[cairo] Custom UserFont sizing questions [RESOLVED]

[Behdad Esfahbod]

Ian Britten wrote:
> Behdad Esfahbod wrote:
> 
>  > What closing?  There is no closing automatically added.
> 
> Well, _someone_ is adding a closing stroke to all my open glyph
> paths, and it certainly isn't me...

It's FreeType then.  So yeah, you should do the same kind of processing on the
cairo_glyph_path() output.

>> If you removed your duplicating hack,
> 
> I have - Sorry I didn't clarify that...
> 
>> it should be just a matter of:
>>
>>   cairo_glyph_path()
>>   cairo_stroke()
>>
>> I still am not sure why you insist that you need special processing :P.
> 
> Are you saying that's all I should need?  If so, then sorry but
> I'm confused what your path suggestions were for... ?

I thought your only problem left was caused by the duplicating (you said your
problem now is to do with the caps).

> Because, using just those two calls was what I originally started
> with, and is why I put in my own calls to FT_Outline_Decompose().

Ok, I suggest:

  cairo_glyph_path()
  cairo_copy_path()
  cairo_new_path()
  process path here and add back to cairo_t
  cairo_stroke()

All this in your UserFont render callback.

> Drawing my (open) glyphs with just those calls (None of the new path
> stuff you were talking about) results in the problematic closing
> strokes.

No I'm not saying.  But I didn't know that's happening.  Previously you said
if you fill them it happens, and that's expected because fill() does implicit
close_path.  But cairo_stroke doesn't.  If it's happening with cairo_stroke(),
then the closing is coming from FreeType, which actually I'm not surprised at
all, since FreeType is designed for contour fonts.


behdad

> Are you saying it shouldn't?  I had naively assumed that
> cairo_glyph_path() eventually called FT_Outline_Decompose()
> itself, since it's an FT font I'm rendering.  If Cairo does, then
> that's who is adding the extra point (And what I'm trying to bypass).
> OTOH, if Cairo is extracting the outlines from the FT_Face itself,
> then it might be in there somewhere...
> 
> [ Note I'm outputting to PDF -  Might there maybe be some sort of
> 'closepath' instruction being issues, as opposed to an extra
> data point?  Dunno (And I don't know how to read a PDF file) ]
> 
> Sorry if I'm being dense, but I suddenly feel like I'm going in
> a circle...
> Ian
> _______________________________________________
> cairo mailing list
> cairo at cairographics.org
> http://lists.cairographics.org/mailman/listinfo/cairo
>