« first day (2424 days earlier)      last day (1404 days later) » 
00:00 - 13:0013:00 - 00:00

1:04 PM
[ SmokeDetector | MS ] Potentially bad keyword in body (1): SQLDelight v1.4 not generating interface by Julien Banse on stackoverflow.com
 
@iBug If there's no index on that field, and a lot of records, that sounds normal
 
^ that
 
tpu- feedback received on I want to know I am okay or not [MS]
 
!!/watch node41--nsxodg\.repl\.co
 
[ SmokeDetector | MS ] Offensive answer detected, toxic answer detected (158): Is there a way to see how many covenant items I have turned in? by Anon on gaming.SE
[ SmokeDetector | MS ] Potentially bad ns for domain in answer (1): i am facing error while running uipath bot related to Excel by Saurabh Nimbade on stackoverflow.com
iBug/Coral: Executing automatic scheduled reboot.
[ SmokeDetector | MS ] Potentially problematic ns configuration in answer (1): Has the lifespan of farmed animals decreased? by Michael Tran on skeptics.SE
 
1:31 PM
@SmokeDetector fpu-
 
!!/watch centralparkpuppies\.com
 
@JeffSchaller You don't have blacklist manager privileges, but I've created PR#4148 for you.
PR#4148 ("Jeff Schaller: Watch centralparkpuppies\.com") opened by SmokeDetector
[ SmokeDetector | MS ] Username similar to website in answer (62): Setting nautilus-open-terminal to launch Terminator rather than gnome-terminal by alvaroferran on askubuntu.com
fp- feedback received on Using Pynput to type into discord [MS]
 
@Makyen I was looking into GitHub Actions a little further, and I think I've determined the problem. GitHub Actions don't get a tty, and so when all the commands we run check to see if they should use colored output, they choose not to because they see we're not in a tty. It seems like there's not really a non-hacky way to get around that, so I think that means we're sticking with Circle. I do like the redundancy though, especially with our sporadic (hopefully gone, but possibly not) failures.
I think the ideal would be if there was a way to turn off notifications for the GH actions.
 
[ SmokeDetector | MS ] Link at beginning of answer (33): godot tilemap tearing between tiles by Артем Постников on stackoverflow.com
[ SmokeDetector | MS ] Potentially bad keyword in body (1): Cannot connect to google compute engine VM ✏️ by Bae HyeonWoo on stackoverflow.com
naa- feedback received on godot tilemap tearing between tiles [MS]
 
1:49 PM
@thesecretmaster There is... I don't get any
can't remember where, though
 
@thesecretmaster The email issue is something I can, and should, handle. It's my problem. It's definitely not something that should prevent us from using something that's useful. My earlier statement about it was mostly an expression of frustration, in part because, at the same time, I also started getting emails about GH Action failures on my fork of the MS repository.
I'm sorry if what I said implied we shouldn't be using GH Actions due to me getting a bit of email. That wasn't my intent.
 
fp- feedback received
 
@thesecretmaster That's ... strange, because there are some escape codes for tty font color changes in the logs that are displayed for that portion of the display. Prior to mentioning the color issue, I had looked at the raw data that was fetched in a GET in order to see if the escape codes were being stripped locally (which I could have corrected with a userscript or browser extension). I assume the output is from multiple different programs, some of which are detecting the lack of a tty.
 
Yeah, that's what's happening. Here's the color code for minitest
I could monkey-patch this into working for minitest, but that doesn't really solve the problem since I'd love colored output in all of the commands, not just tests.
Really, it's just dumb that GH doesn't give you a tty
In theory you're on your own VM, what's stopping you from getting a tty.
 
[ SmokeDetector | MS ] Few unique characters in answer, mostly punctuation marks in answer, repeating characters in answer (203): How long was Tom Paris' prison sentence? by Ted Heavy on scifi.SE
tpu- feedback received
 
2:11 PM
I agree, colored output would be helpful for all of the commands. I'd also noticed its lack on rubocop's output. It is strange that GitHub both supports tty coloring in the output, which is specifically converted to HTML/CSS in the client, but doesn't present a tty to the test environment. I'd consider that a bug/misfeature on their end.
 
[ SmokeDetector | MS ] Post is mostly images (0): how to show URL of SVG in image tag? by Catalin V. on stackoverflow.com
 
Yeah, the nature of the problem is that it does support the color escape codes, but doesn't provide a tty, which many commands take to mean that they shouldn't output color escape codes.
 
teward/Osiris: Executing automatic scheduled reboot.
[ SmokeDetector ] SmokeDetector started at rev 8ca8776 (SmokeDetector: Auto watch of node41--nsxodg\.repl\.co by Machavity) (running on teward/Osiris, Python 3.6.9)
naa feedback received
Restart: API quota is 13750.
 
It kinda makes sense because commands shouldn't be interactive and therefore don't need a tty, but it's also just a fact that many commands use that as a proxy for choosing to be colored or not.
 
Yeah, that's a very common way for command line programs to make that choice.
When they do, there's usually a command line argument to force using color, but it's different for each command, and some don't provide such an option.
 
2:16 PM
Also, I was always under the impression that bash -i -l always allocated a pty, but apparently I was mistaken.
 
It is strange/unfortunate that GH Actions doesn't have the option of presenting a pty to the commands.
I was under the impression that would work under most conditions, but I must admit that it's been quite a while since I've checked that portion of the man page.
 
fp- feedback received
 
As to the color for minitest, I'd suggest adding the config for it to the GH Action, if you haven't already. It's valuable to make the interface as useful as possible.
 
In order for minitest to have colored output, it needs BOTH the --color flag and a tty, which means that I'd need to monkey patch to override that.
 
Which is also relatively stupid, because the color flag should force colors :/
 
Yeah, that is stupid. If you're giving a command line option like that, it's specifically for overriding any automatic detection which the program may be doing. That the program doesn't actually override what it's detecting is ... wrong/a bug, IMO.
 
Although, the "right" solution would be to get GH to fix their bug...
 
True.
 
I could probably write up a little C wrapper to allocate a pty and use it, but I feel like that's kinda hacky and gross.
Here's the relevant issue if anyone's curious.
 
2:31 PM
It is, but if it makes the displayed output better, then it's a win. OTOH, I'd be really surprised that something to do that doesn't already exist.
 
[ SmokeDetector | MS ] No whitespace in title (68): UnexpectedValueException by brooll on magento.SE
 
Also, as a fun note, I stumbled across this issue which, while not useful to us, happens to be related to SE.
 
@SmokeDetector f
 
What I don't get is why GitHub wouldn't support providing a pty interface. Providing one seems like it's an obvious thing to do, at least as an option. One would think that the people developing GitHub Actions would have encountered the issue, if they were actually using GitHub Actions in their own development.
 
2:40 PM
[ SmokeDetector | MS ] Bad keyword with email in body (57): Is Learning this path enough to become a DevOps/SysAdmin? by Izmilia Prastika on stackoverflow.com
tp feedback received
[ SmokeDetector | MS ] Link at beginning of answer (33): What is PDO & why should I use it? by Jean herard Jean on stackoverflow.com
naa- feedback received
 
!!/watch- fctraining\.org
!!/watch- ismilia@hotmail\.com
 
[ SmokeDetector ] Blacklists reloaded at rev c75378f (SmokeDetector: Auto watch of fctraining\.org by Makyen) (running on teward/Osiris)
[ SmokeDetector | MS ] Link at beginning of answer (33): MySQL, MySQLi, MySQL Prepared Statement or PDO? by Jean herard Jean on stackoverflow.com
naa- feedback received
 
!!/approve 4148
 
2:56 PM
@Makyen Looks like Cody destroyed the user. Probably should TP those last 2
 
@Machavity Thanks.
 
[ SmokeDetector | MS ] Potentially bad keyword in body (1): Why ChartOpen function is opening 2 charts of same symbol instead of just 1? by Umar on stackoverflow.com
[ SmokeDetector | MS ] Pattern-matching website in body, potentially bad ns for domain in body, potentially bad keyword in body (98): supplant my GPU by Chris Jackson on askubuntu.com
[ SmokeDetector | MS ] Link at beginning of answer, link at end of answer (93): How do I create a post title that is a link in Blogger interface? by Manthan kushwaha on webapps.SE
[ SmokeDetector | MS ] Link at end of body (99): Do you want to order any cannabis product? by Leafedin Pro on superuser.com
 
3:11 PM
@SmokeDetector tpu-
 
tpu- feedback received on supplant my GPU [MS]
 
!!/watch leafedinpro\.com
 
[ SmokeDetector | MS ] Potentially bad keyword in answer (1): ArXiv/Problems with downloading source ✏️ by PoreyS on tex.SE
[ SmokeDetector | MS ] Potentially bad keyword in answer, toxic answer detected (81): WPF - Webbrowser - getElementById by JAKE PRYCE DAVIES davies on stackoverflow.com
fp- feedback received on ArXiv/Problems with downloading source [MS]
tpu- feedback received
 
3:32 PM
@SmokeDetector k
@SmokeDetector why
 
@πάνταῥεῖ Potentially bad keyword in answer - Position 567-577: extract.me
 
@SmokeDetector f
 
@thesecretmaster I've been assuming that the GH Actions and CircleCI were more-or-less running the same tests and should get similar output. When looking at the GH Actions, I've been looking at the CircleCI and GH Actions results for this commit, because it's the most recent on the master branch which has test errors.
However, the results I'm seeing from the tests are different between GH Actions and CircleCI.
The GH Action rails test results for this commit are here, which show two tests failing and the rest passing. The first run of the CircleCI rails-tests for that commit are here and show the majority of tests failing. The CircleCI failures appeared to be a general testing failure, so I re-ran the workflow.
The results for re-running the CircleCI workflow show all tests passing. That's 3 different results from running the tests 3 times. I haven't yet tried re-running the GH Actions for that commit, as I wanted to point this out to you first, in order to be on the safe side.
I'm not familiar enough with Ruby/this test environment to know if the change you made in the next commit is specifically addressing this, or if it's something that's gotten overlooked.
Is the situation just that the all-passing result from the re-run of the CircleCI workflow is what CircleCI should have gotten and the two tests failing in GH Actions is the only real issue there? Is that difference between the GH Actions result and your expectation of no CircleCI failures (as seen in the re-run of CircleCI) what you addressed in the next commit? If there hadn't been the failures in the initial run of CircleCI, that's what I would have assumed was happening.
However, having specifically seen the discrepancy between the three runs, I felt it was something I should double-check with you.
 
Yeah, what's happening there is (I assume) the Circle tests had the big multithreading error the first time, and then passed the second. The GH tests were failing for a legitimate reason, which I fixed in the next commit. The only mystery remaining is why Circle didn't catch those same failures, which I have no answer to.
But yes, they're running almost the same test verbatim
npm test for JS, bundle exec rubocop for ruby style, and rails test for ruby tests.
 
Yeah, that's what I understood and expected. It was just strange to see the three different results, and I wanted to double-check I wasn't misunderstanding what was supposed to be going on. Thanks.
 
3:44 PM
fp- feedback received
[ SmokeDetector | MS ] Potentially problematic ns configuration in body (1): module-info when one of my external libraries doesn't have a module-info by Calmakestrash on stackoverflow.com
[ SmokeDetector | MS ] No whitespace in title, repeating characters in title (133): dsadasdaaaaaaaaaaaaaaaaaaaaaaaa ✏️ by Dan Volovsky on stackoverflow.com
tp- feedback received
[ SmokeDetector | MS ] Potentially bad keyword in body (1): Atom live server throwing up "internal server error" when used by Luis Santos on stackoverflow.com
fp- feedback received
 
@Makyen That would be great. Is it okay for me to query every 90 seconds in the next ~7.5 hours?
 
user435118
 
@user12986714 That doesn't sound unreasonable.
 
[ SmokeDetector | MS ] Bad keyword in body, pattern-matching product name in body (193): STAY IN THE KNOW AND GET INVOLVED by Anton Ross on superuser.com
tpu- feedback received
 
@SmokeDetector k
 
[ SmokeDetector | MS ] Link at beginning of answer (33): List comprehension in Ruby by Sam Michael on stackoverflow.com
 
!!/watch- colorado4wheel\.com
 
4:15 PM
fp- feedback received
 
@Makyen Now that thesecretmaster has fixed the GraphiQL bug, should the GraphQL link be removed?
 
@double-beep meh, but probably.
 
user435118
4:38 PM
Ooh, yay. Finally a spam call
 
user435118
Saying a car accident wasn't my fault, I don't even drive. It's even illegal for me to drive :)
 
iBug/Sandy: Executing automatic scheduled reboot.
 
How long did you play along for?
 
[ SmokeDetector | MS ] Potentially bad keyword in body (1): Conditional html element only displaying temporarily by docphilstone on stackoverflow.com
 
user435118
@JohnDvorak I managed to do 5 minutes before they put the phone down :P
 
4:51 PM
 
user435118
This is one of the best adverts ever imo: stashmedia.tv/money-supermarket-money-calm-bull-china-shop
 
tpu- feedback received
 
!!/watch- oneglobus\.com
 
[ SmokeDetector ] Blacklists reloaded at rev 901552e (SmokeDetector: Auto watch of oneglobus\.com by Makyen) (running on teward/Osiris)
[ SmokeDetector | MS ] Link following arrow in body (97): I am facing error of python migrate in Django by Muhammad Yeamin Adib on stackoverflow.com
[ SmokeDetector | MS ] Bad ns for domain in body (97): No se como explicar el comportamiento de Spring Boot + Artemis by Developer on es.stackoverflow.com
[ SmokeDetector | MS ] Potentially bad ns for domain in answer, potentially bad keyword in answer (2): Using .Net Core with OpenGL on LInux by Microsoft XNA Framework Redist on stackoverflow.com
 
5:31 PM
!!/watch youfree\.tech
 
@Machavity 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.
 
5:46 PM
[ SmokeDetector | MS ] Potentially bad keyword in body (1): Dynamic PHP Pagination Mysqli by Khemil Jariwala on stackoverflow.com
 
@thesecretmaster If I'm using GraphQL to get information on posts from MS and the overall list of posts for which I want information is > 100 (up to a couple/few hundred), is it better, from an MS performance POV, for me to
A) make a first request listing all the urls/uids/ids for the posts I'm going to want; then, after I get the response containing the first 100, make another identical request, except use offset to receive other/more results; or
B) make one request with 100 posts; get response; make another request with a different list of posts; get response?
Does the answer depend on the total number of posts I want (e.g. 120 posts, 150 posts, 250 posts, 500 posts)?
 
I'd have to check exactly how the GQL queries translate, @Makyen, but if the number is still in the hundreds, either should be fine. If it was 5-digit, then you'd be better off with B to avoid the performance drop of an ID list with SQL OFFSET.
 
6:02 PM
fp- feedback received
 
@ArtOfCode Thanks. I'd figured there was probably some break-point/gradual difference in there where there's a tipping point in a trade-off between doing one query with a moderately higher number of posts, which is then, presumably, cached and further identical requests getting other lines are much faster vs. multiple smaller requests.
 
There isn't any caching on GraphQL, AFAIK
 
and if there was, a different OFFSET value would count as a different query and wouldn't be cached
 
@Makyen Also, if you've got a use case for >100 posts, I can bump up that limit.
IIRC it was pretty arbitrary.
 
[ metasmoke ] ci/circleci success on 5012a38: Your tests passed on CircleCI!
 
6:11 PM
@thesecretmaster The time when I'd use more is for an initial load of a page from The Fire Department, where there can easily be more than 100 SD reports. The limit should be set to whatever makes sense from an MS performance POV. It's easy for me to chunk things (already doing that for the V2 API).
 
I can't see a reason why it'd be particularly expensive MS side. Do you have an example query I could play with?
Also, we should probably put some kind of safeguard on graphql to make it not a DOS risk.
 
I probably wouldn't go much higher than 100, @thesecretmaster - it does take noticeable time to load and deliver the extra posts.
Just as easy to split stuff into chunks of 100
 
[ SmokeDetector | MS ] Potentially bad keyword in body (1): How to update RSS Feed update every 5 seconds in Python using Flask by Sebanti on stackoverflow.com
 
@thesecretmaster It currently doesn't require an API key, so that should probably be added, not that doing so really will stop someone.
 
I might just cap query depth + complexity for folks who aren't core / don't have an API key.
 
6:18 PM
@thesecretmaster The query I'm currently looking at is (you should be able to copy & paste the following into the JavaScript console and get a result):
var query = `query ($ids:[ID], $uids:[String], $urls:[String], $first:Int, $last:Int, $offset:Int) {
  posts (ids:$ids uids:$uids urls:$urls first:$first last:$last offset:$offset) {
    autoflagged,
    body,
    comments {
      created_at,
      id,
      text,
      updated_at,
      user {
        ...userFields
      }
    },
    created_at,
    deleted_at,
    downvote_count,
    feedbacks {
      chat_host,
      chat_user_id,
      created_at,
      feedback_type,
      id,
      message_link,
 
Oh that's a lot of query
 
I can trim it down, somewhat, but the intent was to share a single request between FIRE, AIM, what @Daniil is working on, and have it available for any other scripts running in chat.
 
Yeah, it's not a problem, I was just surprised by how much query there was
 
About the only part in there that FIRE doesn't currently use is the spam_domains. It doesn't use every property that's on Post, but the ones which it doesn't use appeared to be directly in the post table, so, I believe, have very little cost to request. Although, if there's a cost to something like feedbacks_count, I can easily calculate that value, given that I need the actual feedbacks anyway.
 
No that's super cheap, basically if you're requesting a record anyways it costs nothing to look up a few extra fields.
 
6:28 PM
That's how I understood it. :;
 
7:18 PM
[ SmokeDetector | MS ] Bad keyword with email in answer (85): How do airline pilots contact their company while in the air? by Soheir on aviation.SE
teward/Osiris: SE WebSocket: recovered from websocket._exceptions.WebSocketConnectionClosedException: Connection is already closed.
There's been no scan activity for 3 status pings. There may be a problem.
There's been no scan activity for 3 status pings. There may be a problem.
There's been no scan activity for 4 status pings. Going to reboot.
[ SmokeDetector ] SmokeDetector started at rev 901552e (SmokeDetector: Auto watch of oneglobus\.com by Makyen) (running on teward/Osiris, Python 3.6.9)
Restart: API quota is 10259.
 
7:39 PM
[ SmokeDetector | MS ] Bad keyword with email in body (57): I want to hire someone to create chart layouts and workspaces in TradingView by Tony_N on stackoverflow.com
tpu- feedback received
 
7:57 PM
[ SmokeDetector | MS ] Blacklisted user (71): Using Google App Scripts/Sheets for a website CMS / REST API by Richard on stackoverflow.com
[ SmokeDetector | MS ] Potentially bad keyword in body (1): Dynamic PHP Pagination Mysqli ✏️ by Khemil Jariwala on stackoverflow.com
 
@SmokeDetector tp- They've disclosed their affiliation in the past and in their profile, and they don't appear to have been warned about the policy, so I've warned them in a comment on this one.
 
fp- feedback received
 
user435118
@Glorfindel Is there a way to automatically flag comments from your flaggable comments query?
 
standard automation warning: make your automation intelligent or prepare for your account to be suspended
3
 
user435118
@ArtOfCode why suspended? I have done testing and 95%+ of those comments will be removed automatically by immediately raising a comment flag
 
8:08 PM
what about the other 5%
 
user435118
I will be monitoring the site I will automate on and review all flags not handled automatically after the script is run
 
user435118
Would there be an easy way to do this auromation?
 
fp feedback received
 
8:25 PM
[ SmokeDetector | MS ] Potentially bad asn for hostname in answer (1): Pasting an image to outlook 2013 using vba through Late Binding produces a small image by a deleted user on stackoverflow.com
[ SmokeDetector | MS ] Potentially bad keyword in body (1): 'Invalid Column Name' Error After Upgrading from Asp.net core 2.0 to 2.1 by Kevin M on stackoverflow.com
 
8:46 PM
[ SmokeDetector | MS ] Potentially bad keyword in body (1): How to update RSS Feed every 5 seconds in Python using Flask ✏️ by Sebanti on stackoverflow.com
[ SmokeDetector | MS ] Potentially bad keyword in title (1): Passing variables rather than hard coded data to paypal buy now button by Hailrig on stackoverflow.com
 
9:06 PM
@Daniil api.stackexchange.com/docs/create-comment-flag I guess? If you want to flag comments, better talk to Andy ...
 
fp- feedback received
[ SmokeDetector | MS ] Potentially bad ip for hostname in title (1): reshape.cc:55 stretch_dim != -1. Node number X (RESHAPE) failed to prepare by vlad on stackoverflow.com
 
@Daniil you could write yourself a script to use the SE API to flag stuff
 
9:26 PM
[ SmokeDetector | MS ] Potentially problematic ns configuration in answer (1): History of the Wehrmacht after the surrender by berni on history.SE
 
9:40 PM
okay, who linked here from the SO election :P
sudden influx of new faces
 
[ SmokeDetector | MS ] Bad keyword with email in body (57): Creation of characters on demand with Blender by Olivier Saccard on blender.SE
 
user435118
9:57 PM
@Glorfindel That page is confusing, no idea how to use it. Andy seemed quite ignorant of me so I give up then
 
user435118
@ArtOfCode Machavity in their Q&A answers
 
figures
 
user435118
Apparently everything is documented here
 
user435118
It would be even more fun in the main nomination post linked here but the link here isn’t worth the amount of characters it takes up :)
 
[ SmokeDetector | MS ] Potentially bad keyword in body (1): extract data from elementor webhook with php by Roberto Cholaky on stackoverflow.com
fp- feedback received
 
10:16 PM
@Daniil If you're targeting these deliberately... please don't. It's not hard to figure out what the regex for immediate removal is; if we wanted them automatically removed across the board we would just do that.
I frown upon people deliberately triggering the 1-flag criteria en masse.
 
@ArtOfCode I told them you have stickers :P
 
@Machavity *hides stickers*
Stickers? What stickers?
 
!!/watch- play\.story\.creator(?:@gmail\.com)?
!!/watch- play-story-creator\.com
 
user435118
@ArtOfCode The overpriced ones :P
 
10:27 PM
@Daniil did you see my message yesterday about that price being literal bare-minimum-cost-covering?
as in "I do not set the price, I pass it on from the supplier"
and as in "If I go any lower I am losing money"
 
user435118
@Undo There are user(s) who gets a bot to flag a thanks comment every 5 min
 
I know
 
user435118
So why don’t you say that to them?
 
having said it to you doesn't exclude also having said it to them, does it?
 
user435118
@ArtOfCode I think I did, the issue is not the price but the shipping
 
10:32 PM
The shipping that costs a whole 41p?
 
fp- feedback received
iBug/Sandy: Executing automatic scheduled reboot.
 
user435118
@ArtOfCode 1. As if you’re going to send me just 1 sticker 2. I don’t want to give you my address
 
costs 41p whether it's 1 or 10 :P
 
user435118
@ArtOfCode Point 2 still remains
 
yeah well that's your call not mine
 
11:15 PM
[ SmokeDetector | MS ] Link following arrow in body (97): how to change position related box of post by Farzad persian on stackoverflow.com
fp feedback received
Makyen/EC2-linux: Executing automatic scheduled reboot.
 
@SmokeDetector This is either a failed attempt at spam (the link at the end), or an attempt to evade text limits by copy+pasting the contents of the Stack Overflow HTML formatting help. Either way, it's abuse.
 
11:35 PM
[ SmokeDetector | MS ] Potentially bad keyword in body, potentially bad keyword in title (2): Forex trading on WordPress by Remilekun Odeyale on stackoverflow.com
[ SmokeDetector | MS ] Potentially bad keyword in body (1): Cracking SHA1 Hash + Salt Help ✏️ by HAAAASAN on security.SE
 
11:56 PM
fp- feedback received on Forex trading on WordPress [MS]
fp- feedback received
 
00:00 - 13:0013:00 - 00:00

« first day (2424 days earlier)      last day (1404 days later) »