Re: [Freetype] Y advance when a glyp is rotated.

[David Turner]

Hello Olivier,

Olivier Couet wrote:
>  Hi All,
>
>  I am rotating a glyph with something like:
>
>    FT_Glyph aglyph;
>    FT_GlyphSlot slot = face->glyph;
>    FT_Get_Glyph ( slot, &aglyph );
>    FT_Glyph_Transform( aglyph, RotMatrix, NULL );
>    FT_Glyph_To_Bitmap( &aglyph, ft_render_mode_normal, 0, 1);
>
>  It seems not to bad, each glyph is rotated. Now to draw the characters
> one after each other I need the x and y advance. I have no problem for the
> x advance I get it via, int(slot->metrics.horiAdvance>>6) or 
> int(slot->advance.x>>6) both seems equivalent. 
>
>  But I have problems to get a proper y advance. I thought that 
> int(slot->advance.y>>6) is what I need but it is always equal to 0 ...
>
>  I noticed that there is also "advance" values in: aglyph->advance.x and
> aglyph->advance.y. The rotation seems to have some action on these 2
> values. If I have no rotation the y value is zero, and it changes when I
> rotate. But the values in aglyph->advance are really weird. For instance
> on the x axis the aglyph->advance.x is very different from
> int(slot->advance.x >> 6 ) ... I also tried to modify aglyph->advance.x
> with the >> operator but I had no success until now.
>
> Some body have an idea about that ?

When you create a glyph with FT_Get_Glyph, its advance vector
is copied in glyph->advance as a 16.16 fixed-point vector. Each
subsequent call to FT_Glyph_Transform also transforms the advance
vector; you thus should be able to use it.

The reason why 16.16 coordinates are used is to keep a reasonable
precision even in the case of transforms like rotations, indeed.


Regards,

- David Turner
- The FreeType Project  (www.freetype.org)