From: Tomas Frydrych (tomasfrydrych@yahoo.co.uk)
Date: Fri Jun 27 2003 - 13:27:03 EDT
When selection boundary splits the ligature, it is now drawn with
using the two original glyphs. This works pretty well when the width
of the ligature is not greatly different from the combined width of
the original characters (all the currently supported Latin ligatures
fall into this category).
It is less neat when the ligature is much narrower than the combined
widths, since it results in a certain overalp of the decomposed
characters with each other and the text on the sides. Furthermore,
where one of the decomposed characters is combining, it does not get
overlaid, but takes a place of its own, shifting the base charater
(and so making it overalp with its neigbour); the small bonus here is
that it is possible to tell that it is just the combining character
that is sellected. I am aware that this behaviour is nowhere near
perfect, but I think with a little getting used to it makes the
ligatures usable. This is essential for providing rudimentary support
for Arabic, which is really why I bothered with ligatures in the
first place. The Latin ligatures are merely a bonus, and should folk
not like them, they can be just commented out of the data tables.
files: fp_TextRun.cpp
Tomas
This archive was generated by hypermail 2.1.4 : Fri Jun 27 2003 - 13:41:15 EDT