« first day (1752 days earlier)      last day (2364 days later) » 

18:06
[ SmokeDetector | MS ] Potentially bad keyword in answer, blacklisted user: DAO Dividends In Ether? by VladChan on bitcoin.SE
tpu- by Erik the Outgolfer
tpu- by Erik the Outgolfer
[ SmokeDetector | MS ] Link at end of answer, potentially bad keyword in answer: Are calendar invitations rude? by kapiko007 on workplace.SE
tpu- by Erik the Outgolfer
[ SmokeDetector | MS ] Potentially bad keyword in answer: Do any non-human species have packed abs? by man on biology.SE
naa- by Erik the Outgolfer
@Undo ...huh
What about keys sucessful_ci_count/de90a251*?
18:22
@thesecretmaster (empty list or set)
extra huh
keys sucessful_ci_count*?
(yes, I know there's a spelling error)
keys * is empty
... whut?
Don't think it's the wrong Redis instance, either
Might be the wrong database
18:23
doesn't make much sense :P
Try redis-cli -n 0?
no keys
empty
So strange
18:25
and 2 and 3
Are the redis settings in database.yml?
Don't see anything special
How about...
rails c
they're in cable.yml, but that's it
redis = Redis.new
redis.keys "*"
18:26
 => #<Redis client v3.3.5 for redis://127.0.0.1:6379/0>
@thesecretmaster []
Alright
I give up
This is so strange
You're on the ec2, right?
heh
yes
(I don't think you're stupid, I just have no idea what's up)
Because Redis.new.get is literally what the controller does
That was rails c p too, so it should be exactly the same
Any chance there are redis errors in the logs (I doubt it because stuff later in the controller executed, but maybe)
18:29
Which logs?
Got something to grep for?
Actually, I can just throw some logging at this in a custom log file
At work, only have about half a brain right now
It's fine, I'll work it out
No harm done currently, all that happens is we miss the "CI has passed" message when all three tasks pass.
If you're here, wanna sanity check my changes then deploy?
(unrelated)
18:31
master?
Yep
Just kinda don't want to push the big button because the Sites table is important
Also see issue 501
Important, but super easy to restore.
I like that solution. Better than the stuff I was coming up with in my head
LGTM, deploy whenever you want
Still requires some manual database edits from you though
Not urgent
Because now all the old sites are hidden from pretty much everywhere.
18:34
@thesecretmaster Just to migrate some not-closed-but-moved sites?
Yep, e.g. merge Programmers into Software Engineering
All my changes do are stop it from happening in the future and hide any non-existant sites
[ SmokeDetector | MS ] Pattern-matching website in answer, potentially bad keyword in answer: Median in Pivot Table in Excel 2010? ✏️ by Prospertute on superuser.com
[ metasmoke-deploy ] deploy started by thesecretmaster
k. Shouldn't be too hard, I'll just figure out how to replace site_id=x -> y across the whole database.
Yep, prolly AR has a utility for it
Iterate tables, do some ugly stuff
might be migration material
18:37
Maybe. I don't want to poke at it, since it's super low priority, but maybe at some point
yep. Okay to deploy, then maybe throw an issue to resolve the few cases manually?
2 mins ago, by SmokeDetector
[ metasmoke-deploy ] deploy started by thesecretmaster
Also, there's still my ms-deploy PR sitting there
@thesecretmaster merged, pending a couple of builds to see if it works
Ok. Seems like the current build is stalled, which is weird
Webpacker stalled. Haven't seen that before
18:43
Weeerd
Maybe connected to test stalling
server seems like it's choking
I feel like we need some of our puzzling users over here
@Undo CPU or memory?
This might be the issue from before
When it mysteriously ate all the credits a few days ago
18:46
Was that when I was fighting with yarn/npm?
Friday IIRC
Sounds right?
Aug 30 at 14:19, by Undo
Only other thing that could possibly run at 11:00 is check_smokey_status, and I kinda doubt it did this
Oh, I remember that
Froze up before I could get htop to tell me what happened
Restarting, then going to look at log file
18:47
But nobody was deploying at that time?
Don't think so
But webpack only does it's mess during deploys
CPU log will give some insight into what happened
angussidney/droplet received failover signal.
Uh oh
Did MS explode just now?
18:50
yes
Back now
Restart: API quota is 19998.
!!/location
@thesecretmaster teward/Solar Flare
No idea what that was.
@thesecretmaster angussidney/droplet
18:50
Reboot 'fixed' it, eventually, but there's something weird there
Suspect it's Maria, somehow
!!/standby angussidney/droplet
angussidney/droplet is switching to standby
!!/location
@thesecretmaster teward/Solar Flare
JS on the status page needs fixing for admins
Started at a time when the check_smokey_status job was running, and I did see it in htop for a second taking all the CPU
which is not a thing it should do
:(
Chuck that in an issue?
We can soak check_smokey_status in logs and wait until 11 again
[ SmokeDetector | MS ] Bad keyword in link text in body: Hp printer isn't print any documents, what should i do? by jhonperker on askubuntu.com
18:54
505
tpu- by teward
Retrying deploy
now with 100% more link
[ SmokeDetector | MS ] Offensive body detected, offensive title detected, toxic body detected: Why is mysql such a piece of shit? by steve antwan on stackoverflow.com
19:01
@Undo Wanna manually rerun the build, to see if the cache worked?
[ SmokeDetector | MS ] Potentially bad keyword in body: How do I migrate VS 2010 to VS 2015? by Balamurugan.Santhanaraj on stackoverflow.com
@SmokeDetector Technically, that was repairable, but off-topic. It was deleted by a moderator, leaving the spam/rude flags in existence, but not deleted as spam/rude.
Yep
[ metasmoke-deploy ] deploy started by Undo
@thesecretmaster nope, still trying to install
19:03
grumble grumble
Ah, I got the path wrong
needs to be metasmoke/vendor/bundle
Want me to make another PR, or can you just tweak it?
I'll get it
webpacker hang again
veeeeery close to ripping webpacker out
nope, there it goes
Heh. Only the JS folks (aka neither of us) like it :P
@thesecretmaster Does github.com/Undo1/metasmoke-deploy/commit/… look right?
19:06
Perfect
fp- by Makyen
First build running...
Also, you didn't happen to find a way to trigger a new build for the same commit in travis, did you?
I can do it from the UI, and there's a POST request you can make
It's just apparently not in the gem, so I didn't go through the trouble of figuring out the raw request
[ metasmoke-deploy ] deploy started by Undo
19:10
Second run
huh
that's weird
We lost metasmoke?
ah
the cache thing pre-creates the directory structure, then git freaks out
Cache gets applied first
Yep
Can you force a clone-over?
19:12
Haven't done that before
or clone to a different dir and cp over
Maybe it's possible to apply the cache after clone?
Don't think so
It takes the gems from wherever bundle installs them to by default -- system gem dirs.
19:15
oh
The problem is that it expects that it's doing the clone automatically
Right
So it can apply the cache over the clone
But we're not doing that
Yeah, sounds like it needs some combination of copying stuff
git clone https://github.com/Charcoal-SE/metasmoke metasmoke.tmp && cp -r metasmoke.tmp/* metasmoke/
19:17
Not familiar with how cp does overwrites, but if that works then it looks good
yeah, that should work
git doesn't have a vendor folder for cp to clobber with
Are you going to work on this, or should I
If you're looking for something to do, that'd be great.
I'm at work, probably wouldn't get to it for a while anyway
Can do. Gimme maybe a half hour
k, thanks!
@ArtOfCode I think I may have not phrased my last review request well enough. I meant this PR
@SmokeDetector k
@SmokeDetector Prior, clear offense from same user
@Undo How could a user flag that as spam? (even if they know previous post of user)
The previous post by user yes has a link but it does not even work....
19:46
in Tavern on the Meta on Meta Stack Exchange Chat, Aug 16 at 23:35, by Undo
@SonictheInclusiveHedgehog Charcoal feedback, while usually lining up with flags, shouldn't be used to interpret "this should have been marked helpful".
so k!=red flag?
Corollary is also true - tp feedback doesn't always mean "someone should have cast a red flag on this". It means "we would like to have blocked this at the system level"
I'm using a bit of mod judgment to say that was a poor attempt at spam.
(also, I see more than you do)
meeh, ok, I just feel very confused.
@PetterFriberg Ultimately, it's all about making sure our statistics for autoflagging and internal reviews are accurate. Complex subject, and when you try to take the subjectivity out of it you get a complex set of almost-objective stuff.
Normal to be confused if you don't live in this day to day
Mean to link to one that's been around for four days?
#5
Oh oops
Dunno why that was even on my clipboard
@Undo I just feel that your old teaching was to be really careful about being overzealous about flagging. So I'm a bit lost, already the previous answer it's enough with old rulez to NAA (a non working link on a norton question, red flag?)
19:53
Flagging is different than feedback.
...and don't let FIRE or FDSC convince you otherwise :P
I thought of a better solution than copying over
our statistics for autoflagging? so watch out about the feedback
@PetterFriberg This is an edge case, don't read too much into it. I'm making a call as an SO mod that it was a poor attempt to promote... something... and I would have accepted red flags on it.
Less a "charcoal leader" thing and more "SO mod judgement" thing
ok thanks, sorry I was just getting very confused (and maybe a bit worry)
19:55
@thesecretmaster Got a meeting, should be able to look in half an hour or so unless I forget or die.
Don't die
meetings kill
Meeting can be dangerous, make sure to wear the appropriate PPE
@Undo don't die, you're a very cute hairy little puppy, shiny as a diamond
Hrm. What's the appropriate PPE for a meeting... I guess it depends on the type of meeting.
@thesecretmaster Something rated to deflect action items
@thesecretmaster also on where you work
@Undo Prolly ear protection
@thesecretmaster a full vest
it'd be helmets and harnesses for me :P
20:00
but under your clothes, don't forget to look formal!
@ArtOfCode Ah yes, I forgot that you've actually got PPE at work as well :)
spent most of today inspecting it
quite possibly the single most tedious thing ever invented
@ArtOfCode SO review queues?
even worse than that
I can only do 40 tasks in a queue
there's something like 250 bits of equipment we get to inspect
@ArtOfCode SO review queues as a mod?
20:03
N/A, never tried it :)
Art's not a SO mod
Hrm... should this room require ppe?
yes, eye protection and gas masks
yes, definitely, I mean, coal is handled here!
What's the UK equivalent of OSHA?
20:05
@thesecretmaster HSE
@ArtOfCode Just imagine everyone sitting at their computer staring at the screen though their eye protection and gas masks... funny image.
Now, who aside from me actually has a gas mask in their home?
tp by Nisse Engström on DAO Dividends In Ether? [MS]
Well, not a gas mask, but a respirator.
ah, must be sad
20:12
@Mithrandir given that I have full fire gear around somewhere... yeah, count me in :P
20:39
fp by Nisse Engström on Why is mysql such a piece of shit? [MS]
@ByteCommander Your flag conditions were manually disabled by @SmokeDetector: Total accuracy must be >= 99.75%.
@Mithrandir I have a breathing air filter dealie
Sorry to get all technical
21:03
@thesecretmaster travis is unhappy after the most recent PR
[ SmokeDetector | MS ] Potentially bad keyword in body: When and why would one use a group of Executor by user10314481 on stackoverflow.com
@Undo Maybe just symlink to metasmoke/vendor?
Want a PR, or you'll do it?
mobile, but can merge PRs
There's also the GH editor
fp by Nisse Engström
21:11
ick. phone typing.
J F
J F
Anyone else get a 500 when visiting desktop/desktop#5399?
@Undo PRed
@JF Yep
@JF repro'd
@JF yes
J F
J F
strange
21:13
blames Microsoft
@Undo Don't merge me
Half a second late
No worries
Don't deploy
21:15
Travis won't pick it up until I tell it to
@Undo Again?
Started
So far so good
Again?
(to test cache)
Started
No message for some reason
@Undo Message?
Started that second one from metasmoke /status/code, didn't get a message in here
No worky
21:24
"all_teh_gems is not yet cached"
Oh, I need a before_cache: stage
[ SmokeDetector | MS ] Potentially bad keyword in title: Organic or Paid search SEO by userNotFound on stackoverflow.com
fp- by J F
[ metasmoke-deploy ] deploy started by Undo
woah you guys are struggling for hours
21:36
would be solved faster if I was more than 10% engaged
Looks like it worked @thesecretmaster
pets @Undo
hopes he could treat thesecretmaster, but he's somewhere...secret
@Undo \o/
@EriktheOutgolfer I'm struggling with some technical difficulties... my profile pic reflects my life
@EriktheOutgolfer Well, at 2 minutes per build, it'll save... some time?
21:55
@ByteCommander Your flag conditions were manually disabled by @SmokeDetector: Total accuracy must be >= 99.75%.
@SmokeDetector But it is!

« first day (1752 days earlier)      last day (2364 days later) »