Hi all, First of all, thank you for the great project description, the links provided are quite helpful!
I read more about the project, and it got me quite interested in the WOFF2 format and its compression. I have a few observations: * The structure of WOFF2 is quite similar to WOFF fonts (which is already supported) except for the fact that the data for the font tables is compressed in a single data stream and that WOFF2 may contain a 'collection' of fonts, specified in the Collection directory. Also, WOFF2 uses Brotli compression. * The glyf, loca and hmtx tables may have additional transformations applied before they are compressed, so these tables must be reconstructed. * It seems that the Brotli decoder [1] can be (and should be) used to decompress parts of the file. It is fast, does not seem to have any external dependencies and the licensing (MIT) should not be an issue. A few questions: * Are we looking at an implementation similar to WOFF (synthesize an SFNT from the WOFF, and load the SFNT)? This will mean no API to read WOFF2 metadata. * Although I am comfortable with the API, I have not directly worked on the FreeType font processing sources. Is it practically possible to understand it enough to implement this project in the GSoC time frame? I started reading on this subject recently, so apologies for any obvious details I missed! [1] https://github.com/google/brotli/tree/master/c On Tue, Mar 19, 2019 at 8:37 AM Alexei Podtelezhnikov <apodt...@gmail.com> wrote: > > Hi all, > > I have just added another idea for a GSoC project: WOFF2 support. It should > be posted on https://www.freetype.org/gsoc.html shortly. > > Alexei > _______________________________________________ > Freetype-devel mailing list > Freetype-devel@nongnu.org > https://lists.nongnu.org/mailman/listinfo/freetype-devel -- Nikhil _______________________________________________ Freetype-devel mailing list Freetype-devel@nongnu.org https://lists.nongnu.org/mailman/listinfo/freetype-devel
