to control memory debugging with FreeType. See the description of
"FT2_DEBUG_MEMORY", "FT2_ALLOC_TOTAL_MAX" and "FT2_ALLOC_COUNT_MAX"
in DEBUG.TXT
* src/cache/ftccache.c, src/cache/ftccmap.c, src/cache/ftcsbits.c,
ftlru.c: fixed the cache sub-system to correctly deal with out-of-memory
conditions.
* src/pfr/pfrobjs.c, src/pfr/pfrsbits.c: fixing compiler warnings and a
small memory leak
* src/psaux/psobjs.c (t1_reallocate_table): fixed a bug (memory leak) that
only happened when trying to resize an array would end in an OOM.
* src/smooth/ftgrays.c: removed compiler warnings / volatile bug
* src/truetype/ttobjs.c: removed segmentation fault that happened in
tight memory environments.
* include/freetype/config/ftoption.h, src/autohint/ahglobal.h,
src/autohint/ahglobal.c, src/autohint/ahglyph.c,
src/autohint/ahtypes.h:
included David Chester's patches to the auto-hinter in order to
slightly improve the output. Note that everything is controlled
through the new FT_CONFIG_OPTION_CHESTER_HINTS defined at the
end of "ftoption.h", there are also individual FT_CONFIG_CHESTER_XXX
macros to control individual "features".
Note that all improvements are enabled by default, but can be
tweaked for optimization and testing purpose. The configuration
macros will most likely disappear in the short future.
2003-01-11 David Turner <david@freetype.org>
* include/freetype/internal/fnttypes.h: fixed a structure field
definition to avoid memory overwrites
debug mode of template instantiation
* src/cff/cffparse.c: fixed the CFF table loader. It didn't accept
empty arrays, and this prevented the loading of certain fonts.
* src/pshinter/pshalgo2.c, src/pshinter/pshalgo3.c: adding fix to
prevent seg fault when hints are provided in an empty glyph !!
* include/freetype/ftimage.h: removed incorrect "zft_" definitions
and updated constants documentation comments
* include/freetype/freetype.h (FT_FaceRec): updating documentation
comment. The "descender" value is always *negative*, not positive !
want to list them all here. The operations performed were all logical
transformations of the sources:
- trying to convert all enums and constants to CAPITALIZED_STYLE, with
#define definitions like
#define my_old_constants MY_NEW_CONSTANT
- big, big update of the documentation comments
* include/freetype/freetype.h, src/base/ftobjs.c, src/smooth/ftsmooth.c,
include/freetype/ftimage.h: adding support for LCD-optimized rendering
though the new constants/enums:
FT_RENDER_MODE_LCD, FT_RENDER_MODE_LCD_V
FT_PIXEL_MODE_LCD, FT_PIXEL_MODE_LCD_V
this is still work in progress, don't expect everything to work correctly
though most of the features have been implemented.
* adding new FT_LOAD_XXX flags, used to specify both hinting and rendering
targets:
FT_LOAD_TARGET_NORMAL :: anti-aliased hinting & rendering
FT_LOAD_TARGET_MONO :: monochrome bitmaps
FT_LOAD_TARGET_LCD :: horizontal RGB/BGR decimated hinting & rendering
FT_LOAD_TARGET_LCD_V :: vertical RGB/BGR decimated hinting & rendering
note that FT_LOAD_TARGET_NORMAL is 0, which means that the default
behaviour of the font engine is _unchanged_.
preferences. The header freetype.h is not now affected, the interface is
specified via an FT_Parameter, the pointer to the interface is hidden in an
internal part of the face record, and all the definitions are in ftincrem.h.
development forum: "If FT_CONFIG_OPTION_INCREMENTAL is undefined
(this is the default), the TrueType loader crashes in line 852 of
src/truetype/ttgload.c when it tries to access face->glyph_locations.".
changed the names of memory macros. Examples:
MEM_Set => FT_MEM_SET
MEM_Copy => FT_MEM_COPY
MEM_Move => FT_MEM_MOVE
ALLOC => FT_ALLOC
FREE => FT_FREE
REALLOC = >FT_REALLOC
FT_NEW was introduced to allocate a new object from a _typed_
pointer..
note that ALLOC_ARRAY and REALLOC_ARRAY have been replaced
by FT_NEW_ARRAY and FT_RENEW_ARRAY which take _typed_ pointer
arguments.
This results in _lots_ of sources being changed, but makes the
code more generic and less error-prone..
details.
* include/freetype/internal/cfferrs.h,
include/freetype/internal/tterrors.h,
include/freetype/internal/t1errors.h: Removed. Replaced with files
local to the module. All extra error codes have been moved to
`fterrors.h'.
* src/sfnt/ttpost.h: Move error codes to `fterrors.h'.
* src/autohint/aherrors.h, src/cache/ftcerror.h, src/cff/cfferrs.h,
src/cid/ciderrs.h, src/pcf/pcferror.h, src/psaux/psauxerr.h,
src/psnames/psnamerr.h, src/raster/rasterrs.h, src/sfnt/sferrors.h,
src/smooth/ftsmerrs.h, src/truetype/tterrors.h,
src/type1/t1errors.h, src/winfonts/fnterrs.h: New files defining the
error names for the module it belongs to.
* include/freetype/ftmoderr.h: New file, defining the module error
offsets. Its structure is similar to `fterrors.h'.
* include/freetype/fterrors.h (FT_NOERRORDEF): New macro.
(FT_ERRORDEF): Redefined to use module error offsets.
All internal error codes are now public; unused error codes have
been removed, some are new.
* include/freetype/config/ftheader.h (FT_MODULE_ERRORS_H): New
macro.
* include/freetype/config/ftoption.h
(FT_CONFIG_OPTION_USE_MODULE_ERRORS): New macro.
All other source files have been updated to use the new error codes;
some already existing (internal) error codes local to a module have
been renamed to give them the same name as in the base module.
All make files have been updated to include the local error files.
* src/cid/cidtokens.h: Replaced with...
* src/cid/cidtoken.h: This file for 8+3 consistency.
* src/raster/ftraster.c: Use macros for header file names.
* include/freetype/config/ftheader.h: removed obsolete macros like
FT_SOURCE_FILE, etc.. and added cache-specific macro definitions that
were previously defined in <freetype/ftcache.h>. Added comments to be
included in a new API Reference section.
* src/*/*: removed the use of FT_SOURCE_FILE, etc.. now, each component
needs to added its own directory to the include path at compile time.
Modified all "rules.mk" and "descrip.mms" accordingly..
include/freetype/internal/internal.h: fixed header inclusion macros
to use direct definitions. This is the only way to do these things
in a portable way :-( The rest of the code should follow shortly
though everything compiles now..
* builds/compiler/intelc.mk, builds/compiler/watcom.mk,
builds/win32/detect.mk: added support for the Intel C/C++ compiler,
as well as _preliminary_ (read: doesn't work !!) support for Watcom.
Also added a new setup target. Type "make setup list" for a list
of supported command-line compilers on Win32..
Removed. ANSI C doesn't (explicitly) allow macro expansion in
arguments using `##'.
(FT2_PUBLIC_FILE, FT2_CONFIG_FILE, FT2_INTERNAL_FILE): Use directory
names directly. Make them configurable. Use `##' to strip leading
and trailing spaces from arguments.
* builds/unix/ft2unix.h: Adapted.
* src/base/ftsystem.c (ft_alloc, ft_realloc, ft_free, ft_io_stream,
ft_close_stream): Use FT_CALLBACK_DEF.
* builds/unix/ftsystem.c: Use new header scheme.
(FT_Done_Memory): Use free() from FT_Memory structure.
* src/base/ftinit.c, src/base/ftmac.c: Header scheme fixes.
* include/freetype/config/ft2build.h (FT2_CONFIG_ROOT,
FT2_PUBLIC_FILE, FT2_CONFIG_FILE, FT2_INTERNAL_FILE,
FT_SOURCE_FILE): Use `##' operator to be really ANSI C compliant.
in the "builds" directory
- modified a few files because BCC32 didn't like certain comparisons
(UInt compared with a UShort, apparently the UShort is promoted to
an Int)
- updated the INSTALL file to include IDE compilation, prevent against
BSD Make, and clarify the use of "make setup" (instead of "make")
for build configuration
I've made a few modifications:
- there is no new field named "driver" in "FT_SizeRec"
- the new fields in "TT_SizeRec" are:
strike_index :: value 0xFFFF means "no sbit strike selected"
strike_metrics :: the FT_Size_Metrics structure corresponding to
the currently selected strike
- the code in "ttload.c" has been somewhat cleaned up too
thanks a lot, Y-U !!
- David
you should #define FT_STRUCTURE to the structure which will be filled.
Replaced FT_FIELD_REF with FT_FIELD_SIZE, FT_FIELD_SIZE_DELTA, and
FT_FIELD_OFFSET to make the code more readable; additionally, it should be
more portable because we no longer cast a pointer to an FT_UShort (which
e.g. fails with Sun's C++ compiler) but computes the difference between two
pointers which is guaranteed to work.
Fixing warnings (and C++ errors) while using Sun's latest cc and CC
incarnations. Most of them are related to variable shadowing.
Now unix.in uses absolute paths also to overcome problems with libtool.
To do that in a consistent way I had to modify the rules how to handle
$OBJ_DIR (and consequently $LIB_DIR): You should set $OBJ_DIR explicitly
on the command line (or environment).
Moving some internal structures and constants from freetype.h to ftobjs.h.
Finally removing FT_LOAD_ANTI_ALIAS.
Cleaning up all error codes. Only the used ones have survived :-)
Removed unused FT_MAX_GLYPH_FORMATS constant.
T2 error codes are now in the range 0x500-0x5FF (instead of `TrueDoc').
Some minor improvements of error return values.
Finally fixing error code values in ftraster and ftgrays to be compliant
with all other FT error codes.
C++ programs easier:
Renamed FT_WordXX to FT_UIntXX.
Changed a lot of void* to FT_Byte* if related to i/o streams -- FreeType
always accesses streams byte-wise, so this makes sense IMHO.
Added a lot of #ifdef __cplusplus to header files (and removed a few).
Other minor syntax fixes (mostly casts).
Replaced the variable `private' with `private_dict' -- `private' is
reserved in C++.
