[HarfBuzz] hb-shape to output applied features (optionally also an ultra-verbose XML log)?

[Behdad Esfahbod]

Hi Adam,

I fully agree with your assessment in principle.  However, I have to decide
between:

1) Exposing an stable API for such output, which will be inflexible and takes
time, of

2) Figure out a way to enable it at compile time with no stability guarantees,
which would not be as convenient to use.

As I said before, I'm open to the idea, let me think about it and see what I
can offer.

behdad

On 08/28/2012 08:13 PM, Adam Twardoch (List) wrote:
> On 12-08-29 01:20, Behdad Esfahbod wrote:
>> I agree. But that's not the primary goal of HarfBuzz, and distracts us
>> from the main goals. I'll see what I can do. 
> I believe with such an ultra-verbose output, complex-script font
> developers (even of proprietary ones) would be compelled to run the tool
> using the "ot" and "uniscribe" backend, and report inconsistencies --
> which would, after all, help HarfBuzz as well.
> 
> In a way, what I'm proposing is not at all contrary to the comprehensive
> testing effort. And since Uniscribe is probably as much a matching
> matching target as the Pango shapers are, this could be very useful.
> You'd have a chance to tap at developers of high-end complex script
> fonts such as, say, Tiro Typeworks. You give them a useful tool, they
> give you useful feedback which helps improve your product.
> 
> Mostly, their primary deployment target is Uniscribe, so they won't
> spend their time testing HarfBuzz *unless* you provide them with
> something that doesn't exist yet. Such a tool doesn't, and HarfBuzz is
> the perfect place to do such a job. As I mentioned earlier, I don't
> insist exactly on the XML syntax I proposed. It does pretty much
> exemplify in quite some detail what I think would be very useful,
> though. Oh, BTW, the whole "processing" part could be optional. For some
> scenarios, just the input and output, but with exact information, would
> be of great use.
> 
> I believe that a very good way to test implementations of layout engines
> is to collaborate with implementors of fonts (especially: complex fonts)
> which run on such layout engines.
> 
> There won't be many users of the functionality I'm proposing, but the
> ones who will use it will be of very high value for the development of
> HarfBuzz, I believe. I don't develop fonts myself, so I don't actually
> really have a stake in this. But I know from conversations with people
> who do develop complex fonts that this functionality is of great
> potential value for them. BTW, if some financing would be needed so you
> could find a developer who would do this, please let me know and I think
> I'll be able to gather something.
> 
> Best,
> Adam
>