refactor(globals): shrink the context to what actually carries state

[HugoFara]

Stack 3 of 6. Base: #299. Globals goes from 505 lines to 353.

Most of the file's bulk did nothing. Four things come out:

• table() was the identity function — literally return $tableName;, left
  over from a configurable table prefix that no longer exists. It read as though
  it transformed the name, and HomeFacade carried a comment claiming it
  returned "properly prefixed" names. Its 185 call sites are now plain
  strings, and the literal concatenations that produced
  ('DELETE FROM ' . 'words') are folded into single literals.
• query() forwarded one line to QueryBuilder::table(). Callers now say so.
• The error-display flag was set to false by initialize() and never set
  true by anything; its single reader passed it straight to
  setErrorReporting(). That reader now passes false outright, which is what
  it always got. initialize() then had nothing left to do.
• The mysqli handle lived in two places — Connection::$instance and
  Globals::$dbConnection — with getInstance() resyncing from the second on
  every call and setInstance() writing both. Connection now owns it alone
  and the Globals accessors delegate.

Two judgement calls worth a second opinion

The Globals::get/setDbConnection() accessors stay. 291 call sites use them
as an "is the database up?" probe, mostly test skip guards. Rewriting all of
them would be churn for its own sake — the duplicated storage was the problem,
not the name.

Connection::reset() now genuinely clears the handle instead of leaving a
copy for the next getInstance() to recover. This is the one behaviour change
in the PR. testResetClearsInstance was asserting the opposite of its own name;
it now asserts the clear and restores the shared connection for the rest of the
run.

What is deliberately left alone

The user context (currentUserId, multiUserEnabled, isCurrentUserAdmin) is
load-bearing: QueryBuilder reads it at lines 248/287/292/323/328 to scope
every query, so a repository that looks unscoped is in fact filtered. That is
now stated on the class as a security boundary rather than a setting. Threading
a UserId through every repository call is a much larger change where a partial
job is a data-leak bug, so it is not attempted here.

Also drops 29 Globals imports left unused by the above.

Stack: #298 → #299 → 3 → #4 → #5 → #6