Roll HarfBuzz to 1.1.1

third_party/harfbuzz-ng/src/hb-ot-shape-normalize.cc

 /*
  * Copyright © 2011,2012  Google, Inc.
  *
  *  This is part of HarfBuzz, a text shaping library.
  *
  * Permission is hereby granted, without written agreement and without
  * license or royalty fees, to use, copy, modify, and distribute this
  * software and its documentation for any purpose, provided that the
  * above copyright notice and the following two paragraphs appear in
  * all copies of this software.
@@ skipping 44 matching lines @@
  *     though their NFC may be different.
  *
  *   - When a font has a precomposed character for a sequence but the 'ccmp'
  *     feature in the font is not adequate, use the precomposed character
  *     which typically has better mark positioning.
  *
  *   - When a font does not support a combining mark, but supports it precomposed
  *     with previous base, use that.  This needs the itemizer to have this
  *     knowledge too.  We need to provide assistance to the itemizer.
  *
- *   - When a font does not support a character but supports its decomposition,
- *     well, use the decomposition (preferring the canonical decomposition, but
- *     falling back to the compatibility decomposition if necessary).  The
- *     compatibility decomposition is really nice to have, for characters like
- *     ellipsis, or various-sized space characters.
+ *   - When a font does not support a character but supports its canonical
+ *     decomposition, well, use the decomposition.
  *
  *   - The complex shapers can customize the compose and decompose functions to
  *     offload some of their requirements to the normalizer.  For example, the
  *     Indic shaper may want to disallow recomposing of two matras.
- *
- *   - We try compatibility decomposition if decomposing through canonical
- *     decomposition alone failed to find a sequence that the font supports.
- *     We don't try compatibility decomposition recursively during the canonical
- *     decomposition phase.  This has minimal impact.  There are only a handful
- *     of Greek letter that have canonical decompositions that include characters
- *     with compatibility decomposition.  Those can be found using this command:
- *
- *     egrep  "`echo -n ';('; grep ';<' UnicodeData.txt | cut -d';' -f1 | tr '\n' '|'; echo ') '`" UnicodeData.txt
  */
 
 static bool
 decompose_unicode (const hb_ot_shape_normalize_context_t *c,
                    hb_codepoint_t  ab,
                    hb_codepoint_t *a,
                    hb_codepoint_t *b)
 {
   return c->unicode->decompose (ab, a, b);
 }
@@ skipping 10 matching lines @@
 static inline void
 set_glyph (hb_glyph_info_t &info, hb_font_t *font)
 {
   font->get_glyph (info.codepoint, 0, &info.glyph_index());
 }
 
 static inline void
 output_char (hb_buffer_t *buffer, hb_codepoint_t unichar, hb_codepoint_t glyph)
 {
   buffer->cur().glyph_index() = glyph;
-  buffer->output_glyph (unichar);
-  _hb_glyph_info_set_unicode_props (&buffer->prev(), buffer->unicode);
+  buffer->output_glyph (unichar); /* This is very confusing indeed. */
+  _hb_glyph_info_set_unicode_props (&buffer->prev(), buffer);
 }
 
 static inline void
 next_char (hb_buffer_t *buffer, hb_codepoint_t glyph)
 {
   buffer->cur().glyph_index() = glyph;
   buffer->next_glyph ();
 }
 
 static inline void
@@ skipping 46 matching lines @@
   return 0;
 }
 
 static inline void
 decompose_current_character (const hb_ot_shape_normalize_context_t *c, bool shortest)
 {
   hb_buffer_t * const buffer = c->buffer;
   hb_codepoint_t u = buffer->cur().codepoint;
   hb_codepoint_t glyph;
 
-  /* Kind of a cute waterfall here... */
   if (shortest && c->font->get_glyph (u, 0, &glyph))
+  {
     next_char (buffer, glyph);
-  else if (decompose (c, shortest, u))
+    return;
+  }
+
+  if (decompose (c, shortest, u))
+  {
     skip_char (buffer);
-  else if (!shortest && c->font->get_glyph (u, 0, &glyph))
+    return;
+  }
+
+  if (!shortest && c->font->get_glyph (u, 0, &glyph))
+  {
     next_char (buffer, glyph);
-  else
-    next_char (buffer, glyph); /* glyph is initialized in earlier branches. */
+    return;
+  }
+
+  if (_hb_glyph_info_is_unicode_space (&buffer->cur()))
+  {
+    hb_codepoint_t space_glyph;
+    hb_unicode_funcs_t::space_t space_type = buffer->unicode->space_fallback_type (u);
+    if (space_type != hb_unicode_funcs_t::NOT_SPACE && c->font->get_glyph (0x0020u, 0, &space_glyph))
+    {
+      _hb_glyph_info_set_unicode_space_fallback_type (&buffer->cur(), space_type);
+      next_char (buffer, space_glyph);
+      buffer->scratch_flags |= HB_BUFFER_SCRATCH_FLAG_HAS_SPACE_FALLBACK;
+      return;
+    }
+  }
+
+  if (u == 0x2011u)
+  {
+    /* U+2011 is the only sensible character that is a no-break version of another character
+     * and not a space.  The space ones are handled already.  Handle this lone one. */
+    hb_codepoint_t other_glyph;
+    if (c->font->get_glyph (0x2010u, 0, &other_glyph))
+    {
+      next_char (buffer, other_glyph);
+      return;
+    }
+  }
+
+  next_char (buffer, glyph); /* glyph is initialized in earlier branches. */
 }
 
 static inline void
 handle_variation_selector_cluster (const hb_ot_shape_normalize_context_t *c, unsigned int end, bool short_circuit)
 {
   /* TODO Currently if there's a variation-selector we give-up, it's just too hard. */
   hb_buffer_t * const buffer = c->buffer;
   hb_font_t * const font = c->font;
-  for (; buffer->idx < end - 1;) {
+  for (; buffer->idx < end - 1 && !buffer->in_error;) {
     if (unlikely (buffer->unicode->is_variation_selector (buffer->cur(+1).codepoint))) {
       /* The next two lines are some ugly lines... But work. */
       if (font->get_glyph (buffer->cur().codepoint, buffer->cur(+1).codepoint, &buffer->cur().glyph_index()))
       {
         buffer->replace_glyphs (2, 1, &buffer->cur().codepoint);
       }
       else
       {
         /* Just pass on the two characters separately, let GSUB do its magic. */
         set_glyph (buffer->cur(), font);
@@ skipping 15 matching lines @@
   if (likely (buffer->idx < end)) {
     set_glyph (buffer->cur(), font);
     buffer->next_glyph ();
   }
 }
 
 static inline void
 decompose_multi_char_cluster (const hb_ot_shape_normalize_context_t *c, unsigned int end, bool short_circuit)
 {
   hb_buffer_t * const buffer = c->buffer;
-  for (unsigned int i = buffer->idx; i < end; i++)
+  for (unsigned int i = buffer->idx; i < end && !buffer->in_error; i++)
     if (unlikely (buffer->unicode->is_variation_selector (buffer->info[i].codepoint))) {
       handle_variation_selector_cluster (c, end, short_circuit);
       return;
     }
 
-  while (buffer->idx < end)
+  while (buffer->idx < end && !buffer->in_error)
     decompose_current_character (c, short_circuit);
 }
 
 static inline void
 decompose_cluster (const hb_ot_shape_normalize_context_t *c, unsigned int end, bool might_short_circuit, bool always_short_circuit)
 {
   if (likely (c->buffer->idx + 1 == end))
     decompose_current_character (c, might_short_circuit);
   else
     decompose_multi_char_cluster (c, end, always_short_circuit);
@@ skipping 39 matching lines @@
    * separate rounds: decompose, reorder, recompose (if desired).  Currently
    * this makes two buffer swaps.  We can make it faster by moving the last
    * two rounds into the inner loop for the first round, but it's more readable
    * this way. */
 
 
   /* First round, decompose */
 
   buffer->clear_output ();
   count = buffer->len;
-  for (buffer->idx = 0; buffer->idx < count;)
+  for (buffer->idx = 0; buffer->idx < count && !buffer->in_error;)
   {
     unsigned int end;
     for (end = buffer->idx + 1; end < count; end++)
       if (likely (!HB_UNICODE_GENERAL_CATEGORY_IS_MARK (_hb_glyph_info_get_general_category (&buffer->info[end]))))
         break;
 
     decompose_cluster (&c, end, might_short_circuit, always_short_circuit);
   }
   buffer->swap_buffers ();
 
@@ skipping 29 matching lines @@
 
   /* Third round, recompose */
 
   /* As noted in the comment earlier, we don't try to combine
    * ccc=0 chars with their previous Starter. */
 
   buffer->clear_output ();
   count = buffer->len;
   unsigned int starter = 0;
   buffer->next_glyph ();
-  while (buffer->idx < count)
+  while (buffer->idx < count && !buffer->in_error)
   {
     hb_codepoint_t composed, glyph;
     if (/* We don't try to compose a non-mark character with it's preceding starter.
          * This is both an optimization to avoid trying to compose every two neighboring
          * glyphs in most scripts AND a desired feature for Hangul.  Apparently Hangul
          * fonts are not designed to mix-and-match pre-composed syllables and Jamo. */
         HB_UNICODE_GENERAL_CATEGORY_IS_MARK (_hb_glyph_info_get_general_category (&buffer->cur())) &&
         /* If there's anything between the starter and this char, they should have CCC
          * smaller than this character's. */
         (starter == buffer->out_len - 1 ||
          _hb_glyph_info_get_modified_combining_class (&buffer->prev()) < _hb_glyph_info_get_modified_combining_class (&buffer->cur())) &&
         /* And compose. */
         c.compose (&c,
                    buffer->out_info[starter].codepoint,
                    buffer->cur().codepoint,
                    &composed) &&
         /* And the font has glyph for the composite. */
         font->get_glyph (composed, 0, &glyph))
     {
       /* Composes. */
       buffer->next_glyph (); /* Copy to out-buffer. */
       if (unlikely (buffer->in_error))
         return;
       buffer->merge_out_clusters (starter, buffer->out_len);
       buffer->out_len--; /* Remove the second composable. */
       /* Modify starter and carry on. */
       buffer->out_info[starter].codepoint = composed;
       buffer->out_info[starter].glyph_index() = glyph;
-      _hb_glyph_info_set_unicode_props (&buffer->out_info[starter], buffer->unicode);
+      _hb_glyph_info_set_unicode_props (&buffer->out_info[starter], buffer);
 
       continue;
     }
 
     /* Blocked, or doesn't compose. */
     buffer->next_glyph ();
 
     if (_hb_glyph_info_get_modified_combining_class (&buffer->prev()) == 0)
       starter = buffer->out_len - 1;
   }
   buffer->swap_buffers ();
 
 }