Page MenuHomePhabricator

Reconsider skin-specific menu label item behaviour
Open, HighPublicBUG REPORT

Description

Community members noticed the history label suddenly changed (T300052) as well as the label of the add section tab.

Confusingly, SkinTemplate uses wfMessageFallback to allow menu items to be overridden by skin message keys. However this doesn't apply to all menu items.

For example:

  • When creating the label action-viewsource for Vector, vector-action-viewsource message will be consulted first, then action-viewsource.

To add further confusion, in several instances the fallback is not intuitive. For example, the history link first consults vector-view-history and then history_short (not view-history)

Menu items that support this are:

  • edit and create links (view-edit-local, view-create-local, view-edit, view-create
  • $skname-view-foreign / "view-foreign"
  • Add section link action-addsection (new section vs + sign)
  • View source action-viewsource
  • view-history / history_short
  • action-delete / delete
  • action-move / move
  • action-undelete / undelete_short
  • viewdeleted / viewdeleted_short
  • action-protect
  • action-unprotect

TODO

I suggest going forward if this behaviour is necessary:

  • the key should always be predictable (e.g. vector-view-history should fallback to view-history )
  • this should apply to all menu items (for which T272624 is a blocker)
  • we should use update the default menus with old Vectors message keys where the messages in Vector are better e.g. "+" vs the more descriptive "new section"

Event Timeline

T301203 will address the problem on the short term with minimum disruption.

Addressing or not addressing the broader issue here can be deferred to a later date.

The API allows to view authors of deleted revisions (and page creations are logged on most wikis as well), so is the connection between separate "temporary accounts" (without disclosing the IP address itself), which could result from nuking pages from all such accounts simultaneously, an intentional consequence?