Not sure exactly what is wrong with const signals, besides
QMetaObject::activate(...) wanting a non-const this argument.
Collected signals are from a recent clazy run, so changing these will at least
result in less noisy future clazy run ;)
And perhaps consistency in general signal non-constness.
The Clazy const-signal-or-slot check docs only claims
"For signals, it's just pointless to mark them as const."
without giving any reasoning.
So dumping patch as RFC, to be either moved to /dev/null or applied.