@dan1st No, the process was broken prior to Teams moving to the new domain. The issue appears to be in something which only one or two people have access to and hasn't been fixed as of this time. If someone wants to join the Team, it's possible to manually invite them to the Team, but they would need to communicate their email address to at least one of the admins. Probably the best way for that to happen is for them to add their email address to their MS account and then ping an MS admin.
@JeffSchaller That pattern looks like it's already caught by Potentially bad keyword in answer and Potentially bad keyword in body Append -force to the command word(s) if you really want to add the pattern you provided.
@Makyen Hmm, so... that whole process is completely broken now? I was going to submit a PR that changed the language, but it sounds like it should be removed altogether, with a suggestion to ping an admin here in CHQ?
@Makyen Oh, good; there weren't very many messages in here to choose from. :-)
@CodyGray Well, the last time I used a random message as the reply target in order to have a long message, someone pinged me to ask if there was some special reason I was replying to the message I chose. This way, if someone follows the replied-to link it should be self-explanatory. :;
There's likely to be a high amount of overlap between the populations of people who follow back reply chains and the people who understand what you're doing, but OK. :-)
@CodyGray It's been broken since at least October, but I believe it's more like a year. Overall, I believe the intent is that it get fixed. OTOH, it's reasonable to change the text on that page to tell people the automated process isn't currently working and explain what's needed at the moment in order to gain access.
Right, yeah. I remember following exactly that process, and that's how I'm still registered on the Team, so I thought it was just a matter of updating "Channels" to "Teams", which would be a relatively easy change for me or anyone else to make.
If it's a matter of adding an advisory notice that the process is currently broken and what the recommended workaround is, that's also a relatively simple change.
I've switched the "LINKS" message to use normal URLs instead of shortened links. I've also added tooltips to the links to provide slightly more explanation as to what each link goes to. I'm open to suggestions as to what the tooltips should say, if anyone has some ideas. Given that the limit is now the number of characters which chat will display in the sidebar, rather than the length of the chat message, so we have a few characters left, if there's other links people would find helpful.
6
Please ping me if you have any improvements you'd like to suggest/discuss.
[ SmokeDetector | MS ] Potentially bad ip for hostname in answer, potentially bad ns for domain in answer, potentially bad keyword in answer, potentially bad keyword in username (4): Running codeanywhere React Native Cli by qservices on stackoverflow.com
I discovered another issue arising from GitHub having a bad UI and/or me not knowing how to use it: I have a pre-existing fork of the Charcoal userscript repository, and I somehow managed to create a pull-request on my fork, requesting that changes from my fork be pulled into my fork. Why would anyone ever want to do that?
(Obviously, I'd meant to create a PR on the main repository, and I did this using GitHub's built-in prompting, but apparently it led me astray.)
What I really don't understand is why my fork of the Charcoal userscript repository is showing that my "master" branch is 14 commits ahead of Charcoal-SE:master. None of those commits are commits that I made myself, and those commits don't seem to even exist in the main repo. How did they get there?!
@CodyGray It's really no problem. Given the UI, it's something that would be fairly easy to missclick. It's strange that they let you create a fork in the original source account, but don't let you create either an additional fork in an account that already has a fork, or a fork of a fork under the same account as the fork you're forking.
@CodyGray I've done this on purpose, but mostly for testing. It's definitely useful when running a test instance of SD, which will create PRs for watch/blacklist commands when the user doesn't have blacklist manager privileges.
Although, I've seen projects which are forks of another project and accept separate PRs, so it is something that's actually used, and not just in testing.
@Makyen Yeah, I see. That explains how the mistake was made. I already had a fork in my own private namespace, so it wouldn't let me create a second one there, and that caused it to default to "Charcoal-SE" being the owner. I definitely didn't expect it would let me do that, but not let me create a second fork in my own namespace.
@Makyen That doesn't seem to require PRs to come from the same repository. They could still come from another repository (including but not limited to the main repository, of which it is itself a fork).
@CodyGray IIRC, it's a change in their UI that they allow a fork in the source account. I don't recall that being shown as an option before, and they've definitely been changing the UI for that dropdown.
@CodyGray Well, that depends on how the secondary project is handling things. It's quite possible that it uses PRs from branches in the secondary's repository asking to merge into the secondary's master branch. That's actually quite reasonable under some conditions. In fact, I normally do PRs from branches in the main repository, so that could easily be done the same way with a secondary.
I got into the habit of pushing the branch into the main repository and doing a PR from within the main repository, because CI testing was handled a bit differently for that than when the PR comes from a fork.
@CodyGray No really, no. I've also gotten the impression that the CI testing has been reworked so I think the issues I was working to avoid have been fixed for a while, but I haven't actually double-checked.
@CodyGray That looks like a merge went wrong, or slightly differently, in your repo than in the main one (usually this happens when using GitHub's automatic merge conflict resolution, which merges the master branch into the branch you're working on. Doing that can can result in some funky commits). If you don't have anything in that repository that you're wanting to keep, it may be easier to delete it and recreate it. Personally, given that I also have a fork of my fork on my local machine I'd probably open a Git Bash window and do [the following assumes that you have the origin set to your fork and also have the main repository set as a remove] git fetch --all --prune; git checkout deploy; git reset --hard git remote | grep -v origin/deploy; git push origin deploy; git checkout master; git reset --hard git remote | grep -v origin/master; git push origin master and then git push origin master --force; git push origin deploy --force
Yeah, deleting the whole thing and recreating it is what I was planning to do. I was just trying to figure out how this happened first, because if I don't understand it, I can't learn from it.
@CodyGray Actually, I remember this now. This is correct and it's my fault. I'm the one that did the forced push. I'd already spent several hours updating all of the dependencies, but some effectively useless dependabot PRs were merged between when I started working on it and when I was ready, which caused massive merge conflicts. I could have and should have resolved it in a different way.
@CodyGray For other projects, they are usually quite helpful. For the Userscripts repository, they are a negative, IMO. All of the dependencies are only used in CI testing or during development. None of the dependencies are used in the actual userscripts, so there's very little likelihood of a security vulnerability being able to be accessed by anything.
Sigh, and per usual, a quick little tweak to a script turned into a more-than-one-hour project, with excessive overhead and an unwanted diversion into the mechanics of pushes, PRs, forks, and other minutiae of Git/GitHub.
@CodyGray Yeah, something almost always seems to make it such that what should be a very simple issue takes quite a bit longer than it should. I'm sorry that my actions in that forced push caused you to expend quite a bit of time.
@Makyen Nah, that was the least of the problems, really. After what had previously transpired, I was just gun-shy when it came to deleting things and really trying to understand what had happened.
!!/report stackoverflow.com/a/75445599 "spam (answer is posted to a spam-seed question, asked by someone whose answers all promoted this exact same product)"
@CodyGray You can execute the !!/report command again in 11 seconds. To avoid one user sending lots of reports in a few commands and slowing SmokeDetector down due to rate-limiting, you have to wait 30 seconds after you've reported multiple posts in one go.
@CodyGray If SE releases SD from chat rate limiting, we'd be happy to remove that rate limit. :; I know I find that SD reporting rate limit frustrating at times, too.
@CodyGray That pattern looks like it's already caught by Bad keyword in answer and Bad keyword in body Append -force to the command word(s) if you really want to add the pattern you provided.
[ SmokeDetector | MS ] Link at beginning of body, potentially bad asn for hostname in body, potentially problematic ns configuration in body (43): ماشین تصادفی منصف by hosi on stackoverflow.com
> Bad keyword in body, bad keyword in title, bad keyword in username ---------- Bad keyword in body - Position 0-10: skin serum Bad keyword in body - Position 5-10: serum Bad keyword in title - Position 0-10: skin serum Bad keyword in title - Position 0-10: skin serum Bad keyword in title - Position 5-10: serum Bad keyword in username - Position 0-10: skin serum
@CodyGray A normalized version, {{'3012915009', '13012915009'}}, of that pattern is already on the number watchlist. You can use !!/bisect-number +1 301 291 5009 to determine which entries on the number lists match that pattern, which would tell you: Matched by +1 301 291 5009 on line 3459 of watched_numbers.txt: +1 301 291 5009 found verbatim; 3012915009 found normalized
@Mast Done. I'm not bothering with the SO one; there's only one post, it's a question, and it doesn't look like it's intended to promote the site anyway.
