Need to bypass UTF-8 based file names

[Ferdinand Ramirez <ramirez ferdinand yahoo com>]

I know that glib is tightly tied to UTF-8 but I want to write parallel functions in nautilus to bypass UTF-8 
and get byte streams for use in the Desktop icons. This means the name of the file which contains information 
about the icon as well as the contents of the file such as the executable path and icon path need to be 
accessible as byte streams.

I need to be able to read these as byte streams and handle the conversion myself and also be able to write 
them back to these files. Right now, any attempt to create strings in the format I need it to be in gets 
rejected due to the UTF-8 requirement.

Note that I am not looking for a solution to map it to a known charset by setting some environment variables. 
I am looking for a solution that will help me write code to replace the functions that call the UTF-8 based 
functions.

What functions do I need to replace to achieve this? Where is the code for initializing the icons? Is it in 
libnautilus-private? If so, which files contain the functions I would need to replace and which functions are 
these?

Any help would be appreciated.

Thanks!

-Ferdinand