This is just a frontend change, after all. Adding UX

This is a feature that was invaluable on wikimedia's use of phabricator, however, proliferation of a lot of forms can be troublesome / difficult to manage. Ideally the forms system could use some improvements to make things more manageable.

Thanks again @20after4 for the original idea online on Wikimedia since October 2021

Somehow "Normal" in 2023 I think - at least in Maniphest. But probably not so important.

@valerio.bozzolan Thanks a lot for your help, debugging, deeper investigation! Resolving.

For the records, I had the same experience when deleting a repository via ./bin/remove destroy rESHP for Unknown edge constant "25" (not 26 as mentioned in the task title) in downstream https://phabricator.wikimedia.org/T119588

I could reproduce this locally:

1. As an admin, go to http://phorge.localhost/herald/create/
2. Select "Maniphest Tasks"
3. On http://phorge.localhost/herald/create/?adapter=HeraldManiphestTaskAdapter , select "Global Rule"
4. Under "Conditions", select any of and set the three conditions Description | contains | Internet Archive, Description | contains | archive.org, Description | contains | Wayback Machine.
5. Under "Action", select Add projects | someProject
6. Select "Save Rule"
7. Trigger the rule by creating a new task.
8. Wait, wait, and wait, then look at the Herald Transcript linked from the task.

Tagging also as Affects Wikimedia since during Wikimedia Hackaton I've talked with a couple of very friendly staff members from WMF-CiviCRM that would really love this thing.

Thanks again, I really loved this follow-up on your hack

By coincidence, chatting with valerio.bozzolan yesterday made me realize that this rendering rule is merely in place for figlet and cowsay.

Historical commits with some comments about that regex: https://secure.phabricator.com/D7338, https://secure.phabricator.com/D7012

This is a good example of why fixing each case individually might be a bad idea, this fix was from D25202. It appears the old strlen allowed for integers too...

@valerio.bozzolan Here's a small sampling of all the errors that I encountered (I fixed each of them and got to the next one): https://we.phorge.it/P10

@valerio.bozzolan You know, I went on a spree fixing this in my local checkout. But it's super spread out all throughout the repo. The current master is broken even when I just view a revision. Then there are issues that show up in the SSH auth code when trying to push where the stack trace gets hidden (and it'll say something about ENORMOUS CHANGES, but that's a lie)

Thanks @arnold but please also share your error logs with us, otherwise we cannot fix :)

Re: all the null errors.

(Sorry for the ping poing, re-adding Remarkup hoping to be useful since this seems an issue in Remarkup itself, but last word to avivey)

This is unrelated to PHP8. This might turn out to be a duplicate/variation of T15372.

This is not triggered for {{{a}}} a or with a a {{{a}}} a, only for 1 string in front of a string with three squares.

Premising that very probably a fix there would not 100% resolve that upstream issue but surely this deserves a fix :D