bug#54646: 29.0.50; set-fontset-font and font clipping issues

[Eli Zaretskii <eliz@gnu.org>]

> From: Visuwesh <visuweshm@gmail.com>
> Cc: Robert Pluim <rpluim@gmail.com>,  54646@debbugs.gnu.org
> Date: Fri, 01 Apr 2022 20:28:06 +0530
> 
> >   hb_font_t *hb_font
> >     = font->driver->begin_hb_font
> >     ? font->driver->begin_hb_font (font, &position_unit)
> >     : NULL;
> >
> > The value of position_unit then affects the values returned in the
> > Lisp glyph object used to display the grapheme cluster:
> >
> >       xoff = lround (pos[i].x_offset * position_unit);
> >       yoff = - lround (pos[i].y_offset * position_unit);
> >       wadjust = lround (pos[i].x_advance * position_unit);
> >       if (xoff || yoff || wadjust != metrics.width)
> > 	LGLYPH_SET_ADJUSTMENT (lglyph, CALLN (Fvector,
> > 					      make_fixnum (xoff),
> > 					      make_fixnum (yoff),
> > 					      make_fixnum (wadjust)));
> >
> > I'd be interested in what happens there in the "good" vs the "bad"
> > cases.
> >
> > If we pass the same information to HarfBuzz, and it returns different
> > results, then it's probably a problem in HarfBuzz.
> 
> I get the same value for position_unit just after begin_hb_font call and
> just after setting the value of wadjust, in the bad and the good case:
> 0.015625.  In case I was not clear, here's a patch that shows where I
> added the printf calls
> 
> diff --git a/src/hbfont.c b/src/hbfont.c
> index 2721a66120..887e0c0e86 100644
> --- a/src/hbfont.c
> +++ b/src/hbfont.c
> @@ -490,6 +490,7 @@ hbfont_shape (Lisp_Object lgstring, Lisp_Object direction)
>      : NULL;
>    if (!hb_font)
>      return make_fixnum (0);
> +  printf("position_unit begin_hb_font: %f\n", position_unit);
>  
>    hb_bool_t success = hb_shape_full (hb_font, hb_buffer, NULL, 0, NULL);
>    if (font->driver->end_hb_font)
> @@ -593,6 +594,7 @@ hbfont_shape (Lisp_Object lgstring, Lisp_Object direction)
>        xoff = lround (pos[i].x_offset * position_unit);
>        yoff = - lround (pos[i].y_offset * position_unit);
>        wadjust = lround (pos[i].x_advance * position_unit);
> +      printf("position_unit after lround: %f\n", position_unit);
>        if (xoff || yoff || wadjust != metrics.width)
>  	LGLYPH_SET_ADJUSTMENT (lglyph, CALLN (Fvector,
>  					      make_fixnum (xoff),
> 
> So I see "position_unit begin_hb_font: 0.0015625" and "position_unit
> after lround: 0.0015625" in the good and the bad case.

So we pass the same data to HarfBuzz and get back different results in
xoff, yoff, and wadjust?  IOW, the results of shaping are different in
the two cases, although the inputs are identical?  Can you print the
other values involved in the data that gets put into lglyph, and see
whether any of it is different between the two cases?

the lglyph data is shown in this excerpt from the code:

      LGLYPH_SET_CHAR (lglyph, chars[char_idx]);
      LGLYPH_SET_CODE (lglyph, info[i].codepoint);

      unsigned code = info[i].codepoint;
      font->driver->text_extents (font, &code, 1, &metrics);
      LGLYPH_SET_WIDTH (lglyph, metrics.width);
      LGLYPH_SET_LBEARING (lglyph, metrics.lbearing);
      LGLYPH_SET_RBEARING (lglyph, metrics.rbearing);
      LGLYPH_SET_ASCENT (lglyph, metrics.ascent);
      LGLYPH_SET_DESCENT (lglyph, metrics.descent);

      xoff = lround (pos[i].x_offset * position_unit);
      yoff = - lround (pos[i].y_offset * position_unit);
      wadjust = lround (pos[i].x_advance * position_unit);
      if (xoff || yoff || wadjust != metrics.width)
	LGLYPH_SET_ADJUSTMENT (lglyph, CALLN (Fvector,
					      make_fixnum (xoff),
					      make_fixnum (yoff),
					      make_fixnum (wadjust)));

WHat is different between the two cases in this data?  Does the call
to font->driver->text_extents produce different data in 'metrics',
perhaps?  Do the values in pos[i] structure differ?  Something else?