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

[David Turner]

Hi Graham,

address@hidden a écrit :
> 
> FreeType 2 uses the isFixedPitch flag in the PostScript ('post') table in
> TrueType fonts to decide whether to replace the advance width of a
> character by the maximum advance width of any character in the font. I'm
> not sure this is the right thing to do. It gives a very bad appearance to
> (among others) the Chinese font MSSONG.TTF.
>
What do you mean ?

  - font appears monospaces when it should be proportional ?
  - it appears proportional when it should be monospaced ? 

> The TrueType documentation says that the 'post' table contains 'additional
> information needed to use TrueType fonts on PostScript printers', which
> implies to me that the information is not intended for general use. I note
> also that an ad-hoc flag, FT_LOAD_IGNORE_GLOBAL_ADVANCE_WIDTH, was added to
> suppress this behaviour, according to the in-source comment because of
> badly behaved Dynalab fonts; but I suspect that the behaviour may be wrong
> for any font.
>
> I'd be very grateful for comments on this question.
> 
It's not that we could ignore the isFixedPitch flag from the "post" table,
_if_ we have another test that is sufficiently reliable to replace this..

As far as I know, there is no reliable way to quickly tell that a font
is monospaced in the OpenType specification. And I don't think looking
at the PANOSE numbers will do it, unless I'm very mistaken..

Cheers,

- David