some utf8 chars are terrible small.

[Julian Bradfield]

On 2011-06-28, Aidan Kehoe <kehoea at parhasard.net> wrote:
> XEmacs uses a character’s Mule charset to assign a font to it, and it never
> inspects fonts for coverage under X11 (at the time the code was initially
> written, X11 fonts in general covered everything they advertised, and I
> didn’t change that back when I was involved in this code in the middle of
> the last decade, mainly because there wasn’t a sanctioned way to look at a
> server-side font’s coverage.) For the just-in-time Unicode character sets,
> there’s a list of Markus Kuhn’s misc-fixed fonts, which have good coverage,
> built in to XEmacs:

It's difficult. What I do with my own 21.4 fork is this, which works
pretty well for me most of the time.

`display-fonts-per-character' (buffer: *Hyper Apropos*, mode: Hyper-Apropos)


User variable:

  value: 1

  *This integer variable controls the extent to which XEmacs tries to find
  a valid font for a character. If 0, the traditional behaviour of selecting
  fonts according to character-set is used. If 1, then in addition for each
  character XEmacs looks for a font that actually contains that character,
  if the display supports this check. If 2, then in addition XEmacs will
  use legacy fonts when looking for fonts for Unicode characters (that is,
  when proceeding through a list of fonts, any font that matches a charset
  into which the given character can be converted, is considerd). If 3, then
  XEmacs will also try converting legacy characters into UCS characters (this
  may be useful on a system with good Unicode support, but no legacy support).