Re: [ft-devel] longjmp issue 1

[Werner LEMBERG]

> Thus, the easiest solution would be keeping both of FT_INVALID_()
> and FT_INVALID() in ftvalid.h, and, updating FT_INVALID_() in
> otvcommn.h to new syntax.

OK.

> I attached a sample patch. I could reproduce SEGV by
> Roboto-Thin.ttf, and the binary with the sample patch does not
> crash.

Works for me, please apply!  (And please use a one-line sentence as
the first line to get a clean entry for gitk and friends.)

> Considering that the super classes in otvalid/gxvalid includes
> FT_Validator as a member, changing surrounding codes in
> otvalid/gxvalid to call original FT_INVALID() (defined in ftvalid.h)
> would be possible.

Sounds sensible.

Thanks for working on this.


    Werner