Block until the registry mirror agrees with a width reconfigure that has ALREADY happened
in the daemon: every present port name resolves and every retired one no longer does.
Why the retired half is the load-bearing one. The native reconfigure adds the new
leg(s) BEFORE removing the old port (the pw_map id-reuse fix), and the daemon delivers its
registry events in order, so the moment the OLD name disappears from the mirror, every
earlier event — including both new ports' arrival AND the previous flip's removals — has
been applied. Waiting on presence alone is NOT enough: after a stereo→mono→stereo cycle the
mirror can still hold the PREVIOUS input_<n>_L/input_<n>_R objects, whose names resolve to
long-dead port ids.
Returns true when the mirror settled, false on timeout (the caller logs and proceeds —
the strip's ports are real either way, so the head must still be repointed onto them).
Block until the registry mirror agrees with a width reconfigure that has ALREADY happened in the daemon: every
presentport name resolves and everyretiredone no longer does.Why the
retiredhalf is the load-bearing one. The native reconfigure adds the new leg(s) BEFORE removing the old port (the pw_map id-reuse fix), and the daemon delivers its registry events in order, so the moment the OLD name disappears from the mirror, every earlier event — including both new ports' arrival AND the previous flip's removals — has been applied. Waiting on presence alone is NOT enough: after a stereo→mono→stereo cycle the mirror can still hold the PREVIOUSinput_<n>_L/input_<n>_Robjects, whose names resolve to long-dead port ids.Returns
truewhen the mirror settled,falseon timeout (the caller logs and proceeds — the strip's ports are real either way, so the head must still be repointed onto them).