[Development] Fwd: [Qt bugreports] (QTBUG-46505) assertion failure (=> crash) in harfbuzz-shaper

Thiago Macieira thiago.macieira at intel.com
Sat Jun 6 13:25:10 CEST 2015


On Saturday 06 June 2015 10:20:56 René J.V. Bertin wrote:
> On Saturday June 06 2015 00:36:14 Thiago Macieira wrote:
> 
> Yes, I noticed that Qt4 doesn't use the system harfbuzz for whatever reason.
> >> As asked in the bug report below: is it necessary to handle this
> >> situation
> >> with an assert/abort, instead of simply returning (with whatever error
> >> the
> >> void function in question can return)?
> >
> >The assert is inside HB, so it's not our choice.
> 
> Isn't that a bit easy if you're shipping an old version that is no longer
> maintained? Seems the least one could do in this particular case is to
> ensure that in release mode the stuff is built with -DNDEBUG to deactivate
> the assert() side-effects (and "take it like a man" when that leads to
> other crashes and bug reports :) )

If you're hitting an assert, disabling the assert will make the problem worse. 
It means code will assume that the assert succeeded and proceed to make 
decisions on it.

Example:

	assert(ptr);
	ptr->i = 0;

Anyway, if this code is in the old hardbuzz-non-NG, what is preventing us from 
switching to NG?

-- 
Thiago Macieira - thiago.macieira (AT) intel.com
  Software Architect - Intel Open Source Technology Center




More information about the Development mailing list