Refactor shield display logic; changed rules for DMs

[foldleft]

fixes: element-hq/element-web#12484

[foldleft]

This should represent a truth-table for all the behaviours we want to check, given the number of users in a room, DM status, and self-trust status:

  shieldStatusForMembership self-trust behaviour
    ✓ 2 unverified: returns 'normal', self-trust = true, DM = true
    ✓ 2 unverified: returns 'normal', self-trust = true, DM = false (1ms)
    ✓ 2 unverified: returns 'normal', self-trust = false, DM = true
    ✓ 2 unverified: returns 'normal', self-trust = false, DM = false
    ✓ 2 verified: returns 'verified', self-trust = true, DM = true
    ✓ 2 verified: returns 'verified', self-trust = true, DM = false
    ✓ 2 verified: returns 'verified', self-trust = false, DM = true (1ms)
    ✓ 2 verified: returns 'warning', self-trust = false, DM = false
    ✓ 2 mixed: returns 'normal', self-trust = true, DM = true
    ✓ 2 mixed: returns 'normal', self-trust = true, DM = false
    ✓ 2 mixed: returns 'normal', self-trust = false, DM = true
    ✓ 2 mixed: returns 'warning', self-trust = false, DM = false (1ms)
    ✓ 0 others: returns 'verified', self-trust = true, DM = true
    ✓ 0 others: returns 'verified', self-trust = true, DM = false
    ✓ 0 others: returns 'warning', self-trust = false, DM = true
    ✓ 0 others: returns 'warning', self-trust = false, DM = false
    ✓ 1 verified: returns 'verified', self-trust = true, DM = true
    ✓ 1 verified: returns 'verified', self-trust = true, DM = false (1ms)
    ✓ 1 verified: returns 'verified', self-trust = false, DM = true
    ✓ 1 verified: returns 'verified', self-trust = false, DM = false
    ✓ 1 unverified: returns 'normal', self-trust = true, DM = true
    ✓ 1 unverified: returns 'normal', self-trust = true, DM = false (1ms)
    ✓ 1 unverified: returns 'normal', self-trust = false, DM = true
    ✓ 1 unverified: returns 'normal', self-trust = false, DM = false
  shieldStatusForMembership other-trust behaviour
    ✓ 1 verified/untrusted: returns 'warning', DM = true
    ✓ 1 verified/untrusted: returns 'warning', DM = false
    ✓ 2 verified/untrusted: returns 'warning', DM = true (1ms)
    ✓ 2 verified/untrusted: returns 'warning', DM = false
    ✓ 2 unverified/untrusted: returns 'normal', DM = true
    ✓ 2 unverified/untrusted: returns 'normal', DM = false

NOTE: we don't flip "verified" to "normal" if we don't trust ourselves

[jryans]

Thanks for working on this! It's very helpful to have the tests and truth table for this! 😁

✓ 1 verified: returns 'verified', self-trust = false, DM = false

Shouldn't this case be "warning"? The room has someone other than yourself who you have verified, you don't trust yourself, and it's not a DM.

NOTE: we don't flip "verified" to "normal" if we don't trust ourselves

Can you be more specific about which case(s) you mean here?

[foldleft]

Thanks for working on this! It's very helpful to have the tests and truth table for this! 😁

✓ 1 verified: returns 'verified', self-trust = false, DM = false

It's a 1:1 room so it's treated as a DM

NOTE: we don't flip "verified" to "normal" if we don't trust ourselves

Can you be more specific about which case(s) you mean here?

Yeah. If it's a DM or a 1:1 room

[jryans]

✓ 1 verified: returns 'verified', self-trust = false, DM = false

It's a 1:1 room so it's treated as a DM

Ah right, makes sense then!

NOTE: we don't flip "verified" to "normal" if we don't trust ourselves

Can you be more specific about which case(s) you mean here?

Yeah. If it's a DM or a 1:1 room

Okay, I believe this is implementing the logic as expected. For 1:1 / DM rooms, your view of the room trust is meant to "effectively be the same as" your trust of the other user in the room, so nothing about yourself should affect it.

It does mean that each side will see a different shield for the room in 1:1 / DM rooms, but that's known and by design. We'll have to wait for feedback to see if it ends up being confusing.

[jryans]

Looks great, thanks! 😁