Reduce number of allocations during FcSortWalk().

The current behaviour of FcSortWalk() is to create a new FcCharSet on
each iteration that is the union of the previous iteration with the next
FcCharSet in the font set. This causes the existing FcCharSet to be
reproduced in its entirety and then allocates fresh leaves for the new
FcCharSet. In essence the number of allocations is quadratic wrt the
number of fonts required.

By introducing a new method for merging a new FcCharSet with an existing
one we can change the behaviour to be effectively linear with the number
of fonts - allocating no more leaves than necessary to cover all the
fonts in the set.

For example, profiling 'gedit UTF-8-demo.txt'
    Allocator		    nAllocs	    nBytes
Before:
    FcCharSetFindLeafCreate 62886	    2012352
    FcCharSetPutLeaf        9361	    11441108
After:
    FcCharSetFindLeafCreate 1940	    62080
    FcCharSetPutLeaf        281		    190336

The savings are even more significant for applications like firefox-3.0b5
which need to switch between large number of fonts.
Before:
    FcCharSetFindLeafCreate 4461192	    142758144
    FcCharSetPutLeaf	    1124536	    451574172
After:
    FcCharSetFindLeafCreate 80359	    2571488
    FcCharSetPutLeaf	    18940	    9720522

Out of interest, the next most frequent allocations are
    FcPatternObjectAddWithBinding 526029    10520580
    tt_face_load_eblc	    42103	    2529892

src/fccharset.c

@@ -452,6 +452,68 @@ FcCharSetUnion (const FcCharSet *a, const FcCharSet *b)
     return FcCharSetOperate (a, b, FcCharSetUnionLeaf, FcTrue, FcTrue);
 }
 
+FcCharSet *
+FcCharSetMerge (FcCharSet *a, const FcCharSet *b)
+{
+    FcCharSet	    *fcs;
+    FcCharSetIter    ai, bi;
+
+    if (a == NULL) {
+	return FcCharSetCopy ((FcCharSet *) b);
+    } else if (a->ref == FC_REF_CONSTANT) {
+	fcs = FcCharSetCreate ();
+	if (fcs == NULL)
+	    return NULL;
+    } else
+	fcs = a;
+
+    FcCharSetIterStart (a, &ai);
+    FcCharSetIterStart (b, &bi);
+    while (ai.leaf || bi.leaf)
+    {
+	if (ai.ucs4 < bi.ucs4)
+	{
+	    if (!FcCharSetAddLeaf (fcs, ai.ucs4, ai.leaf))
+		goto bail;
+
+	    FcCharSetIterNext (a, &ai);
+	}
+	else if (bi.ucs4 < ai.ucs4)
+	{
+	    if (!FcCharSetAddLeaf (fcs, bi.ucs4, bi.leaf))
+		goto bail;
+
+	    FcCharSetIterNext (b, &bi);
+	}
+	else
+	{
+	    FcCharLeaf  leaf;
+
+	    if (FcCharSetUnionLeaf (&leaf, ai.leaf, bi.leaf))
+	    {
+		if (!FcCharSetAddLeaf (fcs, ai.ucs4, &leaf))
+		    goto bail;
+	    }
+
+	    FcCharSetIterNext (a, &ai);
+	    FcCharSetIterNext (b, &bi);
+	}
+    }
+
+    if (fcs != a)
+	FcCharSetDestroy (a);
+
+    return fcs;
+
+bail:
+    FcCharSetDestroy (fcs);
+
+    if (fcs != a)
+	FcCharSetDestroy (a);
+
+    return NULL;
+}
+
 static FcBool
 FcCharSetSubtractLeaf (FcCharLeaf *result,
 		       const FcCharLeaf *al,

src/fcint.h

@@ -641,6 +641,9 @@ FcNameParseCharSet (FcChar8 *string);
 FcPrivate FcCharLeaf *
 FcCharSetFindLeafCreate (FcCharSet *fcs, FcChar32 ucs4);
 
+FcPrivate FcCharSet *
+FcCharSetMerge (FcCharSet *a, const FcCharSet *b);
+
 FcPrivate FcBool
 FcCharSetSerializeAlloc(FcSerialize *serialize, const FcCharSet *cs);
 

src/fcmatch.c

@@ -601,16 +601,9 @@ FcSortWalk (FcSortNode **n, int nnode, FcFontSet *fs, FcCharSet **cs, FcBool tri
 	    {
                 if (trim || build_cs)
                 {
-                    if (*cs)
-                    {
-                        ncs = FcCharSetUnion (ncs, *cs);
-                        if (!ncs)
+		    *cs = FcCharSetMerge (*cs, ncs);
+		    if (*cs == NULL)
 			return FcFalse;
-                        FcCharSetDestroy (*cs);
-                    }
-                    else
-                        ncs = FcCharSetCopy (ncs);
-                    *cs = ncs;
                 }
 
 		FcPatternReference (node->pattern);