This was a mistake in I53c0283, sorry. There are two users involved:
The "agent" is the user (typically an administrator) that made the
user group change, while the "viewing user" is the user for which the
user group was changed. When the "agent" moves the "viewing user"
into a new user group, the "viewing user" is the one that's now
a member of that group. They are what dictates the gender for the
"group-…-member" message.
Is this mistake bad enough for a backport?
Bug: T368249
Change-Id: I4916de2fb171873b625e51ee8823811e0296d323
This affects primarily the message
"notification-header-user-rights-add-only" which is very priminently
seen by every new editor that made their first few edits and gets
promoted to the next user group a few days later. Turns out the code
was just incomplete. All the information about the user and their
gender is already there, it was just not forwarded correctly.
Notice there are two messages:
* "group-…" messages don't have gender support. This is the (ideally)
gender neutral name of the group. Meant to be used as e.g. section
heading for a list of users.
* "group-…-member" is the same with gender support. To be used in all
contexts that are about a single, specific user with known gender.
Which is exactly what's happening here.
Turns out we can even use a neat convenience function from the
Language class that does exactly what we need.
I can't tell why but the array_values is apparently critical.
Originally added via I49b5fe5. It can't hurt so I keep it.
Bug: T368249
Change-Id: I53c028375d77c93f399538fd38aa8f8af30934b0
Changes to the use statements done automatically via script
Addition of missing use statements done manually
Change-Id: Iad87245bf8082193be72f7e482f29e9f1bad11fc
