Re: Misc. comments

[Fred Kiefer]

I have been giving the font handling some thought, but wanted to wait
for Wims proposal before actually doing anything. We could do a lot of
small improvements, but would be better to have a correct concept. 
Anyway, what I came up with is that in the xgps backend the actual drawn
font should be determined by a combination of the selected font, its
matrix and the current transformation. In general this would lead to a
different X font to use. For this we would need an additional font cache
in the backend, where transformed fonts would be stored (Lets say up to
100 different fonts). This will lead to a possible mismatch between the
font metrics computed from the unscaled font and the metrics of the
really used font.
The question is (beside the one how to do this), if this would be
correct. That is, should text be drawn in the current transformation of
the focused view?