Re: [ft-devel] Starting work on the OpenType code

[Werner LEMBERG]

> Now I'm starting to work for AAT/TrueTypeGX validator, in
> conjunction with Masatake Yamato.

Excellent!

> Taking a glance on OT validator, its error status is only
> FT_INVALID_FORMAT (when version is undefined) or FT_INVALID_DATA
> (everything except of FT_INVALID_FORMAT).

It also uses FT_INVALID_TOO_SHORT and FT_INVALID_OFFSET (see
otvcommn.h).

> I think the data structure of OpenType tables are designed to be
> simple and these 2 error status is enough.

In ftvalid.h, more errors are defined.

> In the case of GX extensions, the tables are much complexed, and
> validation result can be more varied.  [...]

Feel free to add more errors in case the stuff in fterrdef.h isn't
sufficient!

> The validation result should be defined from the viewpoint of font
> format specification?  Or, the validator should consider the fragile
> font driver?

You can control this with FT_ValidationLevel.  The default should
protect the font driver being quite tolerant otherwise.  The highest
level, FT_VALIDATE_PARANOID should make the driver follow the font
specification as close as possible.


    Werner