fp- by DavidPostill
naa by Nisse Engström
naa- by PeterJ
[ SmokeDetector | MS ] Luncheon meat detected: Vectors, dot, cross. MATLAB by MariaxJoe on stackoverflow.com
fp- by PeterJ
[ SmokeDetector | MS ] Potentially bad keyword in answer: How to reseed grass after killing weeds? by Duane Meulners on gardening.SE
fp- by PeterJ
 
6:50 AM
[ SmokeDetector | MS ] URL-only title, bad NS for domain in body, bad NS for domain in title, bad keyword in body, bad keyword in title, +5 more: www.nutritionfit.org/rapid-tone-diet/ by mapicidi on askubuntu.com
tpu- by PeterJ
[ SmokeDetector | MS ] Bad pattern in URL answer, pattern-matching website in answer: Headless Chrome Scalability by Aol customer support on stackoverflow.com
tpu- by PeterJ
fp by Nisse Engström
fp- by PeterJ
 
7:16 AM
[ SmokeDetector | MS ] URL in title, bad NS for domain in body, bad NS for domain in title, bad keyword in body, bad keyword in title, +5 more: www.nutritionfit.org/rapid-tone-diet/ by Fromp1943 on drupal.SE
tpu- by PeterJ
tp by Nisse Engström
[ SmokeDetector | MS ] URL-only title, bad keyword in body, bad keyword in title, blacklisted website in body, blacklisted website in title, +1 more: mumybear.com/rapid-tone-shark-tank/ by john123 on askubuntu.com
tpu- by PeterJ
 
7:47 AM
tp by Praveen on Headless Chrome Scalability [MS]
fp by Praveen on Vectors, dot, cross. MATLAB [MS]
[ SmokeDetector | MS ] Few unique characters in answer, no whitespace in answer, repeating characters in answer: Adding Pagination or Lazy Load to Instagram feed by user9818236 on stackoverflow.com
tpu- by Makyen
[ SmokeDetector | MS ] Blacklisted user: wordpress : How to modify continue reading link? by eitren on stackoverflow.com
fp- by PeterJ
fp by Zoe
[ SmokeDetector | MS ] Blacklisted user: How to modify continue reading link? by eitren on stackoverflow.com
fp- by PeterJ
 
Zoe
8:09 AM
sd - fp-
 
[ SmokeDetector | MS ] Manually reported answer: Windows account ending with $ by Manab roy on superuser.com
tpu- by DavidPostill
[ SmokeDetector | MS ] Bad keyword in body, bad keyword in title, blacklisted website in body, link at end of body, pattern-matching website in body, +1 more: Renuvaline collections and wrinkles etc No females by Dianekelly on askubuntu.com
tpu- by Zoe
[ SmokeDetector | MS ] Potentially bad keyword in answer: Magten 1.9 - Onestepcheckout - Success.phtml not working by Madis Männi on magento.SE
fp- by PeterJ
[ SmokeDetector | MS ] Bad NS for domain in body, bad NS for domain in title, bad keyword in body, bad keyword in title, blacklisted website in body, +5 more: www.healthyorder.org/rapid-tone-diet/ by Quared1952 on drupal.SE
tpu- by Zoe
 
8:51 AM
[ SmokeDetector | MS ] Bad NS for domain in body, bad keyword in body, blacklisted website in body, pattern-matching website in body: Keto Power Diet Want by JohnMondy on drupal.SE
tpu- by PeterJ
[ SmokeDetector | MS ] Link at end of answer: creating custom tableview cells in swift by Waseem Siddiqi on stackoverflow.com
tpu- by DavidPostill
[ SmokeDetector ] SmokeDetector started at rev 45f24c2 (SmokeDetector: Auto unwatch of islamweb\.net by Zoe --autopull) (running on teward/Solar Flare)
Restart: API quota is 18001.
 
00:00 - 09:0009:00 - 00:00

« first day (1645 days earlier)      last day (2185 days later) »