Comment # 25 on bug 967538 from Wolfgang Bauer

(In reply to Stakanov Schufter from comment #24)
> Sorry if I am asking but, could I give you feedback on Monday?

If you prefer to, no problem.

But of course the sooner I submit an update, the sooner this will be fixed for
other users... ;-)

> Or otherwise
> said: are there any "quirks" to be expected?

Not really.

I am not completely sure that everything will continue to behave correctly in
case the patch is triggered, but without it there will be a crash anyway at
that point.

That said, I haven't noticed anything strange here upto now, neither on my 13.2
systems nor on a Leap 42.1 VM. And it fixes the crashes reliably here.

I'll try to explain a bit:
The crash happens in the JavaScript interpreter when a new String object is
allocated but a non-string is passed as value. This is not allowed and should
not happen, but it apparently does in some circumstances.
My patch checks whether the passed value is really a string, and just returns
if not, thus preventing the crash.

So it doesn't fix the actual problem, it only prevents the resulting crash.
But I don't see much point in debugging QMLs JavaScript interpreter code to
find the real bug, especially considering that it seems to be fixed in Qt 5.6
anyway (at least I cannot reproduce the crashes with 5.6).