You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
feat(video-conf): show who is in a call as faces, not as a number
"2 people in the call" answers a worse question than three avatars do. What
decides whether to walk into a call is *who* is already in it.
`CallParticipants` draws a face for each of the people it is given and turns
whatever is left over into a "+N" shaped and sized like one more avatar, so a row
reads as a group of people rather than as faces followed by a statistic. The whole
count stays as the group's label — for anyone who cannot see the avatars, and
because "+2" means nothing without a total.
Two surfaces, one component:
- the sidebar list, under a `Participants` label, from `participants` on the
joinable payload. The server caps that at three, since a call in a busy channel
would otherwise send a roster to draw three avatars.
- the preflight when joining, under a "Participants in the call" label and five at
a time — a screen has more room than a row. These come from the call window's own
copy of the members, so nothing extra travels for them.
`participants` returns to the joinable payload, and `CallParticipants` to the tree;
both were cut from the first release in 12ce0c2, which recorded the design for
exactly this.
The joinable fixture now carries participants, and the spec asserts the cap, that
the payload carries nothing beyond what a face needs, and that people who were
invited but never turned up get no face.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
0 commit comments