- From: Roderick Sheeter <rsheeter@google.com>
- Date: Wed, 9 Mar 2016 09:51:53 -0800
- To: "Levantovsky, Vladimir" <Vladimir.Levantovsky@monotype.com>
- Cc: WebFonts Working Group <public-webfonts-wg@w3.org>
- Message-ID: <CABscrrG4M0kB+rabcWG4V0-AzHiLgg=gHTWEvOOK6NTsMxDcSA@mail.gmail.com>
I like https://www.w3.org/Fonts/WG/wiki/TestPlan20-UserAgent#mustNotRejectIncorrectTotalSize: it should work when totalSfntSize is too big or too small. I might prefer to write https://www.w3.org/Fonts/WG/wiki/TestPlan20-UserAgent#mustNotRejectGlyfSizeMismatch the same way. Make one file where 'glyf' origLength is bigger than decompressed can possibly be and one where it is smaller. To achieve this we might create a woff2 file whose origLength is neither the largest nor the smallest UIntBase128 for some number of bytes then derive two variants from it, one setting origLength to the largest UIntBase128 with the same number of bytes and one that sets the smallest (makes life easier if you don't change all the offsets in the file). Cheers, Rod S. On Fri, Mar 4, 2016 at 11:33 AM, Levantovsky, Vladimir < Vladimir.Levantovsky@monotype.com> wrote: > Folks, > > I updated the spec [1] adding two additional UA conformance requirements > (see end of subclause 3.2 to *not* reject a font due to mismatch between > declared and decoded totalSfntSize and origLength of the 'glyf' table. I > also added two new test case descriptions [2,3] to accommodate them. I'd > like to ask you to review the requirements and the test cases (this one [3] > in particular) - I believe it should work just fine and the fact that an > optimizing transform will always convert '0-contour' glyphs to empty glyphs > gives us an nice "organic" way to test the mismatch between the origLength > and decoded length without having to doctor the font data manually. The > current test description only test for a condition when the 'glyf' > origLength is larger than reconstructed 'glyf' table, not sure if we need > to test both conditions (when origLength is smaller) since it's likely > _will require_ manual adjustment of the test woff2 font file. > > Let me know if you have any comments. > > Thank you, > Vlad > > [1] https://dev.w3.org/webfonts/WOFF2/spec/ > [2] https://www.w3.org/Fonts/WG/wiki/TestPlan20-UserAgent# > mustNotRejectIncorrectTotalSize > [3] https://www.w3.org/Fonts/WG/wiki/TestPlan20-UserAgent# > mustNotRejectGlyfSizeMismatch > > > -----Original Message----- > From: WebFonts Working Group Issue Tracker [mailto:sysbot+tracker@w3.org] > Sent: Wednesday, March 02, 2016 10:48 AM > To: public-webfonts-wg@w3.org > Subject: WOFF-ACTION-198: Add conformance clause for must not reject glyf > size and original size > > WOFF-ACTION-198: Add conformance clause for must not reject glyf size and > original size > > http://www.w3.org/Fonts/WG/track/actions/198 > > Assigned to: Vladimir Levantovsky > > > > > > > > >
Received on Wednesday, 9 March 2016 17:52:23 UTC