SheenFigure

Updates

• Added support for font variations
• Added support for device tables
• Added support for default script
• Fixed context lookup application issues
• Simplified internal structure and file names

New API:

• Variable Font: SFFontRef SFFontCreateWithVariationCoordinates(SFFontRef font, void *object, const SFInt16 *coordArray, SFUInteger coordCount)
• Point Size: void SFArtistSetPPEMSizes(SFArtistRef artist, SFUInt16 ppemWidth, SFUInt16 ppemHeight)

• Added support for applying custom features and values
• Added support for applying features simultaneously
• Fixed glyph map generation issue for backward text mode
• Fixed invalid ignore mask issue for mark to ligature positioning
• Simplified feature record representation in shaping engines
• Wrote builders for script list, feature list, and lookup list tables in the test project
• Made test project compatible with visual c++

• Added support for Reverse Chaining Contextual Substitution
• Used SheenBidi for determining General Category of a code point
• Removed all existing code and data related to general category lookup
• Fixed a bug imposing restrictions on context lookups to match previous/next glyphs in input range only
• Fixed a false assertion on space glyph replacement for zero width code points

• Fixed GSUB and GPOS features collision in pattern building process
• Fixed pair set matching issue in pair positioning implementation
• Ensured mark to base attachment on first base glyph of multiple substitution sequence
• Suppressed mark to mark attachment across different components of a ligature.
• Made Arabic engine explicitly set advance of all mark glyphs to zero
• Added incomplete support for zero width code points
• Added incomplete support for second and third anchor table formats
• Simplified open type table macros

• Added support for alternate substitution subtable
• Updated lookup files to Unicode 10.0
• Fixed a bug related to proper cluster formation
• Fixed a bug related to mark advance calculation

• Updated general category and joining type lookups to Unicode 9.0
• Reduced memory consumed by data lookups
• Fixed a memory leak in album class
• Added a simple engine for handling unrecognized scripts

• Added support for character level mirroring
• Minor bug fixes and improvements

• Added support for all 3 formats of context subtable
• Added support for all 3 formats of chain context subtable
• Fixed a critical bug in value record size calculation
• Fixed a critical bug in mark to mark positioning implementation
• Implemented strong test base for glyph substitution and positioning testing