On Wed, Oct 7, 2015 at 1:18 AM, Werner LEMBERG <[email protected]> wrote:

>
> > I've inserted this file into the corpus, and it added 2 more control flow
> > edges (out of ~15000).
> > Not much, but not zero.
>
> The current function used for the fuzzer doesn't cover GX stuff at
> all.  We have to extend it step by step to get a better FreeType
> coverage.
>

Let's start from here.
I think the more sustainable approach is to commit the target function into
the FreeType git first.
I'll then re-target the bot to use that file instead of my own.
Once that works, one of the domain experts (i.e. not me :)  will be able to
add more stuff there.
This way you will fully control what functionality is actually being
fuzzed, and I will just monitor the infrastructure.
Does that sounds reasonable?

--kcc


>
> > The current target function handles any kind of font, so it may be
> > rather slow in finding problems specific to a single font type.  For
> > that we better write a separate target function. Would you like to
> > help?
>
> Yes, this would be great!  Have a look into the `ftmulti' FreeType
> demo program :-) If you want to test named instances only, a simpler
> approach of looping over all faces might work also with the recent
> 2.6.1 version (we have to add to test TTC support anyways to the
> fuzzing routine).
>
>
>     Werner
>
_______________________________________________
Freetype-devel mailing list
[email protected]
https://lists.nongnu.org/mailman/listinfo/freetype-devel

Reply via email to