Make "Use spoilers when mentioning targets of a rule" feature optional

[MichaelSasser]

Even though I understand that it makes it easier to anonymize or hide user information in screenshots, I would still argue that #559 is a hindrance for some moderators (or at least of us). If we take screenshots, even for internal use, we anonymize them ourselves. We don't need this to be a feature of the bot.

When I start my day on Matrix, the first thing I do is read what the "nightshift" did, which users were affected and why. The moderation room is part of that routine. I usually jump back and forth between rooms to get a better understanding of a situation. Or I just want to scroll back a couple of days and visually get an impression of previous events to confirm a pattern.
I basically want to read the moderation room like a log file. Wrapping mentioning targets in spoilers makes that task harder.

Another argument would be that the bot itself leaks that information under certain conditions anyway:

1. After the message which wraps the mention target in a spoiler, if sends the following message to the moderation room, which has not (yet?) been wrapped:

   some_room:domain.tld (shortcode: my_shotcode) updated with 1 change:
   
        ADDED user (m.ban): @some_user:domain.tld (some_reason)
   

2. Element renders m.policy.rule.* state events in the policy rooms as the following, which in many cases are publicly readable by design.

   @mjolnir:domain.tld created a rule banning users matching @some_user:domain.tld for some_reason
   

I think this feature could be useful to some and should probably be the default, but please make it configurable.

[turt2live]

The intent of the feature isn't for screenshot protection (indeed, you can self-censor as needed). It's to add a layer of safety to moderators - some, if not many given the context, user IDs and such can be abusive on their own.

We're not likely to maintain a configuration flag for this, sorry.

[MichaelSasser]

TBH, I wouldn't have expected this as the intent for the feature. That's actually a fair and important point. I've been moderating for quite a while now, for most of the time outside of Matrix. It's absolutely true that the amount of abusive content I have seen and dealt with is affecting me negatively. I'm guessing many moderators who have been doing this for a while have had similar experiences. Fortunately, Matrix has not been as bad (at least for me) as other places in that regard.

Out of all the things that influence me negatively, user IDs would be at the bottom of my list. Though I wouldn't rule out that I dulled down over the years and that others might feel differently about them. Therefore, having those kinds of features is a very positive step.

The thing that still bugs me about that it is, that it tries to solve an issue that it, in the end, can't fix. At least not the way we use the bot. The only policy rooms our instance of Mjölnir has subscribed to are the ones only it manages. That means we are the only ones adding rules to them using the bot.
For every ban, we have ever issued, the user ID appears in three places in the moderation room. The one where we issue the command, and twice in the bot's responses. Even if the bot consistently wrapped its own ones into spoilers, the one we wrote, that stands out (because user ID is rendered blue) would remain visible. That kind of defies the whole point.

Would it maybe be possible to make this a client feature instead, that renders all user IDs in the timeline as spoilers, where the bot initiates that behavior change by sending a dedicated state event to the room? I'm thinking a little bit about something like org.matrix.room.preview_urls, just for manipulating how user IDs are rendered in a room by default. (I would still like to be able to deactivate it for myself)