Re: fonts with combining characters

[Keith Packard <keithp@xxxxxxxx>]

> The -misc-fixed-* fonts are all of the CharCell category, while the
> fixed-width fonts of Adobe and B&H fall into the Monospaced category,
> probably, because the renderer didn't guarantee that there is no
> out-of-box ink.

Actually, most of them do have ink above the font ascent -- they don't have
hokey looking accented capitals (Aring in particular).

> So it seems that we have no choice for C and M fonts, but would it be
> OK to have zero-width combining characters for P fonts?

Given that C and M fonts will use non-zero width combining characters, it seems
somewhat irrelevant what width you assign for P fonts as applications will have
to handle both styles of fonts.  A very clever client may be able to eliminate
one extra text element in a PolyText call by merging the overstruck character
inline with the remaining text elements, which argues for making them use
negative LSB/RSB and zero width.

Let's also remember to figure out how this should "really" work so we can
generate usable metrics in an extended X rendering system.

-keith


-
Linux-UTF8:   i18n of Linux on all levels
Archive:      http://mail.nl.linux.org/lists/