Re: [ft-devel] Crash because of invalid use of setjmp

[Masatake YAMATO]

> Thank you for info. According to Yamato's post:
> 
> http://lists.nongnu.org/archive/html/freetype-devel/2005-07/msg00008.html
> 
> I think otvalid was originally designed to use ft_setjmp()
> and using ft_validator_run() was not essential. Possibly
> gxvalid is same.
> 
> I've searched the mailing list archive, but I could not find
> the discussion about why ft_validator_run() was introduced:
> possibly David Turner had written it. Anyway, in CVS, I could
> not find any evidence that this function was used before,
> always ft_setjmp() was used. So, now I think, replacing
> ft_validator_run() in otvalid/gxvalid by ft_setjmp() won't be
> problematic.
> 
> Yamato-san, how do you think of?

I agree with you. 

If we can replace a FT_BASE_DEF'ed function with a macro without ABI-breaking,
I think using Jens's macro:

   #define ft_validator_run( valid ) setjmp( (valid)->jump_buffer )

is the best.

I cannot find any situation where ft_validator_run function is useful...

Masatake YAMATO