forked from minhngoc25a/freetype2
added experimental emboldening/outlining code. This
is incomplete and will not compile so turned off
This commit is contained in:
parent
4d3e56392f
commit
f382328131
|
@ -498,3 +498,305 @@
|
|||
;
|
||||
}
|
||||
}
|
||||
|
||||
|
||||
/***************************************************************************/
|
||||
/***************************************************************************/
|
||||
/**** ****/
|
||||
/**** EXPERIMENTAL EMBOLDENING/OUTLINING SUPPORT ****/
|
||||
/**** ****/
|
||||
/***************************************************************************/
|
||||
/***************************************************************************/
|
||||
|
||||
#if 0
|
||||
|
||||
/* Compute the norm of a vector */
|
||||
#ifdef FT_CONFIG_OPTION_OLD_CALCS
|
||||
static
|
||||
FT_Pos ft_norm( FT_Vector* vec )
|
||||
{
|
||||
FT_Int64 t1, t2;
|
||||
|
||||
MUL_64( vec->x, vec->x, t1 );
|
||||
MUL_64( vec->y, vec->y, t2 );
|
||||
ADD_64( t1, t2, t1 );
|
||||
return (FT_Pos)SQRT_64(t1);
|
||||
}
|
||||
#else
|
||||
static
|
||||
FT_Pos ft_norm( FT_Vector* vec )
|
||||
{
|
||||
FT_F26Dot6 u, v, d;
|
||||
FT_Int shift;
|
||||
FT_ULong H, L, L2, hi, lo, med;
|
||||
|
||||
u = vec->x; if (u < 0) u = -u;
|
||||
v = vec->y; if (v < 0) v = -v;
|
||||
|
||||
if (u < v)
|
||||
{
|
||||
d = u;
|
||||
u = v;
|
||||
v = d;
|
||||
}
|
||||
|
||||
/* check that we're not trying to normalise zero !! */
|
||||
if (u==0) return 0;
|
||||
|
||||
/* compute (u*u+v*v) on 64 bits with two 32-bit registers [H:L] */
|
||||
hi = (FT_ULong)u >> 16;
|
||||
lo = (FT_ULong)u & 0xFFFF;
|
||||
med = hi*lo;
|
||||
|
||||
H = hi*hi + (med >> 15);
|
||||
med <<= 17;
|
||||
L = lo*lo + med;
|
||||
if (L < med) H++;
|
||||
|
||||
hi = (FT_ULong)v >> 16;
|
||||
lo = (FT_ULong)v & 0xFFFF;
|
||||
med = hi*lo;
|
||||
|
||||
H += hi*hi + (med >> 15);
|
||||
med <<= 17;
|
||||
L2 = lo*lo + med;
|
||||
if (L2 < med) H++;
|
||||
|
||||
L += L2;
|
||||
if (L < L2) H++;
|
||||
|
||||
/* if the value is smaller than 32-bits */
|
||||
shift = 0;
|
||||
if (H == 0)
|
||||
{
|
||||
while ((L & 0xC0000000) == 0)
|
||||
{
|
||||
L <<= 2;
|
||||
shift++;
|
||||
}
|
||||
return (FT_Sqrt32(L) >> shift);
|
||||
}
|
||||
else
|
||||
{
|
||||
while (H)
|
||||
{
|
||||
L = (L >> 2) | (H << 30);
|
||||
H >>= 2;
|
||||
shift++;
|
||||
}
|
||||
return (FT_Sqrt32(L) << shift);
|
||||
}
|
||||
}
|
||||
#endif
|
||||
|
||||
static
|
||||
int ft_test_extrema( FT_Outline* outline,
|
||||
int n )
|
||||
{
|
||||
FT_Vector *prev, *cur, *next;
|
||||
FT_Pos product;
|
||||
FT_Int first, last;
|
||||
|
||||
/* we need to compute the "previous" and "next" point */
|
||||
/* for this extrema.. */
|
||||
cur = outline->points + n;
|
||||
prev = cur - 1;
|
||||
next = cur + 1;
|
||||
|
||||
first = 0;
|
||||
for ( c = 0; c < outline->n_contours; c++ )
|
||||
{
|
||||
last = outline->contours[c];
|
||||
|
||||
if ( n == first )
|
||||
prev = outline->points + last;
|
||||
|
||||
if ( n == last )
|
||||
next = outline->points + first;
|
||||
|
||||
first = last + 1;
|
||||
}
|
||||
|
||||
product = FT_MulDiv( cur->x - prev->x, /* in.x */
|
||||
next->y - cur->y, /* out.y */
|
||||
0x40 ) -
|
||||
|
||||
FT_MulDiv( cur->y - prev->y, /* in.y */
|
||||
next->x - cur->x, /* out.x */
|
||||
0x40 );
|
||||
|
||||
if (product)
|
||||
product = ( product > 0 ? 1 : -1 );
|
||||
|
||||
return product;
|
||||
}
|
||||
|
||||
|
||||
/* Compute the orientation of path filling. It differs between TrueType */
|
||||
/* and Type1 formats. We could use the 'ft_outline_reverse_fill' flag, */
|
||||
/* but it's better to re-compute it directly (it seems that this flag */
|
||||
/* isn't correctly set for some weird composite glyphs for now).. */
|
||||
/* */
|
||||
/* We do this by computing bounding box points, and computing their */
|
||||
/* curvature.. the function returns either 1 or -1 */
|
||||
/* */
|
||||
static
|
||||
int ft_get_orientation( FT_Outline* outline )
|
||||
{
|
||||
FT_BBox box;
|
||||
FT_BBox indexes;
|
||||
int n, last;
|
||||
|
||||
indexes.xMin = -1;
|
||||
indexes.yMin = -1;
|
||||
indexes.xMax = -1;
|
||||
indexes.yMax = -1;
|
||||
|
||||
box.xMin = box.yMin = 32767;
|
||||
box.xMax = box.yMax = -32768;
|
||||
|
||||
/* is it empty ? */
|
||||
if ( outline->n_contours < 1 )
|
||||
return 1;
|
||||
|
||||
last = outline->contours[outline->n_contours-1];
|
||||
|
||||
for ( n = 0; n <= last; n++ )
|
||||
{
|
||||
FT_Pos x, y;
|
||||
|
||||
x = outline->points[n].x;
|
||||
if ( x < box.xMin )
|
||||
{
|
||||
box.xMin = x;
|
||||
indexes.xMin = n;
|
||||
}
|
||||
|
||||
if ( x > box.xMax )
|
||||
{
|
||||
box.xMax = x;
|
||||
indexes.xMax = n;
|
||||
}
|
||||
|
||||
y = outline->points[n].y;
|
||||
if ( y < box.yMin )
|
||||
{
|
||||
box.yMin = y;
|
||||
indexes.yMin = n;
|
||||
}
|
||||
|
||||
if ( y > box.yMax )
|
||||
{
|
||||
box.yMax = y;
|
||||
indexes.yMax = n;
|
||||
}
|
||||
}
|
||||
|
||||
/* test orientation of the xmin */
|
||||
return ft_test_extrema( outline, indexes.xMin ) ||
|
||||
ft_test_extrema( outline, indexes.yMin ) ||
|
||||
ft_test_extrema( outline, indexes.xMax ) ||
|
||||
ft_test_extrema( outline, indexes.yMax ) ||
|
||||
1; /* this is an empty glyph ?? */
|
||||
}
|
||||
|
||||
|
||||
static
|
||||
FT_Error ft_embolden( FT_Face original,
|
||||
FT_Outline* outline,
|
||||
FT_Pos* advance )
|
||||
{
|
||||
FT_Vector u, v;
|
||||
FT_Vector* points;
|
||||
FT_Vector cur, prev, next;
|
||||
FT_Pos distance;
|
||||
int c, n, first, orientation;
|
||||
|
||||
(void)advance;
|
||||
|
||||
/* compute control distance */
|
||||
distance = FT_MulFix( original->em_size/60,
|
||||
original->size->metrics.y_scale );
|
||||
|
||||
orientation = ft_get_orientation( &original->glyph->outline );
|
||||
|
||||
points = original->glyph->outline.points;
|
||||
|
||||
first = 0;
|
||||
for ( c = 0; c < outline->n_contours; c++ )
|
||||
{
|
||||
int last = outline->contours[c];
|
||||
|
||||
prev = points[last];
|
||||
|
||||
for ( n = first; n <= last; n++ )
|
||||
{
|
||||
FT_Pos norme, delta, d;
|
||||
FT_Vector in, out;
|
||||
|
||||
cur = points[n];
|
||||
if ( n < last ) next = points[n+1];
|
||||
else next = points[first];
|
||||
|
||||
/* compute the in and out vectors */
|
||||
in.x = cur.x - prev.x;
|
||||
in.y = cur.y - prev.y;
|
||||
|
||||
out.x = next.x - cur.x;
|
||||
out.y = next.y - cur.y;
|
||||
|
||||
/* compute U and V */
|
||||
norme = ft_norm( &in );
|
||||
u.x = orientation * FT_DivFix( in.y, norme );
|
||||
u.y = orientation * - FT_DivFix( in.x, norme );
|
||||
|
||||
norme = ft_norm( &out );
|
||||
v.x = orientation * FT_DivFix( out.y, norme );
|
||||
v.y = orientation * - FT_DivFix( out.x, norme );
|
||||
|
||||
d = distance;
|
||||
|
||||
if ( (outline->flags[n] & FT_Curve_Tag_On) == 0 )
|
||||
d *= 2;
|
||||
|
||||
/* Check discriminant for parallel vectors */
|
||||
delta = FT_MulFix( u.x, v.y ) - FT_MulFix( u.y, v.x );
|
||||
if ( delta > FT_BOLD_THRESHOLD || delta < - FT_BOLD_THRESHOLD )
|
||||
{
|
||||
/* Move point - compute A and B */
|
||||
FT_Pos x, y, A, B;
|
||||
|
||||
A = d + FT_MulFix( cur.x, u.x ) + FT_MulFix( cur.y, u.y );
|
||||
B = d + FT_MulFix( cur.x, v.x ) + FT_MulFix( cur.y, v.y );
|
||||
|
||||
x = FT_MulFix( A, v.y ) - FT_MulFix( B, u.y );
|
||||
y = FT_MulFix( B, u.x ) - FT_MulFix( A, v.x );
|
||||
|
||||
outline->points[n].x = distance + FT_DivFix( x, delta );
|
||||
outline->points[n].y = distance + FT_DivFix( y, delta );
|
||||
}
|
||||
else
|
||||
{
|
||||
/* Vectors are nearly parallel */
|
||||
FT_Pos x, y;
|
||||
|
||||
x = distance + cur.x + FT_MulFix( d, u.x + v.x )/2;
|
||||
y = distance + cur.y + FT_MulFix( d, u.y + v.y )/2;
|
||||
|
||||
outline->points[n].x = x;
|
||||
outline->points[n].y = y;
|
||||
}
|
||||
|
||||
prev = cur;
|
||||
}
|
||||
|
||||
first = last+1;
|
||||
}
|
||||
|
||||
if (advance)
|
||||
*advance = (*advance + distance*4) & -64;
|
||||
|
||||
return 0;
|
||||
}
|
||||
|
||||
#endif /* 0 - EXPERIMENTAL STUFF !! */
|
||||
|
|
Loading…
Reference in New Issue