« first day (1450 days earlier)      last day (2371 days later) » 

4:03 PM
sd k
 
Something's wrong with CircleCI because of MS being down I think, Github thinks it's going to run on PRs
We need to seriously start splitting off all these little cron jobs and lambdas
 
or we need a better metasmoke :p
@quartata where are you getting the problem with CircleCI from? I don't see any failures?
 
@Undo @Andy any more research into Serverless for these tasks?
@ThomasWard Check any PR, it'll say one expected check
"Waiting on status..." I'm pretty sure MS plays a role in controlling what CircleCI runs on
 
No, that's not an MS issue
that's a GH reporting issue
and if GH is having issues processing the return result then GH staff needs to be made known about the issue
 
CircleCI is not supposed to run on our PRs
 
4:15 PM
then the system is misconfigured.
CircleCI runs the PRs, so does Travis ,so...
I don't have access to revise the specific settings for a given hook, like CircleCI. I've poked GH support about the CircleCI issue.
 
[ SmokeDetector | MS ] URL in title, bad keyword in body, bad keyword in title, blacklisted website in body, blacklisted website in title, +3 more: musclech3mistry.org/explosive-muscle/ by jordanrqaudi on apple.SE
tpu- by Glorfindel
 
shut up
2 messages deleted
@SmokeDetector k
 
[ SmokeDetector | MS ] Few unique characters in body: Worker with several executors- internal comunication by rom.w on stackoverflow.com
 
@SmokeDetector v
 
@quartata looks like Circle is set to not build forked PRs.
it doesn't have an option to disable pull reqs with origins within the project it seems.
 
4:18 PM
and there's no other webhook set
I don't see anything in particular in GithubController that would cause this ither than the metasmoke/ci hook
 
@quartata that wouldn't affect Circle
Circle shows under services
not under custom webhooks
here's an example where that's important
or rather
under webhooks
an example of the way that shows in webhooks for Circle integration is there for my imaplibext python CI integration
 
@quartata Not from me.
 
[ SmokeDetector | MS ] Manually reported answer (batch report: post 1 out of 2): Old photography effect in Photoshop by Damagedphotorestoration on photo.SE
[ SmokeDetector | MS ] Manually reported answer (batch report: post 2 out of 2): Old photography effect in Photoshop by Damagedphotorestoration on photo.SE
 
4:29 PM
sd 2k
 
@quartata yes, but I'm a bit handicapped at the moment. I'm waiting for drive recovery on my failed drive. I am working on it though and hope to really get back to it late this week or next week
 
@Ferrybig Fixed. Postmortem: (1) Log exporter working as expected; log files were like 50MB total. (2) Some idiot put a 4GB swapfile on that server.
 
@Andy Did they ever tell you if it was a spinner failure or a PCB failure?
 
Filesystem      Size  Used Avail Use% Mounted on
/dev/xvda1       12G  4.8G  6.4G  43% /
Fixed now. Probably a bunch of temp files too
184K tmp maybe not
(from before reboot)
 
[ SmokeDetector | MS ] Pattern-matching website in answer: Some facts about Lord Jagannath temple by Neel on hinduism.SE
fp- by Glorfindel
 
4:35 PM
@Magisch One of the heads failed. The good news is that means 7 out of 8 were still working so they could recover most of the data. The orphaned files there were finding though were some of the more important ones I needed
So, once again: Back up your data
correctly
 
Head crash?
The programmer in me says "switch the platter to a head that works" :P
 
@Andy Ahh that means you'd have been pressed to do anything yourself
@Undo You can do that, that's how the company my company hired did it the last time
The problem is you need a super clean clean room and a lot of care for that to work
 
@Undo I asked that too. Apparently, it doesn't work that way. shrugs That's why I paid the professionals.
 
probably scratched the heck out of that platter or something
 
It does work that way, but you'd be paying a lot more
And it's still no guarantee if the head failed in a way that messed up the platter first
 
4:37 PM
See, this is why SSDs are better: People like me know they're black magic and don't try to fix them ourselves
 
I'm paying quite a bit. :) When I talked to them late last week they were getting a replacement and moving it to the clean room.
 
Isn't a SSD like a ram with memory permanence?
 
at a super high level, yeah
 
@Andy Ahh, so they're doing a swap
 
@Undo Heh. That's like me with cars. opens hood and sees engine is still present No idea what's wrong.
 
4:39 PM
Honestly, cars are just like computers. You can use the exact same debugging methodology on them
 
@Andy Aka me with anything more then menial mechanica
I tried to change a bicycle wheel and failed
I can re-solder a motherboard if need be but apparently changing a bicycle tire is beyond me
 
Just like computers. I just drove with grandparents about 800 miles in a diesel truck pulling a 10k+ lb trailer.
ding ding ding "Change fuel filter", followed by "engine power is reduced"
Good news: this had happened before, so we had a fuel filter in the back seat.
bad news: This is the US, so we were 80+ miles away from anything. Literally anything.
(in Europe, I guess 80 miles is like a whole country)
 
@Undo that's a very remote place
I'm not sure there is any point in germany where I can be 80 miles from civilization
 
pretty easy to do out here
"Engine power is reduced" is code for "now you can only go about 20mph on a curvy highway with speed limit 70mph in the dark"
slower on hills, of which there are many
 
o_O
 
4:43 PM
So what'd we do? Drive 80 miles at 20mph in the dark. Never again.
 
so 4 hours of driving in the dark?
 
[ SmokeDetector | MS ] Link at end of body: Who owns this patent? by Jay Patent on patents.SE
 
more like forty hours, but yes
 
Woo! Middle of nowhere!
 
@Undo 20 miles per hour / 80 miles = 4 hours?
 
4:44 PM
Stopped at a shady-as-heck gas station (in the dark), loaded the pistol, hunted around for a long time for that bucket of tools we knew we had but no one actually threw in, etc.
 
am I missing something?
 
@Magisch time perception
 
wtf you had a pistol
why
 
see: middle of nowhere
 
what are you gonna shoot in the middle of absolutely nowhere?
 
4:45 PM
Hopefully nothing :P
 
Yeah but why do you want a pistol there
 
Anyway, took about an hour to change that filter. Truck runs file. Make it the next five miles to a motel, stayed there overnight. Wake up, drive 200mi to next big city, "engine power is reduced"
Now, a fuel filter in a diesel should last about thirty years. Not two hundred miles.
So we stopped, did it again.
Fifty miles later, same thing.
So we finally did the smart thing and bought one of those little OBD reader things. Plugged it in and got a bunch of codes, as is usual for a truck where the engine light is always on and no one cares
Now I've got error codes. Hey, I know what to do with those.
Root issue is still unknown, something about a fuel injector throwing a bad pattern or clutch slipping. Pulling a 10klb trailer up a hill for 3+ miles puts a ton of strain on stuff.
But, the computer throws it into "limp mode", which is "something is wrong and we're going to keep you from shooting yourself in the foot and blowing your engine up"
 
@Magisch whispers U...S...A
 
How do you fix computers?
Next time it did it, I pulled over to the side of the road for TWO MINUTES at the suggestion of some random Internet guy (tried one minute, didn't work). That two-minute stop had the exact same effect as stopping for an hour to change the darn filter
1 hour changing a filter is more than two minutes, so the computer resets and you think changing the filter fixed it.
 
Overheating issue?
 
4:51 PM
nah, temps were fine
 
Ahh so the filter wasn't the problem
 
Exactly
Something else was wrong, causing the computer to slow us down and say "usually this is a filter issue, so maybe change it"
illustration:
 
Solution:
 
@Undo So what ended up being the problem?
 
anyway, moral of story: There are lots of people on the internet that have cars, some of them have the same issues you do. Just like programming.
 
4:53 PM
 
Or did you just keep doing 2min stops and take the truck to a body shop later?
 
@Magisch Don't know. Could be a clutch slip, injector pattern issue (computer gets irritated when there's too much bypass for too long), etc.
@Magisch Did two more 2min stops over the next 300+ miles.
We've been driving it since then (more than a week) on 100+ mi trips to town. It runs fine if you're not pulling a trailer double the truck's weight up a hill
It's well within the load capacity, there's just some issue that only manifests itself on long hard pulls
Planning to take it somewhere eventually. None of the codes are super high priority (glow plug #8, who cares really)
 
@Magisch @Undo it literally is - they're just different types of RAM. System-memory RAM is DRAM; SSDs are SRAM.
 
@Magisch I do that all the time. I end up just making everything worse every time I attempt to fix the thing.
 
System RAM is faster than SSD RAM because of that - otherwise, you could legitimately use an SSD as system memory
 
5:09 PM
@ArtOfCode Optane!
 
@bwDraco still unlikely to be as fast as system RAM
 
[ SmokeDetector | MS ] Blacklisted website in answer: Making dynamic content searchable in Google by Angularjs Examples on stackoverflow.com
[ SmokeDetector | MS ] Repeating characters in answer: What's the width of the average mountain? by user44470 on worldbuilding.SE
naa- by ByteCommander
 
@ThomasWard I meant hooks into Circle, not GH
 
also I forgot Smokey only pulls code privileges when it doesn't have it :/
 
5:26 PM
Um... Did this one get caught: ell.stackexchange.com/a/146545/16318
 
@Catija nope.
 
Ah. OK.
 
[ SmokeDetector | MS ] Bad keyword in body: How to disposal from sexual desire? by user10548941 on health.SE
fp- by Glorfindel
 
5:48 PM
PSA: SOCVR Roommeeting in ten-ish minutes
@quartata the room will be open when the meeting starts, just refresh the screen once the chair kicked of the meeting. No need to request access.
 
Oh, all right
 
np, we know we are confusing
 
Probably dont need to also timeout it then :p
 
tpu- by KazWolfe
 
5:54 PM
lul
 
@quartata do know that word has meaning in Dutch. I almost flagged you ...
 
third-fastest invalidated flag on the network
 
6:25 PM
@rene ...sorry. I didn't actually know that
 
@quartata meh, doesn't matter. You're fine ...
 
6:37 PM
[ SmokeDetector | MS ] Bad keyword in link text in body: How do I set up live Twitch streaming from RetroArch on a Pi 3? by LeoPGiles on raspberrypi.SE
[ SmokeDetector | MS ] Bad keyword in link text in answer: How do I set up live Twitch streaming from RetroArch on a Pi 3? by LeoPGiles on raspberrypi.SE
[ SmokeDetector | MS ] Offensive answer detected: What can com.sec.android.app.launcher be used for? by user8895943 on stackoverflow.com
tpu- by Glorfindel
@QPaysTaxes HEAD isn't at tip of origin's master branch
tpu- by ByteCommander
naa- by Glorfindel
 
It's definitely good to have a record of it, but I won't flag that as spam.
 
!!/gitstatus
 
On branch deploy
Your branch is up-to-date with 'origin/deploy'.
nothing to commit, working directory clean
@QPaysTaxes HEAD isn't at tip of origin's master branch
 
Should it be on master instead of deploy?
 
@QPaysTaxes angussidney/EC2
 
6:49 PM
!!/master
 
Restart: API quota is 16712.
[ SmokeDetector | MS ] Username similar to website in answer: Xamarin.Forms .NET Standard 2.0 Live Player Error by hmheng on stackoverflow.com
@QPaysTaxes HEAD isn't at tip of origin's master branch
 
On branch deploy
Your branch is up-to-date with 'origin/deploy'.
nothing to commit, working directory clean
 
ergh...
f seems appropriate
 
!!/master
 
!!/gitstatus
 
On branch deploy
Your branch is up-to-date with 'origin/deploy'.
nothing to commit, working directory clean
Restart: API quota is 16670.
@QPaysTaxes HEAD isn't at tip of origin's master branch
  File "/home/ubuntu/charcoal/SmokeDetector/chatcommands.py", line 374, in do_blacklist
    code_permissions=is_code_privileged(kwargs['ev_room'], kwargs['ev_user_id'], kwargs['wrap2'])

  File "/home/ubuntu/charcoal/SmokeDetector/datahandling.py", line 138, in is_code_privileged
    metasmoke.Metasmoke.update_code_privileged_users_list()

  File "/home/ubuntu/charcoal/SmokeDetector/metasmoke.py", line 311, in update_code_privileged_users_list
    data=json.dumps(payload), headers=headers).json()['items']
 
!!/pull
 
Restart: API quota is 16660.
 
!!/watch severalnines
 
6:56 PM
@quartata HEAD isn't at tip of origin's master branch
 
no it probably won't make a difference
I suspect metasmoke screwed something up when rebasing deploy
since it was down
I need to pull to compare histories, one moment
 
le sigh
still happening?
 
> commit c0a7979d2f69c2046f0a9689544df7d8dbcc2d15 (HEAD -> master, origin/master, origin/deploy, origin/HEAD)
maybe not
 
@QPaysTaxes HEAD isn't at tip of origin's master branch
 
7:01 PM
!!/standby angus
 
bwDraco/Linode received failover signal.
 
wtf
 
Restart: API quota is 19972.
 
metasmoke pls
 
teward/Solar Flare received failover signal.
 
7:01 PM
!!/standby bwDraco
 
ok
!!/pull
 
good
 
@QPaysTaxes You don't have code privileges, but I've created PR#1204 for you.
Restart: API quota is 19999.
 
7:02 PM
by the way @ArtOfCode did you change something with circleci
 
Restart: API quota is 19998.
 
!!/standby teward
 
why
 
@quartata ?
 
7:03 PM
both it and travis are trying to run now
 
@quartata because I'm here and can watch it
 
Henders/EC2 received failover signal.
 
@quartata they always have
 
take a look at the checks
@ArtOfCode not on PRs
 
Restart: API quota is 19999.
 
7:04 PM
!!/location
 
@quartata Henders/EC2
@quartata ArtOfCode/UPC
 
!!/standby Henders
 
@quartata then probably, yes
 
teward/Solar Flare received failover signal.
Restart: API quota is 19985.
 
omfg
ok
 
7:05 PM
Merged SmokeDetector #1204.
 
I don't think your instance is pinging metasmoke
 
Restart: API quota is 19987.
 
!!/standby teward
@quartata tis now
 
cool
 
7:05 PM
@QPaysTaxes ... did I miss something?
 
helps if you put the right token in the config file
 
okay, do your worst
 
CI on 24304f9 succeeded. Message contains 'autopull', pulling...
Restart: API quota is 19973.
 
next question: why does circle not run on PRs
 
7:08 PM
[ SmokeDetector | MS ] Pattern-matching website in answer: Wordpress permalink and HTTPS by Sandeep goswami on stackoverflow.com
 
answer: because settings
anyway
@quartata once the conflicts are resolved I can merge it
 
things are a bit broken but should fix themselves on the next PR
 
CI on 4372388 succeeded.
 
there we go
 
7:20 PM
CI on 48fc607 succeeded.
[ SmokeDetector | MS ] Repeating characters in answer: Playing multiple sound files at same time with Windows Forms Visual C++ by AUSTIN PITTS on stackoverflow.com
[ SmokeDetector | MS ] Pattern-matching website in body: Trying to convert users message into hex from decimal using Discord bot by Jake Pavek on stackoverflow.com
[ SmokeDetector | MS ] Potentially bad keyword in answer: WKCrownDelegate doesn't seem to be working with Xcode 9 GM by Larry Drebes on stackoverflow.com
CI on 69b4519 succeeded. Message contains 'autopull', pulling...
Restart: API quota is 19800.
 
@SmokeDetector f
 
@SmokeDetector f
 
fp by Mithrandir
 
(question asks for it)
 
7:42 PM
!!/location
 
@Henders ArtOfCode/UPC
 
@ArtOfCode I haven't been following, do I need to do something to my instance?
 
@ArtOfCode I think we should think about this actually
In order to merge a PR both CIs will have to succeed
i.e you have to wait for both
maybe we should make circle optional or something
idk
 
How long does Circle take, generally?
 
Actually Circle is (usually) faster
 
7:43 PM
I haven't seen any CI things that actually take longer than two minutes
 
But Travis is the one that submits coverage data to codecov
@Mithrandir That's because on commits both run, and whichever one succeeds first is what Metasmoke uses to tell SmokeDetector to pull
But on PRs it's different -- both are mandatory checks as of now, so you have to wait for both
 
yup
 
sd k
 
^ should be up to date now, I'll failover it later if that's alright to test...
 
8:21 PM
I suspect something blew up last night while I was asleep?
 
meatsmoke
wait a second
is the reason why MS keeps going down because it's actually a piece of bacon
that seems like a poor design choice
 
J F
@quartata It’s actually a can of spam.
 
!!
 
@JF Does it detect itself?
 
@ArtOfCode So aside from the various hooks in github.com/Charcoal-SE/metasmoke/blob/master/app/controllers/…, what other webhooks is Metasmoke in charge of currently?
 
J F
8:30 PM
@DJMcMayhem Let’s find out.
 
> Url in title
----------
Title - Position 1-36: https://metasmoke.erwaysoftware.com
[ SmokeDetector | MS ] Few unique characters in answer, no whitespace in answer, repeating characters in answer: How did life change for the Average American as a result of the end of World War 1? by user27810 on history.SE
tpu- by J F
 
@JF Hmm, it looks like you're affiliated with that site, and clicking on it gives more links to lots of spammy sites. I'm gonna go ahead and say that's a tpu-
 
@quartata many
 
@DJMcMayhem One does not simply click on a spam link...
 
honestly, shrug
 
8:35 PM
would probably be better to look at routes.rb
 
@JF Well played..
 
J F
@DJMcMayhem What would make you think that I am affiliated with metasmoke and Charcoal?
 
honestly I think that's it
I'm looking at the routes
 
@JF It's just a hunch
 
there's some hooks for OAuth but those need to stay on MS obviously
 
9:31 PM
[ SmokeDetector | MS ] Offensive body detected: Miami Miami Miami Miami Miami by user811429 on superuser.com
tpu- by DJMcMayhem
[ SmokeDetector | MS ] Blacklisted user: Miami Miami Miami Miami Miami by user811429 on superuser.com
tpu- by Glorfindel
 
hello
 
9:52 PM
[ SmokeDetector | MS ] Potentially bad keyword in answer: Do I need to prevent preemption while measuring performance by Razvan C on stackoverflow.com
 
sd why
 
@DJMcMayhem [:40988275] Body - Position 931-942: millionaire
fp- by DJMcMayhem
[ SmokeDetector | MS ] Username similar to website in answer: Xamarin.Forms .NET Standard 2.0 Live Player Error by hmheng on stackoverflow.com
fp- by ByteCommander
 
ok we need to get our shit together
first time that was reported, it was tp'd by 3 people
then it was reported again 3 times in a row and fp'd every time
 
It has proper disclosure for the link and looks useful to me, therefore fp. It also has a score of +2|-4, where the -4 probably come from misplaced red flags...
 
they are different revisions but not much difference -- he just put more of the tutorial in the answer
sd - ignore-
 
@ArtOfCode Why does checking for conflicting feedback/fp on autoflagged occur in update_feedback_cache and not when the feedback model is created
I'm assuming there's a Rails reason
 
10:21 PM
links?
 
[ SmokeDetector | MS ] Potentially bad keyword in body: JavaScript for online quiz with countdown timer by ojwiya on stackoverflow.com
fp- by ByteCommander
[ SmokeDetector | MS ] Bad keyword with email in answer: is K-Means clustering suited to real time applications? by Grant Goris on stats.SE
 
tp or naa?
 
10:36 PM
fp- by rene
 
I'm not entirely sure where that is then called from but
 
that basically is an after_create IIRC
 
Oh OK
 
yeah, it gets called from an after_create on Feedback
 
And nowhere else?
Looks like it
Funny that the "It seems like this is your first time" message is then in FeedbacksController.create
Although granted that's operating a single feedback, this is an overall post status thing
 

« first day (1450 days earlier)      last day (2371 days later) »