2 * Copyright © 2012 Google, Inc.
4 * This is part of HarfBuzz, a text shaping library.
6 * Permission is hereby granted, without written agreement and without
7 * license or royalty fees, to use, copy, modify, and distribute this
8 * software and its documentation for any purpose, provided that the
9 * above copyright notice and the following two paragraphs appear in
10 * all copies of this software.
12 * IN NO EVENT SHALL THE COPYRIGHT HOLDER BE LIABLE TO ANY PARTY FOR
13 * DIRECT, INDIRECT, SPECIAL, INCIDENTAL, OR CONSEQUENTIAL DAMAGES
14 * ARISING OUT OF THE USE OF THIS SOFTWARE AND ITS DOCUMENTATION, EVEN
15 * IF THE COPYRIGHT HOLDER HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH
18 * THE COPYRIGHT HOLDER SPECIFICALLY DISCLAIMS ANY WARRANTIES, INCLUDING,
19 * BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND
20 * FITNESS FOR A PARTICULAR PURPOSE. THE SOFTWARE PROVIDED HEREUNDER IS
21 * ON AN "AS IS" BASIS, AND THE COPYRIGHT HOLDER HAS NO OBLIGATION TO
22 * PROVIDE MAINTENANCE, SUPPORT, UPDATES, ENHANCEMENTS, OR MODIFICATIONS.
24 * Google Author(s): Behdad Esfahbod
27 #ifndef HB_SET_DIGEST_HH
28 #define HB_SET_DIGEST_HH
33 * The set digests here implement various "filters" that support
34 * "approximate member query". Conceptually these are like Bloom
35 * Filter and Quotient Filter, however, much smaller, faster, and
36 * designed to fit the requirements of our uses for glyph coverage
39 * Our filters are highly accurate if the lookup covers fairly local
40 * set of glyphs, but fully flooded and ineffective if coverage is
43 * The frozen-set can be used instead of a digest, to trade more
44 * memory for 100% accuracy, but in practice, that doesn't look like
45 * an attractive trade-off.
48 template <typename mask_t, unsigned int shift>
49 struct hb_set_digest_lowest_bits_t
51 static constexpr unsigned mask_bytes = sizeof (mask_t);
52 static constexpr unsigned mask_bits = sizeof (mask_t) * 8;
53 static constexpr unsigned num_bits = 0
54 + (mask_bytes >= 1 ? 3 : 0)
55 + (mask_bytes >= 2 ? 1 : 0)
56 + (mask_bytes >= 4 ? 1 : 0)
57 + (mask_bytes >= 8 ? 1 : 0)
58 + (mask_bytes >= 16? 1 : 0)
61 static_assert ((shift < sizeof (hb_codepoint_t) * 8), "");
62 static_assert ((shift + num_bits <= sizeof (hb_codepoint_t) * 8), "");
64 void init () { mask = 0; }
66 void add (hb_codepoint_t g) { mask |= mask_for (g); }
68 bool add_range (hb_codepoint_t a, hb_codepoint_t b)
70 if ((b >> shift) - (a >> shift) >= mask_bits - 1)
73 mask_t ma = mask_for (a);
74 mask_t mb = mask_for (b);
75 mask |= mb + (mb - ma) - (mb < ma);
81 void add_array (const T *array, unsigned int count, unsigned int stride=sizeof(T))
83 for (unsigned int i = 0; i < count; i++)
86 array = (const T *) (stride + (const char *) array);
90 bool add_sorted_array (const T *array, unsigned int count, unsigned int stride=sizeof(T))
92 for (unsigned int i = 0; i < count; i++)
95 array = (const T *) (stride + (const char *) array);
100 bool may_have (hb_codepoint_t g) const
101 { return !!(mask & mask_for (g)); }
105 static mask_t mask_for (hb_codepoint_t g)
106 { return ((mask_t) 1) << ((g >> shift) & (mask_bits - 1)); }
110 template <typename head_t, typename tail_t>
111 struct hb_set_digest_combiner_t
119 void add (hb_codepoint_t g)
125 bool add_range (hb_codepoint_t a, hb_codepoint_t b)
127 head.add_range (a, b);
128 tail.add_range (a, b);
131 template <typename T>
132 void add_array (const T *array, unsigned int count, unsigned int stride=sizeof(T))
134 head.add_array (array, count, stride);
135 tail.add_array (array, count, stride);
137 template <typename T>
138 bool add_sorted_array (const T *array, unsigned int count, unsigned int stride=sizeof(T))
140 head.add_sorted_array (array, count, stride);
141 tail.add_sorted_array (array, count, stride);
145 bool may_have (hb_codepoint_t g) const
147 return head.may_have (g) && tail.may_have (g);
159 * This is a combination of digests that performs "best".
160 * There is not much science to this: it's a result of intuition
163 typedef hb_set_digest_combiner_t
165 hb_set_digest_lowest_bits_t<unsigned long, 4>,
166 hb_set_digest_combiner_t
168 hb_set_digest_lowest_bits_t<unsigned long, 0>,
169 hb_set_digest_lowest_bits_t<unsigned long, 9>
174 #endif /* HB_SET_DIGEST_HH */