From: Tomas Frydrych (firstname.lastname@example.org)
Date: Sun Apr 21 2002 - 12:44:43 EDT
> Joaquin Cuenca Abela wrote:
> Tomas, can you also clarify this point?
> I don't see why UTF-8 is unsuitable for the piecetable.
I want the piecetable and layout engine to be optimised for speed;
variable character width requires extra processing, that's why I do
not think it is a good encoding to use here.
> UTF-8 has another advantage over UTF-32. In the gtk & the qnx
> frontend, the format used to output text is UTF-8, so we'll not need
> to do any conversion to get text displayed. win32 uses UTF-16/UCS-2
> (in funcion of the registry?) so anyway you need to do a conversion
> from UTF-8 or UCS-4 to UCS-2.
We need to make a distinction between the GUI (the menus and
dialogues, etc) and the main editing space. We are going to
implement the main editing space using FreeType, which means
we will not use any calls to the native text drawing routines.
FreeType API uses unsigned long for Unicode characters. For the
GUI we will need to use what-ever the OS requires, and here I
agree that UTF-8 is the logical way to store the static strings.
This archive was generated by hypermail 2.1.4 : Sun Apr 21 2002 - 12:50:00 EDT