The merge no longer refuses when NAS is unreachable. DB changes always
run; per-side file operations are best-effort. When the NAS rename can't
happen it's recorded in a new pending_nas_moves table, and nas:auto-sync
drains the queue (rename + best-effort parent prune) once NAS is back,
retaining attempts/last_error on failure. Also plugs missing isEnabled()
guards on renameNasPath and deleteFolder so the merge cleanup path can't
hang on smbclient timeouts.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
847020fd02