« first day (1137 days earlier)      last day (2668 days later) » 

12:00 AM
otherwise we have the head mismatch problem
 
Well, we have a plan
Would you like to do it, or shall I?
 
I'm already working on it
I have PyCharm in another window :P
which is how I should've caught the Flake8 failure, if I bothered to read the wavy line alerts xD
 
Oh, and I haven't installed any of my dev tools on this computer yet, so it would be a bit hard for me to do it :P
 
@angussidney :P
@angussidney Considering I do a lot of python work for my one workplace, and I'm permitted to use my own laptop and my own GitLab instance, I keep PyCharm + VCS handy all the time
And fun fact, I get it for free, because I'm still in college xD
!!/alive
uhm....
whoops.
 
I could get pycharm for free too, but I prefer Atom
 
note to self: stop killing screen
 
Restart: API quota is 3192.
 
!!/gitstatus
 
On branch deploy
Your branch is up-to-date with 'origin/deploy'.
nothing to commit, working directory clean
 
!!/blacklist-website this-is-a-test\.io
grrr now why are you dying
ooo that's interesting
fatal: could not read Username for 'https://github.com': No such device or address
oh i think i know why
 
12:06 AM
STDIN read error?
 
maybe
 
Morning @art
 
what's weird is it's for the push command...
 
@angussidney evening :)
 
@ArtOfCode I thought it was morning for you now too? Past 0000hrs?
 
12:09 AM
@angussidney oh hey, so it is
 
ooo there's the problem
 
alright so this should work now...
 
Restart: API quota is 9999.
 
!!/blacklist-website this-is-a-test\.io
 
12:12 AM
@AquariusOne You don't have code privileges, but I've created a pull request for you.
 
yay!
 
[ SmokeDetector | MS ] No whitespace in answer: How to use Xfce4-display-settings terminal commands? by Hdjdh on askubuntu.com (@ThomasWard)
 
hey it's working now :)
 
tpu- by teward
 
@angussidney so, all's working as it should, now. We just have to "pull" master, and it'll either work or fail.
@ArtOfCode Turns out that the box rebooted, and I forgot, so it 'forgot' the GH creds for "push"
and since it wasn't in interactive mode, it blew up.
though
it's not on the right branch...
!!/gitstatus
 
12:16 AM
On branch master
Your branch is up-to-date with 'origin/master'.
nothing to commit, working directory clean
 
and I know why.
and that one's an easy fix xD
 
Restart: API quota is 9971.
 
!!/blacklist-website this-is-a-test-gotta-get-deploy-not-master-after-pr\.io
 
@AquariusOne You don't have code privileges, but I've created a pull request for you.
 
!!/gitstatus
 
12:20 AM
@AquariusOne You are not a privileged user. Please see the privileges wiki page for information on what privileges are and what is expected of privileged users.
 
grr
 
!!/gitstatus
 
On branch deploy
Your branch is up-to-date with 'origin/deploy'.
nothing to commit, working directory clean
 
THERE we go!
correct branch, oorah!
 
@angussidney so, I've fixed the mismatched Master HEAD problem, and fixed a bug I thought was present, in where the PR is created, but the "return" triggers before we switch back to the Deploy branch.
i can rest easy now.
 
12:21 AM
^ that's still me
 
@ArtOfCode I seem to remember that you were interested in the pics of my completed build, so here they are: imgur.com/a/DNqTn
 
@Undo for situational awareness: thanks to @angussidney, the approach to the workflow was to pull Master, when we do the check, if the pull fails we trigger the HEAD error as normal. All's pushed up and fixed.
 
Well, not really thanks to me, you did all the hard work :)
 
@angussidney true, but Undo was going to go to a PC and then restate the workflow
then you came along while I was messing with the code :P
egads I spammed my inbox with GH alerts
 
@angussidney Nice. Clean cabling, too, which is more than I managed.
 
12:29 AM
CI on 064bd8a succeeded. Message contains 'autopull', pulling...
Restart: API quota is 9930.
 
!!/gitstatus
 
On branch deploy
Your branch is up-to-date with 'origin/deploy'.
nothing to commit, working directory clean
 
yay
small wins are small, but wonderful :)
In any case, problem solved, blacklist works again :)
And should anyone majorly torpedo Smokey and provide an invalid blacklist type somehow, we won't have an explosion.
 
@ArtOfCode the case has room for cable management along with hook n' loop straps at the back, so all I really had to do was stuff everything behind the motherboard, strap it down and shove the sidepanel on :)
 
@angussidney damn, that is sweet
what case?
 
12:33 AM
Phanteks Enthoo Pro
When it arrived I realised that it was much, much bigger than it looked in the photo
 
is that ATX or mini/micro ATX form factor?
 
@angussidney same with mine
 
@ThomasWard it fits EATX, but I have an ATX board in there
 
nice!
!!/gitstatus
 
On branch deploy
Your branch is up-to-date with 'origin/deploy'.
nothing to commit, working directory clean
 
12:35 AM
lies
that is a lie
!!/gitstatus
 
On branch deploy
Your branch is up-to-date with 'origin/deploy'.
Changes not staged for commit:
(use "git add <file>..." to update what will be committed)
(use "git checkout -- <file>..." to discard changes in working directory)

modified: gitmanager.py

no changes added to commit (use "git add" and/or "git commit -a")
 
interesting
 
!!/gitstatus
 
On branch deploy
Your branch is up-to-date with 'origin/deploy'.
Changes not staged for commit:
(use "git add <file>..." to update what will be committed)
(use "git checkout -- <file>..." to discard changes in working directory)

modified: gitmanager.py

no changes added to commit (use "git add" and/or "git commit -a")
Restart: API quota is 9904.
 
12:37 AM
yes i know it's modified, i think this is working...
 
!!/blacklist-website testing-123-testing\.io
oops
@ArtOfCode are we sure the gitmanager methods aren't static methods?
 
AIUI ClassMethods, they're good for alternate class constructors.
with the exception of the dirty branch, that's running as staticmethods, we would get rid of self since it doesn't refer to self.anything
(and save about 9 bytes xD)
 
@ThomasWard you should be an expert on Ubuntu (I hope), do you reccomend 14.04 or 16.04 for a new install?
 
@angussidney how old is the system, how old is the hardware, and specs please
@angussidney second question, do you prefer upstart/sysvinit, or SystemD?
@angussidney third question, how much do you hate VPNs and overriding DNS servers for use based on what connection you're on?
 
Restart: API quota is 9899.
 
12:47 AM
final test, before I turn off my alter ego
 
@ThomasWard hardware brand new (the build pictured), definitely fast enough to run any version. i5-6600k, GTX 1060 + 16GB ram. Desktop ubuntu btw, not server
 
!!/blacklist-website testing-123-testing\.io
 
@AquariusOne You don't have code privileges, but I've created a pull request for you.
 
YOU WORK! YAY!
I saved us 10 bytes xD
 
PPCG.SE
 
12:47 AM
I know, right?
 
@ThomasWard not much
@ThomasWard don't know much about either of those
 
@angussidney that said, the gitmanager.py methods, they don't need to be initialized with an instance of the class itself, it doesn't refer to either of the methods within each other, so therefore the self is just extra.
@angussidney final question:
AMD/ATI/Radeon or nVidia graphics?
 
nVidia
 
try 16.04 first, the Nouveau drivers might work, if not you'll have to download the binary drivers from nVidia and install them.
if that fails, 14.04
Use the LiveUSB first, to test things
It should work with 16.04
in fact I'd put 16.04 over 14.04, if only because I prefer the latest LTS :P
 
@ArtOfCode I seem to remember that either you or Undo were having troubles with 16.04 with MS a few months back, does that only apply if you're hosting, or does it matter if you're developing too?
 
12:52 AM
@AquariusOne not a clue
 
@ArtOfCode well, I got angry at the thing, so I converted to @staticmethod and it's all working right
 
@angussidney I use 16.04 locally for dev without any issues, and I think the MS server is on 16.04 now too
 
(AquariusOne is my unpriv. sockpuppet xD)
 
@ThomasWard I know :)
 
@ArtOfCode fun fact, I found unreachable code in ws.py
 
12:54 AM
16.04.1 it is then
 
o/ @Undo
 
The issue was something in 16.04 not playing nice with MySQL, but someone patched that
 
hail the commit storm >:D
 
@ThomasWard Oy!
 
@ThomasWard get rid of it :)
 
12:54 AM
@ThomasWard there is a lot of unreachable code all over our codebase
 
Finally at a real machine. With big monitor and all.
 
If you find useless code, just get rid of it if you're sure
 
@ArtOfCode Heck no
 
@Undo you can thank @angussidney for helping me deploy a fix for the mismatched/detatched head stuff
 
if things break we'll just rollback
 
12:55 AM
@ArtOfCode brilliance of VCS :)
 
@Undo oh, I thought it was
 
Thanks @angussidney :)
Welcome to Ubuntu 14.04.5 LTS (GNU/Linux 3.13.0-91-generic x86_64)
 
Fair enough. I don't see why it couldn't be 16.04, though, given that my local works
 
@angussidney Yeah, I (believe I) was specifically having issues with how 16.04's management of disk resources was playing with MySQL.
I wouldn't worry about it for a desktop box.
 
Still don't know why people are thanking me, I only restated this: i.stack.imgur.com/CCh9G.png
 
12:56 AM
@angussidney you're important now remember ;)
 
fun fact: 25" monitors are really, really nice after using a lap-mounted 13" for a week
 
@Undo Yep. I feel the difference with a 21.5" monitor and a 15.4" laptop, so I can imagine.
 
@Undo you don't need that last part. Just "really, really nice"
 
we use dev, test and prod
dev is master
 
@YvetteColomb This sounds fairly sane
Might be overkill for what we need, though, since we don't really have a dedicated test environment
@Andy I actually like it better than a 27"
 
1:01 AM
I gave a 27" monitor I had to a friend because it didn't line up nicely with my dual 24s. He loves it, but only uses the single monitor, whereas I have moved between dual and triple setups but don't like using a single monitor when doing dev work
 
@Andy Yeah, I usually run the 25 and 13 below it. Not ideal, but it works fine for what I do.
also @ThomasWard, you really don't need anyone's permission to test stuff on prod. The extent of possible collateral damage is "there are a bunch of emails in Undo's inbox", which is a couple clicks for me and doesn't matter.
 
I need my monitors to match/line up nicely and 27/24 ones just didn't do that
 
I want a 27" Oreo.
 
Nice auto-correct there :)
Mmmm...oreos
 
@Undo and no other food for a week, is it?
 
1:04 AM
> Where diamonds are made, smoke is detected, and we break things by developing on production.
 
OK. I'm done on the phone for a while.
 
The worst that could happen is @hichris123 would complain that you have spammed his inbox with GH emails
:P
 
@Undo oh, indeed?
Cool.
 
@Undo hahaha I just emailed someone saying I pushed to test, no one has checked it and he's saying we should push it to production - without being checked!!!! The timing is impeccable.
 
Well, the extent of collateral damage is I fubar my box, in which case I keep a backup for good reason ;)
 
1:05 AM
CI on c8ebabf succeeded.
 
there's two commits, I expect it to whine :P
 
@ThomasWard Yeah. We should probably have some way to back up user blacklists, etc. but... they really don't matter much.
 
!!/pull
@Undo thankfully I rsync weekly between Aurora and esx2
wat "Invalid base"
what did I break...
 
That sounds like a fun Git error
 
1:07 AM
@Undo don't we do a sync with MS with the website blacklists etc? We could probably set up something similar with user blacklists/whitelists
 
@angussidney That was mostly abandoned before it reached a functional state. We put them in text files synced to the repo instead.
 
i had to clear the breakage manually XD
 
Restart: API quota is 9834.
 
@angussidney we don't, at the moment. We have the blacklist functionality there on metasmoke, but it's not used.
 
1:07 AM
@Undo actually they're on gitignore
so I'm not sure they directly sync
not 100% certain on that though
 
It is functional, it's just unused.
 
@ThomasWard Oh, I mean the website blacklists.
 
ah
!!/pull
 
Good clarification
 
1:08 AM
yay you didn't die this time.
@Undo this is one of those commit / merge sets where I had to manually resolve it xD
 
Restart: API quota is 9829.
 
Yeah, we need to fix that. Whoever's running it shouldn't have to babysit it.
Back in the day, I could run it on a Pi in my closet for months at a time without touching it
 
indeed.
 
Back in the day when we wern't considering integration
 
and then the Pi broke :P
 
1:10 AM
and when your SD card wasn't corrupted
 
@Undo it might be one of those things where we have an API call to a separate shell script that executes the pull, then calls the restart command to restart Smokey
 
I got ninja'd
 
just a thought :P
 
@angussidney I don't think it's actually corrupted, really. More 'Undo broke something with how he mounts the disk Smokey runs off'
 
like a !!/forcepull xD
 
1:10 AM
... hypothetically, can a Pi boot off a USB drive or similar?
 
@Undo Hypothetically, yes, but it's been evil when I've tried
 
nothing's been autoflagged in the last 12 hours
 
@ArtOfCode Smokey's been getting beaten to hell and back the past 8
 
@ArtOfCode Not spammy hour yet.
 
@Undo probably, but it's a slower interface
 
1:11 AM
also not spam hour yet
 
@ArtOfCode Meh on that one for my purposes.
huh. "raspberry pi boot from usb" autocompletes, so maybe I've been down this road before
 
@Undo sure, but yesterday we still had a trickle of posts over the day outside of spammy hour
 
@ArtOfCode Last one that shows up on metasmoke.erwaysoftware.com/flagging/conditions/3/edit is from 12 hours ago, and I mostly trust that logic
 
I'm still boggling at the 11683/5 tp/fp rate
 
@Undo if you have a Pi 3, they introduced a bunch of network and USB boot features
 
1:13 AM
@angussidney Pi 2 (B?)
 
@ArtOfCode I don't get how meta can complain about that
@Undo I think you might be able to flash it
Lemme find the blog post
 
@ArtOfCode Yeah, testament to how well folks have been tuning it
Is there any good test for whether the card's corrupted in the first place?
It boots and I haven't found any oddities other than mounting the drive, so...
[feature-suggestion-slash-rfc] Post a thing in here any time a post is autoflagged, then subsequently receives negative feedback.
It shouldn't be all that noisy. If it is, we deserve it.
 
@Undo stick the card into your computer, fsck it
Input/Output errors indicate card death
unfixable I/O errors*
 
CI on 80ec58b succeeded.
 
1:18 AM
@ThomasWard And therein lies the issue. I've lost the microSD adapter thing
 
In the "can't find it after five minutes of looking" sense. It'll turn up eventually.
 
@Undo you in the USA? I'll send you one of my spares
if not, then bleh
 
@Undo +1
 
:P
!!/gitstatus
 
1:19 AM
On branch deploy
Your branch is up-to-date with 'origin/deploy'.
nothing to commit, working directory clean
 
!!/rev
 
@ThomasWard Yeah, but that's kinda unnecessary.
I mean really, it's $4
 
This will probably explode, in which case I'll revert.
!!/pull
 
[ SmokeDetector ] SmokeDetector started at rev 80ec58b (Thomas Ward: Convert off of ClassMethods.) (running on teward/Aurora)
 
1:20 AM
Holy crap you work
 
You should probably call the sloMo guys to record if you want to blow something up
 
How does staticmethod work?
 
Restart: API quota is 9798.
 
1685
Q: What is the difference between @staticmethod and @classmethod in Python?

Daryl SpitzerWhat is the difference between a function decorated with @staticmethod and one decorated with @classmethod?

 
@Undo basically the same as a standard def foo(self, bar,baz): ... but without the self argument
 
1:21 AM
oh, well that makes a heck of a lot of sense
 
Where self is never called in a method, you can make it a Static Method, which then means the method doesn't also have to be able to call to its parent class
 
Should have been doing that since the beginning.
 
@Undo yeah, well, the constant wavy lines about 'cls' vs. 'self' were angering me
 
IDE warnings drive behavior quite well
 
Rob
Having never used python before - I was under the impression all methods were static since you need to explicitly pass in the self (aka this) argument?
 
1:22 AM
and, 'cls' passes an instance of the class, so unless we're using a classmethod to allow for the class to be constructed differently, we should avoid the use of classmethod
We don't even need a decorator for something that calls (self, foo, bar)
 
Yeah, that makes sense. Thanks for that explanation.
 
@ThomasWard Python uses self sometimes, plus some of us are Rubyists which also uses self, so that'll be why
 
AFK now, to return eventually.
 
@Rob probably, but it helps to decorate it to avoid confusion.
@ArtOfCode True, though in none of the places the last two commits change, is self.something called
 
Rob
Fair enough
 
1:24 AM
so for those functions, they're staticmethod now. That said, the ones that do call self.something are just standard instance methods
without being decorated as a classmethod
and so far, Smokey's running well
someone should spam xD
Ooh, fun fact, Smokey's taking up less RAM now xD
 
Should I report a random question again?
 
@JanDvorak up to you, but if it processes something and breaks, I need to revert my changes :)
 
My Inbox ATM... it's a good time to be a gamer
 
and the only place I can flag right now is Ask Ubuntu (E: Flag Exhaustion) and a flag from me there is the equivalent of a nuke, so....
 
oh yay ubuntu download finished
 
1:27 AM
heheheh
nice, @angussidney
!!/report
oops
 
@ThomasWard Not enough arguments.
 
[ SmokeDetector | MS ] Manually reported question: Disregard, this is a test authorized by a moderator by AquariusOne on askubuntu.com (@ThomasWard)
 
oh there we go :)
@SmokeDetector gone
 
Restart: API quota is 9758.
 
1:30 AM
wat
 
I think SD was gone for a sec
 
Was this a legitimate test ? askubuntu.com/q/865219
 
I believe Thomas authorized his bot to perform it.
 
@Serg yes, AquariusOne is my sockpuppet
 
OK. Just czech-ing
 
1:32 AM
note the 'Deleted by Thomas Ward" button on it
 
Well, I noticed that. I didn't know you had a bot
 
CI on d3f2d03 succeeded.
 
what was this test for?
 
@KazWolfe I made codebase changes to Smokey
I needed to make sure all the functions work right
they don't, so...
foobarbazREVERT
 
oh.
 
1:34 AM
OK then. Back to AU general . . .
 
@Serg @KazWolfe yep, shoulda said something, but I lazy :)
alrighty, so I've reverted the DeletionWatcher changes...
 
Maybe we should wait until there is actually spam before we test next time, to avoid confusion?
 
[ SmokeDetector ] SmokeDetector started at rev 42c8275 (SmokeDetector: Revert "Convert off of ClassMethods.") (running on teward/Aurora)
Restart: API quota is 9729.
 
@angussidney Probably a good idea. DeletionWatcher is the only thing that was dying
at least in that run. At least we fixed that part xD
s/we/I/
if it dies again there's only one more commit to undo for that one.
oops, damn
i forgot to run in screen >.<
 
lol oops
 
1:39 AM
[ SmokeDetector ] SmokeDetector started at rev 42c8275 (SmokeDetector: Revert "Convert off of ClassMethods.") (running on teward/Aurora)
Restart: API quota is 9720.
 
well, in any case, Smokey's not dying, we'll see if it works :)
 
metasmoke is not happy.
and are you testing in prod?
 
Real programmers don't have no time for testing.
 
um...
 
> and we break things by developing on production
 
1:43 AM
@KazWolfe read the topic.
@KazWolfe Yes, MS is not happy, but this is an unfortunate evil. It will eventually calm itself.
 
nothing should even go on the testing server until 95% unit test coverage with pass, and documentation is fully complete.
 
ooh that's interesting
apparently, there was a hung Smokey instance
 
We can revert any changes on production within about 10 seconds, so testing doesn't matter too much for us
 
i think that was breaking things :P
 
Bad practise, but hey, it works for us
 
1:45 AM
@angussidney faster when the person who made the change is the one who fubar'd and is immediately reverting their breakage
esp. with direct access to the Smokey console :)
well, in any case, it's behaving... I think.
!!/alive
 
@ThomasWard Of course
 
Also @ThomasWard creating bootable USB drive with rufus, should I format it as FAT32 or NTFS?
 
@angussidney I'd use FAT32 for the drive, not NTFS, because of weirdness
 
@angussidney FAT32.
 
I always format my bootable drives as FAT32, because BIOS can read it easy XD
back in a bit
and I am returned :)
 
1:50 AM
Who returned you? :-)
 
the goat.
 
@JanDvorak /dev/urandom finally spit me back out xD
 
How did you end up down there?
 
he got put into the entropy pool by the kernel.
 
I tripped into /dev/null, which redistributed my memory structure throughout the kernel, and /dev/urandom finally spat me out after the kernel reconstituted me into the entropy pool :)
 
1:54 AM
Now I finally know why I fear the Tux :-)
 
actually that's a good idea. anything sent to /dev/null should enter the entropy pool.
 
I feel like this might yield some vulnerabilities.
 
it probably would
unless sent through five SHA512 encryption passes with five separate random salts from the entropy pool
 
i wonder if there's an ACE bug in smokey. that would be interesting.
 
@KazWolfe I'm tired, ACE = ?
 
1:57 AM
Arbitrary Code Execution.
 
95% chance of that
RCE-as-a-feature
though I don't see any ACE risks... yet.
 
That would be bad. Much worse once autoflagging's on
 
mhm
 
have a spammer to figure out how to flag everything, and ignore their own posts.
that is some next-level spamming.
 
Talking about that, we should probably set up Metasmoke to send an email to admins when too many posts get flagged
 
1:59 AM
or use smokey (and its privileges) to spamblast the entire network. Comments, answers, questions, everything.
go spam meta for once.
 

« first day (1137 days earlier)      last day (2668 days later) »