and empty lines in output.
* docs/docmaker.py: Improved script to generate table of contents
and index pages. It also supports wildcards on non Unix systems.
* include/freetype/*.h, include/freetype/cache/*.h: Updated comments
to include section definitions/delimitations for the API Reference
generator.
* include/freetype/freetype.h: Moved declaration of
`FT_Generic_Finalizer' and the `FT_Generic' structure to...
* include/freetype/fttypes.h: here.
* builds/os2/detekt.mk, builds/win32/detekt.mk: Moved comment to
avoid trailing spaces in variable.
* builds/freetype.mk: Use $(D) instead of $D to make statement more
readable.
* docs/docmaker.py: Formatting.
* src/psaux/psauxmod.c: Fixed a broken inclusion of component
header files (an FT_FLAT_COMPILE test was missing).
* src/cache/ftcmanag.c (FTC_Manager_Done): Fixed a bug that caused
an occasional crash when the function was called (due to a dangling
pointer).
* src/base/ftsystem.c (FT_Done_Memory): Fixed an obvious bug:
The ANSI "free()" function was called instead of "memory->free()".
* docs/docmaker.py: Added section filtering, multi-page generation
(index page generation is still missing though).
* builds/unix/install.mk, builds/unix/ft2unix.h: The file "ft2unix.h"
is now installed as <ft2build.h> for Unix systems. Note that we
still use the "freetype2/freetype" installation path for now.
* many files: Now using <ft2build.h> as the default build and setup
configuration file in all public headers. Internal source files
still need some changes though.
* builds/devel/ft2build.h, builds/devel/ftoption.h: Created a new
Win32 developer builds.
builds/win32/w32-dev.mk: Changed the developer build targets to
"devel-gcc" and "devel-bcc" in order to be able to develop with the
Borland C++ compiler.
the basic parsing routines seem to work ok
we now generate a list of DocBlock objects from
a list of input file, we now need to sort them
by "kind" (i.e. type/macro/functions) to generate
several web pages, as well as a global index
There is still some work on it to allow the following:
- multiple input files
- block classification according to block's first markers
(e.g. types, functions, constants, etc..)
- indexing and cross-linking
- better layout ;-)
David Turner