hb-ot-shape-normalize.cc revision c311d852080b50ffc85e80168de62abb05a6be59
1/*
2 * Copyright © 2011  Google, Inc.
3 *
4 *  This is part of HarfBuzz, a text shaping library.
5 *
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.
11 *
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
16 * DAMAGE.
17 *
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.
23 *
24 * Google Author(s): Behdad Esfahbod
25 */
26
27#include "hb-ot-shape-private.hh"
28#include "hb-ot-shape-complex-private.hh"
29
30HB_BEGIN_DECLS
31
32/*
33 * HIGHLEVEL DESIGN:
34 *
35 * This file exports one main function: _hb_ot_shape_normalize().
36 *
37 * This function closely reflects the Unicode Normalization Algorithm,
38 * yet it's different.  The shaper an either prefer decomposed (NFD) or
39 * composed (NFC).
40 *
41 * In general what happens is that: each grapheme is decomposed in a chain
42 * of 1:2 decompositions, marks reordered, and then recomposed if desires,
43 * so far it's like Unicode Normalization.  However, the decomposition and
44 * recomposition only happens if the font supports the resulting characters.
45 *
46 * The goals are:
47 *
48 *   - Try to render all canonically equivalent strings similarly.  To really
49 *     achieve this we have to always do the full decomposition and then
50 *     selectively recompose from there.  It's kinda too expensive though, so
51 *     we skip some cases.  For example, if composed is desired, we simply
52 *     don't touch 1-character clusters that are supported by the font, even
53 *     though their NFC may be different.
54 *
55 *   - When a font has a precomposed character for a sequence but the 'ccmp'
56 *     feature in the font is not adequate, form use the precomposed character
57 *     which typically has better mark positioning.
58 *
59 *   - When a font does not support a character but supports its decomposition,
60 *     well, use the decomposition.
61 *
62 *   - The Indic shaper requests decomposed output.  This will handle splitting
63 *     matra for the Indic shaper.
64 */
65
66static void
67output_glyph (hb_ot_shape_context_t *c,
68	      hb_codepoint_t glyph)
69{
70  hb_buffer_t *buffer = c->buffer;
71
72  buffer->output_glyph (glyph);
73  hb_glyph_info_set_unicode_props (&buffer->out_info[buffer->out_len - 1], buffer->unicode);
74}
75
76static bool
77decompose (hb_ot_shape_context_t *c,
78	   bool shortest,
79	   hb_codepoint_t ab)
80{
81  hb_codepoint_t a, b, glyph;
82
83  if (!hb_unicode_decompose (c->buffer->unicode, ab, &a, &b) ||
84      (b && !hb_font_get_glyph (c->font, b, 0, &glyph)))
85    return FALSE;
86
87  bool has_a = hb_font_get_glyph (c->font, a, 0, &glyph);
88  if (shortest && has_a) {
89    /* Output a and b */
90    output_glyph (c, a);
91    if (b)
92      output_glyph (c, b);
93    return TRUE;
94  }
95
96  if (decompose (c, shortest, a)) {
97    if (b)
98      output_glyph (c, b);
99    return TRUE;
100  }
101
102  if (has_a) {
103    output_glyph (c, a);
104    if (b)
105      output_glyph (c, b);
106    return TRUE;
107  }
108
109  return FALSE;
110}
111
112static void
113decompose_current_glyph (hb_ot_shape_context_t *c,
114			 bool shortest)
115{
116  if (decompose (c, shortest, c->buffer->info[c->buffer->idx].codepoint))
117    c->buffer->skip_glyph ();
118  else
119    c->buffer->next_glyph ();
120}
121
122static void
123decompose_single_char_cluster (hb_ot_shape_context_t *c,
124			       bool will_recompose)
125{
126  hb_codepoint_t glyph;
127
128  /* If recomposing and font supports this, we're good to go */
129  if (will_recompose && hb_font_get_glyph (c->font, c->buffer->info[c->buffer->idx].codepoint, 0, &glyph)) {
130    c->buffer->next_glyph ();
131    return;
132  }
133
134  decompose_current_glyph (c, will_recompose);
135}
136
137static void
138decompose_multi_char_cluster (hb_ot_shape_context_t *c,
139			      unsigned int end)
140{
141  /* TODO Currently if there's a variation-selector we give-up, it's just too hard. */
142  for (unsigned int i = c->buffer->idx; i < end; i++)
143    if (unlikely (is_variation_selector (c->buffer->info[i].codepoint)))
144      return;
145
146  while (c->buffer->idx < end)
147    decompose_current_glyph (c, FALSE);
148}
149
150void
151_hb_ot_shape_normalize (hb_ot_shape_context_t *c)
152{
153  hb_buffer_t *buffer = c->buffer;
154  bool recompose = !hb_ot_shape_complex_prefer_decomposed (c->plan->shaper);
155  bool has_multichar_clusters = FALSE;
156  unsigned int count;
157
158  /* We do a fairly straightforward yet custom normalization process in three
159   * separate rounds: decompose, reorder, recompose (if desired).  Currently
160   * this makes two buffer swaps.  We can make it faster by moving the last
161   * two rounds into the inner loop for the first round, but it's more readable
162   * this way. */
163
164
165  /* First round, decompose */
166
167  buffer->clear_output ();
168  count = buffer->len;
169  for (buffer->idx = 0; buffer->idx < count;)
170  {
171    unsigned int end;
172    for (end = buffer->idx + 1; end < count; end++)
173      if (buffer->info[buffer->idx].cluster != buffer->info[end].cluster)
174        break;
175
176    if (buffer->idx + 1 == end)
177      decompose_single_char_cluster (c, recompose);
178    else {
179      decompose_multi_char_cluster (c, end);
180      has_multichar_clusters = TRUE;
181    }
182  }
183  buffer->swap_buffers ();
184
185
186  /* Technically speaking, two characters with ccc=0 may combine.  But all
187   * those cases are in languages that the indic module handles (which expects
188   * decomposed), or in Hangul jamo, which again, we want decomposed anyway.
189   * So we don't bother combining across cluster boundaries.
190   *
191   * TODO: Am I right about Hangul?  If I am, we should add a Hangul module
192   * that requests decomposed. */
193
194  if (!has_multichar_clusters)
195    return; /* Done! */
196
197
198  /* Second round, reorder (inplace) */
199
200  count = buffer->len;
201  for (unsigned int i = 0; i < count; i++)
202  {
203    if (buffer->info[i].combining_class() == 0)
204      continue;
205
206    unsigned int end;
207    for (end = i + 1; end < count; end++)
208      if (buffer->info[end].combining_class() == 0)
209        break;
210
211    /* We are going to do a bubble-sort.  Only do this if the
212     * sequence is short.  Doing it on long sequences can result
213     * in an O(n^2) DoS. */
214    if (end - i > 10) {
215      i = end;
216      continue;
217    }
218
219    unsigned int k = end - i - 1;
220    do {
221      hb_glyph_info_t *pinfo = buffer->info + i;
222      unsigned int new_k = 0;
223
224      for (unsigned int j = 0; j < k; j++)
225	if (pinfo[j].combining_class() > pinfo[j+1].combining_class()) {
226	  hb_glyph_info_t t;
227	  t = pinfo[j];
228	  pinfo[j] = pinfo[j + 1];
229	  pinfo[j + 1] = t;
230
231	  new_k = j;
232	}
233      k = new_k;
234    } while (k);
235
236    i = end;
237  }
238
239
240  if (!recompose)
241    return;
242
243  /* Third round, recompose */
244
245  /* As noted in the comment earlier, we don't try to combine
246   * ccc=0 chars with their previous Starter. */
247
248  buffer->clear_output ();
249  count = buffer->len;
250  unsigned int starter = 0;
251  buffer->next_glyph ();
252  while (buffer->idx < count)
253  {
254    if (buffer->info[buffer->idx].combining_class() == 0) {
255      starter = buffer->out_len;
256      buffer->next_glyph ();
257      continue;
258    }
259
260    hb_codepoint_t composed, glyph;
261    if ((buffer->out_info[buffer->out_len - 1].combining_class() >=
262	 buffer->info[buffer->idx].combining_class()) ||
263	!hb_unicode_compose (c->buffer->unicode,
264			     buffer->out_info[starter].codepoint,
265			     buffer->info[buffer->idx].codepoint,
266			     &composed) ||
267	!hb_font_get_glyph (c->font, composed, 0, &glyph))
268    {
269      /* Blocked, or doesn't compose. */
270      buffer->next_glyph ();
271      continue;
272    }
273
274    /* Composes. Modify starter and carry on. */
275    buffer->out_info[starter].codepoint = composed;
276    hb_glyph_info_set_unicode_props (&buffer->out_info[starter], buffer->unicode);
277
278    buffer->skip_glyph ();
279  }
280  buffer->swap_buffers ();
281
282}
283
284HB_END_DECLS
285