@Mast I agree; we want to block that post. I don't think blacklisting the user is necessary though in this case, so I would probably have just gone with tp.
@Xnero I'd be OK with it, if it restricted what can be approved with the command to: changing, adding, or removing no more than N lines in the watch/blacklist files in total, where N is small, say 5 maybe 10, and only when A) GitHub indicates an automatic merge is possible without merge conflicts (this is different than what's currently done for SD's PRs) and after B) the full CI suite has passed on the PR prior to the merge.
The merge should, preferably, be done on GitHub, rather than locally (this is not a strict requirement). The code resulting from the merge should not be pulled or used by SD unless the full CI suite passes after the merge (this is different than current).
@Makyen My reason for restricting the number of lines changed is that I want people to definitely be looking at the changes prior to approving them. Beyond that complexity of change, it's usually something which should be looked at more closely, rather than tossed off with a quick chat command (and it makes it less likely for someone to approve the wrong PR, which does happen from time to time).
Perhaps require that more than one person approves when the changes are longer.
Part of the reason for wanting to restrict the length is also the main reason for wanting to see the full CI suite run prior to the code being used: git merge sometimes messes up in ways which don't produce an error from the command, but the actual code ends up being fubared. Such situations will, hopefully, be caught by the CI testing.
@Mast thanks. It's not plagiarism; I'm glad they've learned to cite the composer. Whether such a string of questions is desired or not should be discussed on Chess Meta. Personally, I like the puzzles, because they're not as complicated as most other chess problems and they have a kind of mathematical beauty about them.
@SmokeDetector While it might be reasonable to block this at the system level, as it could be read as spam, it's written much more like just a user that doesn't understand what's necessary for a question on Stack Exchange (i.e. that questions need to be self-contained). In addition, a moderator has commented that they don't feel it's spam. I've raised a flag suggesting closure and normal deletion, which allows the user to edit while deleted and to raise a flag if want evaluation for undeletion.
OTOH, I wouldn't have a problem with watching the domain to be sure that they don't repeat posting about their blog. If they do repeat a couple of times, then I'd be much more inclined to call it spam.
@Xnero I have not explored it from a technical standpoint. However, I would be surprised if it wasn't possible to do what I've described, with the possible exception of performing the merge on GitHub, rather than locally, as doing that will depend on what's available within the GitHub API. However, I would have been surprised if it wasn't something the GitHub API could do.
Is it fine to add links to MS reports in custom flags? I don't see any issues with it, but maybe I'm missing something. Also, I assume the answer would be the same for any site on the network.
