[MPEG-OTSPEC] Minor ballot comment - change a field from 16 to 24 bits

[Behdad Esfahbod]

On Wed, Apr 2, 2025 at 7:46 AM mpsuzuki via mpeg-otspec <
mpeg-otspec at lists.aau.at> wrote:

> Dear Liam,
>
> I apologize to ask a few questions to your post in 2 months ago...
>
> As far as I could check the current specifications, the field DataOffset
> has been 16-bit for a long time, in Microsoft OpenType spec, Apple TrueType
> spec, and current 14496-22.
>
>
> https://learn.microsoft.com/ja-jp/typography/opentype/spec/otvarcommonformats
>
> https://developer.apple.com/fonts/TrueType-Reference-Manual/RM06/Chap6gvar.html
> (in Apple's spec, offsetToData in glyphVariationData structure might be the
> field we discuss)
>
> It is reasonable attitude to care about the overflow of this field in
> 24-bit
> GID world, but simple change of this field to 24-bit could be incompatible
> with existing implementations. At least, the proposed change in the draft
> ballot comment does not mention about a technique to keep the compatibility
> with existing fonts & font drivers. Maybe I'm missing something, please
> let me ask a few questions.
>
> 1) Adding a new version of "gvar" table whose DataOffset is 24-bit is bad
> option?
> Also, reservation of a bit in "flags" field of "gvar" header might be
> another
> option.
>

The proposal is to change the DataOffset to 24bit only in the newly
proposed `GVAR` table. We are not touching the existing `gvar`.

behdad


> 2) Some fonts with 24-bit DataOffset are already shipped to the market?
>
> 3) Some existing implementations are already capable to handle 24-bit
> DataOffset?
> HarfBuzz was already mentioned, but Behdad commented that it is not enabled
> by default.
>
> Regards,
> mpsuzuki
>
> On 2025/02/06 12:02, Liam R. E. Quin via mpeg-otspec wrote:
> > In 7.2.2.2 GlyphVariationDataHeader, there is a field called
> > DataOffset; this is defined to be 16-bit. In testing, people building
> > actual fonts found this overflowed and needed to be 24 bits,
> > so we have a proposed change, Change dataOffset from Offset16 to
> > Offset24.
> >
> > My understanding is that this has already been implemented in harfbuzz,
> > since its an ISO draft and since it was necessary...
> >
> > No other problems were found so far.
> >
> > The PDF file enclosed says the same thing, using the ISO template for
> > comments (numbered as if made sa a US comment; the Canadian mirror
> > committee meeting is tomorrow)
> >
> > liam
> >
>
> _______________________________________________
> mpeg-otspec mailing list
> mpeg-otspec at lists.aau.at
> https://lists.aau.at/mailman/listinfo/mpeg-otspec
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.aau.at/pipermail/mpeg-otspec/attachments/20250402/75a12469/attachment.htm>