Re: [Devel] fixed pitch fonts: should isFixedPitch be honoured?

[Werner LEMBERG]

> I should have explained in a bit more detail. What happens is that,
> because the Chinese ideograms are about 1 em wide, and I guess
> that's the width of the widest character, all the Roman characters
> look ridiculously widely spaced l i k e t h i s. It *doesn't* happen
> on Microsoft Windows NT applications, so it looks like Microsoft's
> font rendering logic is ignoring the flag.
> 
> I am not sure whether the rasterizer needs to do anything with a
> fixed-pitch flag except maybe avoid hinting the characters in a way
> that might make them different widths. I am probably rather
> ignorant, but surely a fixed-width font would set all its glyphs to
> the same advance width anyway. If it doesn't then an application is
> free to ignore the returned widths and impose a standard advance
> width, but that is another matter.
> 
> What do you think?

Sorry for not responding earlier.  Do you know whether this problem is
TrueType specific?  Or can this affect other font formats also?  I can
imagine that Windows FON/FNT fonts exhibit a similar `feature' (i.e.,
non-CJK characters in DBCS fonts are exactly half-width).

What about TrueType CJK fonts on the Mac?

Marc wrote

  In general, for monospaced DBCS fonts, the Latin glyphs are spaced
  as half-width glyphs.

Do you know how the algorithm in Windows work?  Something like

  . Check the Unicode range bits.

  . If the CJK bit(s) are set, and the isFixedPitch flag is set in the
    font, handle all glyphs not in the CJK range as half-width glyphs.


    Werner