Formatting, fixing descriptions.
changed the SFNT driver slightly to add more robust
Completely checked and fully formatted.
oops, forgot to commit these files
moved a lot of things from the TrueType driver to the SFNT
added a new demonstration program called "ftmulti" to
finalised the multiple masters support
a small change to allow the "type1" driver to detect
added support for multiple master fonts in "type1z". It is
important modifications to the Type1z driver
EXPORT_DEF renamed to FT_EXPORT_DEF +
minor fix to the Type1 driver(s) to apply the font matrix when
fixed some header files inclusions
some fixes for 64-bit systems. Mainly changed some
fixed the sbit loader (src/base/sfnt/ttsbit.c)
removed TT_MAX_SUBGLYPHS macro, there is no static
formatting - removed trailing spaces
removed obsolete header file
another rather large update
at last, fixed the bug, and re-enabled 5-gray levels support
removed obsolete files + update
created new header files, formatting and other minor
added the tag for the "DSIG" table
implemented FT_Select_Charmap and FT_Set_Charmap
small update to the memory tester
fixed a potential memory leak when loading embedded bitmaps
fixed stupid math lib bug in the Makefile
another massive changes in order to completely avoid
a new massive grunt work. Redefined the EXPORT_DEF,
additional changes, this time in order to pass extra parameters
another fix for the build system (the module list wasn't
I knew the first CVS commit wouldn't be good :o)
re-added the "freetype.mk" that was mistakenly deleted
logged directory re-organisation
major re-organisation of the FreeType 2 directory hierarchy
added experimental emboldening/outlining code. This
closed most of the memory leaks in the Type 1 driver(s)
changes the type of the "mode" parameter in FT_Set_Raster_Mode
finally a auto-hinter module interface I'm satisfied with,
fixed a compilation problem that occurred with the
an update to FT_Outline_Funcs used to improve
at last, the bug seems to be solved. Now, the source code
very preliminray CID support
still working on that damn rasterizer bug !! ;-)
"ftrast.c" now works for monochrome bitmaps, and doesn't
various updates to reduce the compilation warnings
added FT_CONFIG_OPTION_OLD_CALCS to ensure that
removed the smooth renderer from the demos directory
updates needed to compile the source programs now that
added support for version 2 of the OS/2 table
in order to ensure that the bytecode interpretation is exactly
added a simple line to set "glyph.control_len" when loading
changes to reflect the new files
first version of "t1tables", which will be used to access the
added the definition of the MaxProfile table
added the declaration of the smooth renderer to "include"
defined FT_Pointer as a typeless pointer
a few updates to the OpenType Layout code
moved the smooth renderer to the base layer directory
added FT_Sqrt64 to ensure that all bytecode operations are
removed an over-zealous assertation that halted the engine
added macros to read little endian words from a byte stream
raised TT_MAX_SUBGLYPHS to 96 to avoid problems
implemented extra interface for "FT_Get_Sfnt_Table"
added FT_Get_Sfnt_Table from "include/tttables.h"
added new auto-hinter driver interface !
added a new include file to access TrueType tables
added the FT_LOAD_FORCE_AUOHINT flag
fixed another small bug. The metrics were not computed
fixed a small bug that prevented the loading of certain
raised the value of TT_MAX_SUBGLYPHS, as it prevented
Fixed a compiler warning.
Removed compiler warnings.
Adding LDFLAGS. I don't know yet a good solution (within GNU make) to
Added pre-liminary autoconf support files for Unix. Note that these don't work
a small improvement to the Type 1 hinter, that
formatting, and remove the "interface" identifier
Consider the smooth renderer completed ! It will probably
simple fix required by the auto-hinting module
minor improvements, mainly to the PFB testing code...
fixed the display on 24-bit and 32-bit displays
major changes to the library:
support for FT_LOAD_NO_RECURSE needed by the auto-hinter
ftoutln.c is now mandatory. The optional parts of this
fixed a stupid bug in FT_MulFix