This has returned with the webtool. Queries against logging_userindex are relatively expensive timewise. I may need to write a caching function so at least we only have to experience them once. SQLQuery me!23:56, 23 December 2017 (UTC)[reply]
I have an ugly / hackish workaround for this. Screen-scraping "https://en.wikipedia.org/w/index.php?title=Special:Contributions/" . $username . "&offset=&&limit=1" for "This account is currently blocked" seems to work. These blocks do not show up in API:Blocks, Special:IPBlocklist, nor the Block log - so this was a truly last resort. Running a test case (72's had a few of these) now to confirm. SQLQuery me!01:23, 6 December 2017 (UTC)[reply]
This one shows as a never been blocked, but was again a typo. The actual account name was corrected by another editor. This time, I inadvertently left the "e" on the end of "username" in when I selected "usernam" (minus the trailing 'e') for replacing the username. No idea how to account for a typo like that :/ Leave it to a human to really mess things up.
This one is a special case; I made this request when that account was already blocked; I was asking for a change in the block settings. The bot immediately removed the request, so I then made this request with an intentionally wrong username (adding 'NNN' to the end) so the bot wouldn't immediately remove it. The change in block settings I requested was done [1].
This one was just a typo done by accidentally copy/pasting {{Vandal|username}} per the instructions and not changing "username". Maybe it could ignore "username" as an account name as that is obviously in error. This request immediately preceded the above request, and was done in error. The same text for a request appeared in the next AIV report I made, albeit with the proper username put in.
At the moment, the script does not check to see WHEN the block happened, just IF it did. At some point, I will probably refactor it to see if the user was blocked within 24 hours of the AIV report. SQLQuery me!22:40, 5 December 2017 (UTC)[reply]
If a user re-adds a report from another user (accidentally removed / etc), it does count against that user. For the moment this can't be fixed because doing so would not count edits made by renamed users. SQLQuery me!22:36, 5 December 2017 (UTC)[reply]
Per feedback on my talkpage - when used as a webapp, allowing user-configurable (maybe regex, maybe substr()-ish) fields for keywords would be the way to go SQLQuery me!23:37, 5 December 2017 (UTC)[reply]