The following is required:
Generate a script to take a ttf font and generate a bdf bitmap font.
The script must be capable of setting the rasterized font height and gray scale depth (either 1, 2, 4 or 8 bit gray scale depth)
The script must be capable of converting every code point present in the source file or only selected code points (supplied with a text file of Unicode code points)
The rasterisation MUST support the use of diacritical marks.
Open source font editors such as fontforge ([url removed, login to view]) or freetype ([url removed, login to view]) for example may be used for this step.
Script must be capable of running on a Linux platform. Standard Linux tools such as bash, perl, etc will be available.
Generate a rendering application
Using the intermediate bdf file, generate an example application. Take a Unicode string on the console and generate a graphic image with the rendered string.
The code must be written in c. Source files and makefile must be compatible with GCC.
The final binary must be statically compiled and not rely on any resources from the file system.
It must be capable of supporting diacritical marks, in particular the Khmer alphabet ([url removed, login to view])
It must not use excessive resources; the final code will be run on an embedded system. The target architecture is ARM7. No operating system is present, malloc is not available, stack usage should be limited as much as possible.
The example application should store the font data compressed using a lightweight compression algorithm (RLE perhaps). Benchmark tests should be provided to show the relative rendering time and the memory consumption for compressed and uncompressed data at different sizes (8, 16, 32 and 64 pixel heights)