Synchronised line endinge with release branch

1 files changed, 441 insertions, 441 deletions

diff --git a/libXfont/src/fontfile/fontscale.c b/libXfont/src/fontfile/fontscale.c
index 524db98d0..eb8d0bb58 100644
--- a/libXfont/src/fontfile/fontscale.c
+++ b/libXfont/src/fontfile/fontscale.c
@@ -1,441 +1,441 @@
-/*

-

-Copyright 1991, 1998  The Open Group

-

-Permission to use, copy, modify, distribute, and sell this software and its

-documentation for any purpose is hereby granted without fee, provided that

-the above copyright notice appear in all copies and that both that

-copyright notice and this permission notice appear in supporting

-documentation.

-

-The above copyright notice and this permission notice shall be included in

-all copies or substantial portions of the Software.

-

-THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR

-IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,

-FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL THE

-OPEN GROUP BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN

-AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN

-CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

-

-Except as contained in this notice, the name of The Open Group shall not be

-used in advertising or otherwise to promote the sale, use or other dealings

-in this Software without prior written authorization from The Open Group.

-

-*/

-

-/*

- * Author:  Keith Packard, MIT X Consortium

- */

-

-#ifdef HAVE_CONFIG_H

-#include <config.h>

-#endif

-#include    <X11/fonts/fntfilst.h>

-#include <math.h>

-

-#ifdef _MSC_VER

-#define hypot _hypot

-#endif

-

-Bool

-FontFileAddScaledInstance (FontEntryPtr entry, FontScalablePtr vals, 

-			   FontPtr pFont, char *bitmapName)

-{

-    FontScalableEntryPtr    scalable;

-    FontScalableExtraPtr    extra;

-    FontScaledPtr	    new;

-    int			    newsize;

-

-    scalable = &entry->u.scalable;

-    extra = scalable->extra;

-    if (extra->numScaled == extra->sizeScaled)

-    {

-	newsize = extra->sizeScaled + 4;

-	new = realloc (extra->scaled, newsize * sizeof (FontScaledRec));

-	if (!new)

-	    return FALSE;

-	extra->sizeScaled = newsize;

-	extra->scaled = new;

-    }

-    new = &extra->scaled[extra->numScaled++];

-    new->vals = *vals;

-    new->pFont = pFont;

-    new->bitmap = (FontEntryPtr) bitmapName;

-    if (pFont)

-	pFont->fpePrivate = (pointer) entry;

-    return TRUE;

-}

-

-/* Must call this after the directory is sorted */

-

-void

-FontFileSwitchStringsToBitmapPointers (FontDirectoryPtr dir)

-{

-    int	    s;

-    int	    b;

-    int	    i;

-    FontEntryPtr	    scalable;

-    FontEntryPtr	    nonScalable;

-    FontScaledPtr	    scaled;

-    FontScalableExtraPtr    extra;

-    

-    scalable = dir->scalable.entries;

-    nonScalable = dir->nonScalable.entries;

-    for (s = 0; s < dir->scalable.used; s++)

-    {

-	extra = scalable[s].u.scalable.extra;

-	scaled = extra->scaled;

-	for (i = 0; i < extra->numScaled; i++)

-	    for (b = 0; b < dir->nonScalable.used; b++)

-		if (nonScalable[b].name.name == (char *) scaled[i].bitmap)

-		    scaled[i].bitmap = &nonScalable[b];

-    }

-}

-

-void

-FontFileRemoveScaledInstance (FontEntryPtr entry, FontPtr pFont)

-{

-    FontScalableEntryPtr    scalable;

-    FontScalableExtraPtr    extra;

-    int			    i;

-

-    scalable = &entry->u.scalable;

-    extra = scalable->extra;

-    for (i = 0; i < extra->numScaled; i++)

-    {

-	if (extra->scaled[i].pFont == pFont)

-	{

-	    if (extra->scaled[i].vals.ranges)

-		free (extra->scaled[i].vals.ranges);

-	    extra->numScaled--;

-	    for (; i < extra->numScaled; i++)

-		extra->scaled[i] = extra->scaled[i+1];

-	}

-    }

-}

-

-Bool

-FontFileCompleteXLFD (FontScalablePtr vals, FontScalablePtr def)

-{

-    FontResolutionPtr res;

-    int		num_res;

-    double	sx, sy, temp_matrix[4];

-    double	pixel_setsize_adjustment = 1.0;

-    /*

-     * If two of the three vertical scale values are specified, compute the

-     * third.  If all three are specified, make sure they are consistent

-     * (within a pixel)

-     *

-     * One purpose of this procedure is to complete XLFD names in a

-     * repeatable manner.  That is, if the user partially specifies

-     * a name (say, pixelsize but not pointsize), the results generated

-     * here result in a fully specified name that will result in the

-     * same font.

-     */

-

-    res = GetClientResolutions(&num_res);

-

-    if (!(vals->values_supplied & PIXELSIZE_MASK) ||

-	!(vals->values_supplied & POINTSIZE_MASK))

-    {

-	/* If resolution(s) unspecified and cannot be computed from

-	   pixelsize and pointsize, get appropriate defaults. */

-

-	if (num_res)

-	{

-	    if (vals->x <= 0)

-		vals->x = res->x_resolution;

-	    if (vals->y <= 0)

-		vals->y = res->y_resolution;

-	}

-

-	if (vals->x <= 0)

-	    vals->x = def->x;

-	if (vals->y <= 0)

-	    vals->y = def->y;

-    }

-    else

-    {

-	/* If needed, compute resolution values from the pixel and

-	   pointsize information we were given.  This problem is

-	   overdetermined (four equations, two unknowns), but we don't

-	   check for inconsistencies here.  If they exist, they will

-	   show up in later tests for the point and pixel sizes.  */

-

-	if (vals->y <= 0)

-	{

-	    double x = hypot(vals->pixel_matrix[1], vals->pixel_matrix[3]);

-	    double y = hypot(vals->point_matrix[1], vals->point_matrix[3]);

-	    if (y < EPS) return FALSE;

-	    vals->y = (int)(x * 72.27 / y + .5);

-	}

-	if (vals->x <= 0)

-	{

-	    /* If the pixelsize was given as an array, or as a scalar that

-	       has been normalized for the pixel shape, we have enough

-	       information to compute a separate horizontal resolution */

-

-	    if ((vals->values_supplied & PIXELSIZE_MASK) == PIXELSIZE_ARRAY ||

-	        (vals->values_supplied & PIXELSIZE_MASK) ==

-		    PIXELSIZE_SCALAR_NORMALIZED)

-	    {

-		double x = hypot(vals->pixel_matrix[0], vals->pixel_matrix[2]);

-		double y = hypot(vals->point_matrix[0], vals->point_matrix[2]);

-		if (y < EPS) return FALSE;

-		vals->x = (int)(x * 72.27 / y + .5);

-	    }

-	    else

-	    {

-		/* Not enough information in the pixelsize array.  Just

-		   assume the pixels are square. */

-		vals->x = vals->y;

-	    }

-	}

-    }

-

-    if (vals->x <= 0 || vals->y <= 0) return FALSE;

-

-    /* If neither pixelsize nor pointsize is defined, take the pointsize

-       from the defaults structure we've been passed. */

-    if (!(vals->values_supplied & PIXELSIZE_MASK) &&

-	!(vals->values_supplied & POINTSIZE_MASK))

-    {

-	if (num_res)

-	{

-	    vals->point_matrix[0] =

-	    vals->point_matrix[3] = (double)res->point_size / 10.0;

-	    vals->point_matrix[1] =

-	    vals->point_matrix[2] = 0;

-	    vals->values_supplied = (vals->values_supplied & ~POINTSIZE_MASK) |

-				    POINTSIZE_SCALAR;

-	}

-	else if (def->values_supplied & POINTSIZE_MASK)

-	{

-	    vals->point_matrix[0] = def->point_matrix[0];

-	    vals->point_matrix[1] = def->point_matrix[1];

-	    vals->point_matrix[2] = def->point_matrix[2];

-	    vals->point_matrix[3] = def->point_matrix[3];

-	    vals->values_supplied = (vals->values_supplied & ~POINTSIZE_MASK) |

-				    (def->values_supplied & POINTSIZE_MASK);

-	}

-	else return FALSE;

-    }

-

-    /* At this point, at least two of the three vertical scale values

-       should be specified.  Our job now is to compute the missing ones

-       and check for agreement between overspecified values */

-

-    /* If pixelsize was specified by a scalar, we need to fix the matrix

-       now that we know the resolutions.  */

-    if ((vals->values_supplied & PIXELSIZE_MASK) == PIXELSIZE_SCALAR)

-    {

-	/* pixel_setsize_adjustment used below to modify permissible

-	   error in pixel/pointsize matching, since multiplying a

-	   number rounded to integer changes the amount of the error

-	   caused by the rounding */

-

-	pixel_setsize_adjustment = (double)vals->x / (double)vals->y;

-	vals->pixel_matrix[0] *= pixel_setsize_adjustment;

-	vals->values_supplied  = (vals->values_supplied & ~PIXELSIZE_MASK) |

-				 PIXELSIZE_SCALAR_NORMALIZED;

-    }

-

-    sx = (double)vals->x / 72.27;

-    sy = (double)vals->y / 72.27;

-

-    /* If a pointsize was specified, make sure pixelsize is consistent

-       to within 1 pixel, then replace pixelsize with a consistent

-       floating-point value.  */

-

-    if (vals->values_supplied & POINTSIZE_MASK)

-    {

-    recompute_pixelsize: ;

-	temp_matrix[0] = vals->point_matrix[0] * sx;

-	temp_matrix[1] = vals->point_matrix[1] * sy;

-	temp_matrix[2] = vals->point_matrix[2] * sx;

-	temp_matrix[3] = vals->point_matrix[3] * sy;

-	if (vals->values_supplied & PIXELSIZE_MASK)

-	{

-	    if (fabs(vals->pixel_matrix[0] - temp_matrix[0]) >

-		    pixel_setsize_adjustment ||

-		fabs(vals->pixel_matrix[1] - temp_matrix[1]) > 1 ||

-		fabs(vals->pixel_matrix[2] - temp_matrix[2]) > 1 ||

-		fabs(vals->pixel_matrix[3] - temp_matrix[3]) > 1)

-		return FALSE;

-	}

-	if ((vals->values_supplied & PIXELSIZE_MASK) == PIXELSIZE_ARRAY &&

-	    (vals->values_supplied & POINTSIZE_MASK) == POINTSIZE_SCALAR)

-	{

-	    /* In the special case that pixelsize came as an array and

-	       pointsize as a scalar, recompute the pointsize matrix

-	       from the pixelsize matrix. */

-	    goto recompute_pointsize;

-	}

-

-	/* Refresh pixel matrix with precise values computed from

-	   pointsize and resolution.  */

-	vals->pixel_matrix[0] = temp_matrix[0];

-	vals->pixel_matrix[1] = temp_matrix[1];

-	vals->pixel_matrix[2] = temp_matrix[2];

-	vals->pixel_matrix[3] = temp_matrix[3];

-

-	/* Set values_supplied for pixel to match that for point */

-	vals->values_supplied =

-	    (vals->values_supplied & ~PIXELSIZE_MASK) |

-	    (((vals->values_supplied & POINTSIZE_MASK) == POINTSIZE_ARRAY) ?

-		PIXELSIZE_ARRAY : PIXELSIZE_SCALAR_NORMALIZED);

-    }

-    else

-    {

-	/* Pointsize unspecified...  compute from pixel size and

-	   resolutions */

-    recompute_pointsize: ;

-	if (fabs(sx) < EPS || fabs(sy) < EPS) return FALSE;

-	vals->point_matrix[0] = vals->pixel_matrix[0] / sx;

-	vals->point_matrix[1] = vals->pixel_matrix[1] / sy;

-	vals->point_matrix[2] = vals->pixel_matrix[2] / sx;

-	vals->point_matrix[3] = vals->pixel_matrix[3] / sy;

-

-	/* Set values_supplied for pixel to match that for point */

-	vals->values_supplied =

-	    (vals->values_supplied & ~POINTSIZE_MASK) |

-	    (((vals->values_supplied & PIXELSIZE_MASK) == PIXELSIZE_ARRAY) ?

-		POINTSIZE_ARRAY : POINTSIZE_SCALAR);

-

-	/* If we computed scalar pointsize from scalar pixelsize, round

-	   pointsize to decipoints and recompute pixelsize so we end up

-	   with a repeatable name */

-	if ((vals->values_supplied & POINTSIZE_MASK) == POINTSIZE_SCALAR)

-	{

-	    /* Off-diagonal elements should be zero since no matrix was

-	       specified. */

-	    vals->point_matrix[0] =

-		(double)(int)(vals->point_matrix[0] * 10.0 + .5) / 10.0;

-	    vals->point_matrix[3] =

-		(double)(int)(vals->point_matrix[3] * 10.0 + .5) / 10.0;

-	    goto recompute_pixelsize;

-	}

-    }

-

-    /* We've succeeded.  Round everything to a few decimal places

-       for repeatability. */

-

-    vals->pixel_matrix[0] = xlfd_round_double(vals->pixel_matrix[0]);

-    vals->pixel_matrix[1] = xlfd_round_double(vals->pixel_matrix[1]);

-    vals->pixel_matrix[2] = xlfd_round_double(vals->pixel_matrix[2]);

-    vals->pixel_matrix[3] = xlfd_round_double(vals->pixel_matrix[3]);

-    vals->point_matrix[0] = xlfd_round_double(vals->point_matrix[0]);

-    vals->point_matrix[1] = xlfd_round_double(vals->point_matrix[1]);

-    vals->point_matrix[2] = xlfd_round_double(vals->point_matrix[2]);

-    vals->point_matrix[3] = xlfd_round_double(vals->point_matrix[3]);

-

-    /* Fill in the deprecated fields for the benefit of rasterizers

-       that do not handle the matrices. */

-    vals->point = vals->point_matrix[3] * 10;

-    vals->pixel = vals->pixel_matrix[3];

-

-    return TRUE;

-}

-

-static Bool

-MatchScalable (FontScalablePtr a, FontScalablePtr b)

-{

-    int i;

-

-    /* Some asymmetry here:  we assume that the first argument (a) is

-       the table entry and the second (b) the item we're trying to match

-       (the key).  We'll consider the fonts matched if the relevant

-       metrics match *and* if a) the table entry doesn't have charset

-       subsetting or b) the table entry has identical charset subsetting

-       to that in the key.  We could add logic to check if the table

-       entry has a superset of the charset required by the key, but

-       we'll resist the urge for now.  */

-

-#define EQUAL(a,b) ((a)[0] == (b)[0] && \

-                    (a)[1] == (b)[1] && \

-                    (a)[2] == (b)[2] && \

-                    (a)[3] == (b)[3])

-

-    if (!(a->x == b->x &&

-	  a->y == b->y &&

-	  (a->width == b->width || a->width == 0 || b->width == 0 || b->width == -1) &&

-	  (!(b->values_supplied & PIXELSIZE_MASK) ||

-	    ((a->values_supplied & PIXELSIZE_MASK) ==

-	     (b->values_supplied & PIXELSIZE_MASK) &&

-	    EQUAL(a->pixel_matrix, b->pixel_matrix))) &&

-	  (!(b->values_supplied & POINTSIZE_MASK) ||

-	    ((a->values_supplied & POINTSIZE_MASK) ==

-	     (b->values_supplied & POINTSIZE_MASK) &&

-	    EQUAL(a->point_matrix, b->point_matrix))) &&

-	  (a->nranges == 0 || a->nranges == b->nranges)))

-      return FALSE;

-

-    for (i = 0; i < a->nranges; i++)

-	if (a->ranges[i].min_char_low != b->ranges[i].min_char_low ||

-	    a->ranges[i].min_char_high != b->ranges[i].min_char_high ||

-	    a->ranges[i].max_char_low != b->ranges[i].max_char_low ||

-	    a->ranges[i].max_char_high != b->ranges[i].max_char_high)

-		return FALSE;

- 

-    return TRUE;

-}

-

-FontScaledPtr

-FontFileFindScaledInstance (FontEntryPtr entry, FontScalablePtr vals, 

-			    int noSpecificSize)

-{

-    FontScalableEntryPtr    scalable;

-    FontScalableExtraPtr    extra;

-    FontScalablePtr	    mvals;

-    int			    dist, i;

-    int			    mini;

-    double		    mindist;

-    register double	    temp, sum=0.0;

-

-#define NORMDIFF(a, b) ( \

-    temp = (a)[0] - (b)[0], \

-    sum = temp * temp, \

-    temp = (a)[1] - (b)[1], \

-    sum += temp * temp, \

-    temp = (a)[2] - (b)[2], \

-    sum += temp * temp, \

-    temp = (a)[3] - (b)[3], \

-    sum + temp * temp )

-

-    scalable = &entry->u.scalable;

-    extra = scalable->extra;

-    if (noSpecificSize && extra->numScaled)

-    {

-	mini = 0;

-	mindist = NORMDIFF(extra->scaled[0].vals.point_matrix,

-			   vals->point_matrix);

-	for (i = 1; i < extra->numScaled; i++)

-	{

-	    if (extra->scaled[i].pFont &&

-		!extra->scaled[i].pFont->info.cachable) continue;

-	    mvals = &extra->scaled[i].vals;

-	    dist = NORMDIFF(mvals->point_matrix, vals->point_matrix);

-	    if (dist < mindist)

-	    {

-		mindist = dist;

-		mini = i;

-	    }

-	}

-	if (extra->scaled[mini].pFont &&

-	    !extra->scaled[mini].pFont->info.cachable) return 0;

-	return &extra->scaled[mini];

-    }

-    else

-    {

-    	/* See if we've scaled to this value yet */

-    	for (i = 0; i < extra->numScaled; i++)

-    	{

-	    if (extra->scaled[i].pFont &&

-		!extra->scaled[i].pFont->info.cachable) continue;

-	    if (MatchScalable (&extra->scaled[i].vals, vals))

-	    	return &extra->scaled[i];

-    	}

-    }

-    return 0;

-}

+/*
+
+Copyright 1991, 1998  The Open Group
+
+Permission to use, copy, modify, distribute, and sell this software and its
+documentation for any purpose is hereby granted without fee, provided that
+the above copyright notice appear in all copies and that both that
+copyright notice and this permission notice appear in supporting
+documentation.
+
+The above copyright notice and this permission notice shall be included in
+all copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL THE
+OPEN GROUP BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
+AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+
+Except as contained in this notice, the name of The Open Group shall not be
+used in advertising or otherwise to promote the sale, use or other dealings
+in this Software without prior written authorization from The Open Group.
+
+*/
+
+/*
+ * Author:  Keith Packard, MIT X Consortium
+ */
+
+#ifdef HAVE_CONFIG_H
+#include <config.h>
+#endif
+#include    <X11/fonts/fntfilst.h>
+#include <math.h>
+
+#ifdef _MSC_VER
+#define hypot _hypot
+#endif
+
+Bool
+FontFileAddScaledInstance (FontEntryPtr entry, FontScalablePtr vals, 
+			   FontPtr pFont, char *bitmapName)
+{
+    FontScalableEntryPtr    scalable;
+    FontScalableExtraPtr    extra;
+    FontScaledPtr	    new;
+    int			    newsize;
+
+    scalable = &entry->u.scalable;
+    extra = scalable->extra;
+    if (extra->numScaled == extra->sizeScaled)
+    {
+	newsize = extra->sizeScaled + 4;
+	new = realloc (extra->scaled, newsize * sizeof (FontScaledRec));
+	if (!new)
+	    return FALSE;
+	extra->sizeScaled = newsize;
+	extra->scaled = new;
+    }
+    new = &extra->scaled[extra->numScaled++];
+    new->vals = *vals;
+    new->pFont = pFont;
+    new->bitmap = (FontEntryPtr) bitmapName;
+    if (pFont)
+	pFont->fpePrivate = (pointer) entry;
+    return TRUE;
+}
+
+/* Must call this after the directory is sorted */
+
+void
+FontFileSwitchStringsToBitmapPointers (FontDirectoryPtr dir)
+{
+    int	    s;
+    int	    b;
+    int	    i;
+    FontEntryPtr	    scalable;
+    FontEntryPtr	    nonScalable;
+    FontScaledPtr	    scaled;
+    FontScalableExtraPtr    extra;
+    
+    scalable = dir->scalable.entries;
+    nonScalable = dir->nonScalable.entries;
+    for (s = 0; s < dir->scalable.used; s++)
+    {
+	extra = scalable[s].u.scalable.extra;
+	scaled = extra->scaled;
+	for (i = 0; i < extra->numScaled; i++)
+	    for (b = 0; b < dir->nonScalable.used; b++)
+		if (nonScalable[b].name.name == (char *) scaled[i].bitmap)
+		    scaled[i].bitmap = &nonScalable[b];
+    }
+}
+
+void
+FontFileRemoveScaledInstance (FontEntryPtr entry, FontPtr pFont)
+{
+    FontScalableEntryPtr    scalable;
+    FontScalableExtraPtr    extra;
+    int			    i;
+
+    scalable = &entry->u.scalable;
+    extra = scalable->extra;
+    for (i = 0; i < extra->numScaled; i++)
+    {
+	if (extra->scaled[i].pFont == pFont)
+	{
+	    if (extra->scaled[i].vals.ranges)
+		free (extra->scaled[i].vals.ranges);
+	    extra->numScaled--;
+	    for (; i < extra->numScaled; i++)
+		extra->scaled[i] = extra->scaled[i+1];
+	}
+    }
+}
+
+Bool
+FontFileCompleteXLFD (FontScalablePtr vals, FontScalablePtr def)
+{
+    FontResolutionPtr res;
+    int		num_res;
+    double	sx, sy, temp_matrix[4];
+    double	pixel_setsize_adjustment = 1.0;
+    /*
+     * If two of the three vertical scale values are specified, compute the
+     * third.  If all three are specified, make sure they are consistent
+     * (within a pixel)
+     *
+     * One purpose of this procedure is to complete XLFD names in a
+     * repeatable manner.  That is, if the user partially specifies
+     * a name (say, pixelsize but not pointsize), the results generated
+     * here result in a fully specified name that will result in the
+     * same font.
+     */
+
+    res = GetClientResolutions(&num_res);
+
+    if (!(vals->values_supplied & PIXELSIZE_MASK) ||
+	!(vals->values_supplied & POINTSIZE_MASK))
+    {
+	/* If resolution(s) unspecified and cannot be computed from
+	   pixelsize and pointsize, get appropriate defaults. */
+
+	if (num_res)
+	{
+	    if (vals->x <= 0)
+		vals->x = res->x_resolution;
+	    if (vals->y <= 0)
+		vals->y = res->y_resolution;
+	}
+
+	if (vals->x <= 0)
+	    vals->x = def->x;
+	if (vals->y <= 0)
+	    vals->y = def->y;
+    }
+    else
+    {
+	/* If needed, compute resolution values from the pixel and
+	   pointsize information we were given.  This problem is
+	   overdetermined (four equations, two unknowns), but we don't
+	   check for inconsistencies here.  If they exist, they will
+	   show up in later tests for the point and pixel sizes.  */
+
+	if (vals->y <= 0)
+	{
+	    double x = hypot(vals->pixel_matrix[1], vals->pixel_matrix[3]);
+	    double y = hypot(vals->point_matrix[1], vals->point_matrix[3]);
+	    if (y < EPS) return FALSE;
+	    vals->y = (int)(x * 72.27 / y + .5);
+	}
+	if (vals->x <= 0)
+	{
+	    /* If the pixelsize was given as an array, or as a scalar that
+	       has been normalized for the pixel shape, we have enough
+	       information to compute a separate horizontal resolution */
+
+	    if ((vals->values_supplied & PIXELSIZE_MASK) == PIXELSIZE_ARRAY ||
+	        (vals->values_supplied & PIXELSIZE_MASK) ==
+		    PIXELSIZE_SCALAR_NORMALIZED)
+	    {
+		double x = hypot(vals->pixel_matrix[0], vals->pixel_matrix[2]);
+		double y = hypot(vals->point_matrix[0], vals->point_matrix[2]);
+		if (y < EPS) return FALSE;
+		vals->x = (int)(x * 72.27 / y + .5);
+	    }
+	    else
+	    {
+		/* Not enough information in the pixelsize array.  Just
+		   assume the pixels are square. */
+		vals->x = vals->y;
+	    }
+	}
+    }
+
+    if (vals->x <= 0 || vals->y <= 0) return FALSE;
+
+    /* If neither pixelsize nor pointsize is defined, take the pointsize
+       from the defaults structure we've been passed. */
+    if (!(vals->values_supplied & PIXELSIZE_MASK) &&
+	!(vals->values_supplied & POINTSIZE_MASK))
+    {
+	if (num_res)
+	{
+	    vals->point_matrix[0] =
+	    vals->point_matrix[3] = (double)res->point_size / 10.0;
+	    vals->point_matrix[1] =
+	    vals->point_matrix[2] = 0;
+	    vals->values_supplied = (vals->values_supplied & ~POINTSIZE_MASK) |
+				    POINTSIZE_SCALAR;
+	}
+	else if (def->values_supplied & POINTSIZE_MASK)
+	{
+	    vals->point_matrix[0] = def->point_matrix[0];
+	    vals->point_matrix[1] = def->point_matrix[1];
+	    vals->point_matrix[2] = def->point_matrix[2];
+	    vals->point_matrix[3] = def->point_matrix[3];
+	    vals->values_supplied = (vals->values_supplied & ~POINTSIZE_MASK) |
+				    (def->values_supplied & POINTSIZE_MASK);
+	}
+	else return FALSE;
+    }
+
+    /* At this point, at least two of the three vertical scale values
+       should be specified.  Our job now is to compute the missing ones
+       and check for agreement between overspecified values */
+
+    /* If pixelsize was specified by a scalar, we need to fix the matrix
+       now that we know the resolutions.  */
+    if ((vals->values_supplied & PIXELSIZE_MASK) == PIXELSIZE_SCALAR)
+    {
+	/* pixel_setsize_adjustment used below to modify permissible
+	   error in pixel/pointsize matching, since multiplying a
+	   number rounded to integer changes the amount of the error
+	   caused by the rounding */
+
+	pixel_setsize_adjustment = (double)vals->x / (double)vals->y;
+	vals->pixel_matrix[0] *= pixel_setsize_adjustment;
+	vals->values_supplied  = (vals->values_supplied & ~PIXELSIZE_MASK) |
+				 PIXELSIZE_SCALAR_NORMALIZED;
+    }
+
+    sx = (double)vals->x / 72.27;
+    sy = (double)vals->y / 72.27;
+
+    /* If a pointsize was specified, make sure pixelsize is consistent
+       to within 1 pixel, then replace pixelsize with a consistent
+       floating-point value.  */
+
+    if (vals->values_supplied & POINTSIZE_MASK)
+    {
+    recompute_pixelsize: ;
+	temp_matrix[0] = vals->point_matrix[0] * sx;
+	temp_matrix[1] = vals->point_matrix[1] * sy;
+	temp_matrix[2] = vals->point_matrix[2] * sx;
+	temp_matrix[3] = vals->point_matrix[3] * sy;
+	if (vals->values_supplied & PIXELSIZE_MASK)
+	{
+	    if (fabs(vals->pixel_matrix[0] - temp_matrix[0]) >
+		    pixel_setsize_adjustment ||
+		fabs(vals->pixel_matrix[1] - temp_matrix[1]) > 1 ||
+		fabs(vals->pixel_matrix[2] - temp_matrix[2]) > 1 ||
+		fabs(vals->pixel_matrix[3] - temp_matrix[3]) > 1)
+		return FALSE;
+	}
+	if ((vals->values_supplied & PIXELSIZE_MASK) == PIXELSIZE_ARRAY &&
+	    (vals->values_supplied & POINTSIZE_MASK) == POINTSIZE_SCALAR)
+	{
+	    /* In the special case that pixelsize came as an array and
+	       pointsize as a scalar, recompute the pointsize matrix
+	       from the pixelsize matrix. */
+	    goto recompute_pointsize;
+	}
+
+	/* Refresh pixel matrix with precise values computed from
+	   pointsize and resolution.  */
+	vals->pixel_matrix[0] = temp_matrix[0];
+	vals->pixel_matrix[1] = temp_matrix[1];
+	vals->pixel_matrix[2] = temp_matrix[2];
+	vals->pixel_matrix[3] = temp_matrix[3];
+
+	/* Set values_supplied for pixel to match that for point */
+	vals->values_supplied =
+	    (vals->values_supplied & ~PIXELSIZE_MASK) |
+	    (((vals->values_supplied & POINTSIZE_MASK) == POINTSIZE_ARRAY) ?
+		PIXELSIZE_ARRAY : PIXELSIZE_SCALAR_NORMALIZED);
+    }
+    else
+    {
+	/* Pointsize unspecified...  compute from pixel size and
+	   resolutions */
+    recompute_pointsize: ;
+	if (fabs(sx) < EPS || fabs(sy) < EPS) return FALSE;
+	vals->point_matrix[0] = vals->pixel_matrix[0] / sx;
+	vals->point_matrix[1] = vals->pixel_matrix[1] / sy;
+	vals->point_matrix[2] = vals->pixel_matrix[2] / sx;
+	vals->point_matrix[3] = vals->pixel_matrix[3] / sy;
+
+	/* Set values_supplied for pixel to match that for point */
+	vals->values_supplied =
+	    (vals->values_supplied & ~POINTSIZE_MASK) |
+	    (((vals->values_supplied & PIXELSIZE_MASK) == PIXELSIZE_ARRAY) ?
+		POINTSIZE_ARRAY : POINTSIZE_SCALAR);
+
+	/* If we computed scalar pointsize from scalar pixelsize, round
+	   pointsize to decipoints and recompute pixelsize so we end up
+	   with a repeatable name */
+	if ((vals->values_supplied & POINTSIZE_MASK) == POINTSIZE_SCALAR)
+	{
+	    /* Off-diagonal elements should be zero since no matrix was
+	       specified. */
+	    vals->point_matrix[0] =
+		(double)(int)(vals->point_matrix[0] * 10.0 + .5) / 10.0;
+	    vals->point_matrix[3] =
+		(double)(int)(vals->point_matrix[3] * 10.0 + .5) / 10.0;
+	    goto recompute_pixelsize;
+	}
+    }
+
+    /* We've succeeded.  Round everything to a few decimal places
+       for repeatability. */
+
+    vals->pixel_matrix[0] = xlfd_round_double(vals->pixel_matrix[0]);
+    vals->pixel_matrix[1] = xlfd_round_double(vals->pixel_matrix[1]);
+    vals->pixel_matrix[2] = xlfd_round_double(vals->pixel_matrix[2]);
+    vals->pixel_matrix[3] = xlfd_round_double(vals->pixel_matrix[3]);
+    vals->point_matrix[0] = xlfd_round_double(vals->point_matrix[0]);
+    vals->point_matrix[1] = xlfd_round_double(vals->point_matrix[1]);
+    vals->point_matrix[2] = xlfd_round_double(vals->point_matrix[2]);
+    vals->point_matrix[3] = xlfd_round_double(vals->point_matrix[3]);
+
+    /* Fill in the deprecated fields for the benefit of rasterizers
+       that do not handle the matrices. */
+    vals->point = vals->point_matrix[3] * 10;
+    vals->pixel = vals->pixel_matrix[3];
+
+    return TRUE;
+}
+
+static Bool
+MatchScalable (FontScalablePtr a, FontScalablePtr b)
+{
+    int i;
+
+    /* Some asymmetry here:  we assume that the first argument (a) is
+       the table entry and the second (b) the item we're trying to match
+       (the key).  We'll consider the fonts matched if the relevant
+       metrics match *and* if a) the table entry doesn't have charset
+       subsetting or b) the table entry has identical charset subsetting
+       to that in the key.  We could add logic to check if the table
+       entry has a superset of the charset required by the key, but
+       we'll resist the urge for now.  */
+
+#define EQUAL(a,b) ((a)[0] == (b)[0] && \
+                    (a)[1] == (b)[1] && \
+                    (a)[2] == (b)[2] && \
+                    (a)[3] == (b)[3])
+
+    if (!(a->x == b->x &&
+	  a->y == b->y &&
+	  (a->width == b->width || a->width == 0 || b->width == 0 || b->width == -1) &&
+	  (!(b->values_supplied & PIXELSIZE_MASK) ||
+	    ((a->values_supplied & PIXELSIZE_MASK) ==
+	     (b->values_supplied & PIXELSIZE_MASK) &&
+	    EQUAL(a->pixel_matrix, b->pixel_matrix))) &&
+	  (!(b->values_supplied & POINTSIZE_MASK) ||
+	    ((a->values_supplied & POINTSIZE_MASK) ==
+	     (b->values_supplied & POINTSIZE_MASK) &&
+	    EQUAL(a->point_matrix, b->point_matrix))) &&
+	  (a->nranges == 0 || a->nranges == b->nranges)))
+      return FALSE;
+
+    for (i = 0; i < a->nranges; i++)
+	if (a->ranges[i].min_char_low != b->ranges[i].min_char_low ||
+	    a->ranges[i].min_char_high != b->ranges[i].min_char_high ||
+	    a->ranges[i].max_char_low != b->ranges[i].max_char_low ||
+	    a->ranges[i].max_char_high != b->ranges[i].max_char_high)
+		return FALSE;
+ 
+    return TRUE;
+}
+
+FontScaledPtr
+FontFileFindScaledInstance (FontEntryPtr entry, FontScalablePtr vals, 
+			    int noSpecificSize)
+{
+    FontScalableEntryPtr    scalable;
+    FontScalableExtraPtr    extra;
+    FontScalablePtr	    mvals;
+    int			    dist, i;
+    int			    mini;
+    double		    mindist;
+    register double	    temp, sum=0.0;
+
+#define NORMDIFF(a, b) ( \
+    temp = (a)[0] - (b)[0], \
+    sum = temp * temp, \
+    temp = (a)[1] - (b)[1], \
+    sum += temp * temp, \
+    temp = (a)[2] - (b)[2], \
+    sum += temp * temp, \
+    temp = (a)[3] - (b)[3], \
+    sum + temp * temp )
+
+    scalable = &entry->u.scalable;
+    extra = scalable->extra;
+    if (noSpecificSize && extra->numScaled)
+    {
+	mini = 0;
+	mindist = NORMDIFF(extra->scaled[0].vals.point_matrix,
+			   vals->point_matrix);
+	for (i = 1; i < extra->numScaled; i++)
+	{
+	    if (extra->scaled[i].pFont &&
+		!extra->scaled[i].pFont->info.cachable) continue;
+	    mvals = &extra->scaled[i].vals;
+	    dist = NORMDIFF(mvals->point_matrix, vals->point_matrix);
+	    if (dist < mindist)
+	    {
+		mindist = dist;
+		mini = i;
+	    }
+	}
+	if (extra->scaled[mini].pFont &&
+	    !extra->scaled[mini].pFont->info.cachable) return 0;
+	return &extra->scaled[mini];
+    }
+    else
+    {
+    	/* See if we've scaled to this value yet */
+    	for (i = 0; i < extra->numScaled; i++)
+    	{
+	    if (extra->scaled[i].pFont &&
+		!extra->scaled[i].pFont->info.cachable) continue;
+	    if (MatchScalable (&extra->scaled[i].vals, vals))
+	    	return &extra->scaled[i];
+    	}
+    }
+    return 0;
+}