3 """Generator of the function to prohibit certain vowel sequences.
5 It creates ``_hb_preprocess_text_vowel_constraints``, which inserts dotted
6 circles into sequences prohibited by the USE script development spec.
7 This function should be used as the ``preprocess_text`` of an
8 ``hb_ot_complex_shaper_t``.
11 from __future__ import absolute_import, division, print_function, unicode_literals
15 from HTMLParser import HTMLParser
17 print (s.encode ('utf-8'), end='')
19 from html.parser import HTMLParser
22 sys.stdout.buffer.write (s.encode ('utf-8'))
27 if len (sys.argv) != 3:
28 print ('usage: ./gen-vowel-constraints.py HBIndicVowelConstraints.txt Scripts.txt', file=sys.stderr)
31 with io.open (sys.argv[2], encoding='utf-8') as f:
32 scripts_header = [f.readline () for i in range (2)]
39 fields = [x.strip () for x in line.split (';')]
42 uu = fields[0].split ('..')
43 start = int (uu[0], 16)
49 for u in range (start, end + 1):
51 if script not in script_order:
52 script_order[script] = start
54 class ConstraintSet (object):
55 """A set of prohibited code point sequences.
58 constraint (List[int]): A prohibited code point sequence.
61 def __init__ (self, constraint):
62 # Either a list or a dictionary. As a list of code points, it
63 # represents a prohibited code point sequence. As a dictionary,
64 # it represents a set of prohibited sequences, where each item
65 # represents the set of prohibited sequences starting with the
66 # key (a code point) concatenated with any of the values
70 def add (self, constraint):
71 """Add a constraint to this set."""
76 if isinstance (self._c, list):
77 if constraint == self._c[:len (constraint)]:
79 elif self._c != constraint[:len (self._c)]:
80 self._c = {self._c[0]: ConstraintSet (self._c[1:])}
81 if isinstance (self._c, dict):
83 self._c[first].add (rest)
85 self._c[first] = ConstraintSet (rest)
87 def _indent (self, depth):
88 return (' ' * depth).replace (' ', '\t')
90 def __str__ (self, index=0, depth=4):
92 indent = self._indent (depth)
93 if isinstance (self._c, list):
94 if len (self._c) == 0:
95 s.append ('{}matched = true;\n'.format (indent))
96 elif len (self._c) == 1:
97 s.append ('{}matched = 0x{:04X}u == buffer->cur ({}).codepoint;\n'.format (indent, next (iter (self._c)), index or ''))
99 s.append ('{}if (0x{:04X}u == buffer->cur ({}).codepoint &&\n'.format (indent, self._c[0], index))
100 s.append ('{}buffer->idx + {} < count &&\n'.format (self._indent (depth + 2), len (self._c)))
101 for i, cp in enumerate (self._c[1:], start=1):
102 s.append ('{}0x{:04X}u == buffer->cur ({}).codepoint{}\n'.format (
103 self._indent (depth + 2), cp, index + i, ')' if i == len (self._c) - 1 else ' &&'))
104 s.append ('{}{{\n'.format (indent))
105 for i in range (len (self._c)):
106 s.append ('{}buffer->next_glyph ();\n'.format (self._indent (depth + 1)))
107 s.append ('{}_output_dotted_circle (buffer);\n'.format (self._indent (depth + 1)))
108 s.append ('{}}}\n'.format (indent))
110 s.append ('{}switch (buffer->cur ({}).codepoint)\n'.format(indent, index or ''))
111 s.append ('{}{{\n'.format (indent))
112 cases = collections.defaultdict (set)
113 for first, rest in sorted (self._c.items ()):
114 cases[rest.__str__ (index + 1, depth + 2)].add (first)
115 for body, labels in sorted (cases.items (), key=lambda b_ls: sorted (b_ls[1])[0]):
116 for i, cp in enumerate (sorted (labels)):
118 s.append (self._indent (depth + 1))
121 s.append ('case 0x{:04X}u:{}'.format (cp, '\n' if i % 4 == 3 else ''))
122 if len (labels) % 4 != 0:
125 s.append ('{}break;\n'.format (self._indent (depth + 2)))
126 s.append ('{}}}\n'.format (indent))
130 with io.open (sys.argv[1], encoding='utf-8') as f:
131 constraints_header = [f.readline ().strip () for i in range (2)]
136 constraint = [int (cp, 16) for cp in line.split (';')[0].split ()]
137 if not constraint: continue
138 assert 2 <= len (constraint), 'Prohibited sequence is too short: {}'.format (constraint)
139 script = scripts[constraint[0]]
140 if script in constraints:
141 constraints[script].add (constraint)
143 constraints[script] = ConstraintSet (constraint)
144 assert constraints, 'No constraints found'
146 print ('/* == Start of generated functions == */')
148 print (' * The following functions are generated by running:')
150 print (' * %s use Scripts.txt' % sys.argv[0])
152 print (' * on files with these headers:')
154 for line in constraints_header:
155 print (' * %s' % line.strip ())
157 for line in scripts_header:
158 print (' * %s' % line.strip ())
161 print ('#include "hb-ot-shape-complex-vowel-constraints.hh"')
163 print ('static void')
164 print ('_output_dotted_circle (hb_buffer_t *buffer)')
166 print (' hb_glyph_info_t &dottedcircle = buffer->output_glyph (0x25CCu);')
167 print (' _hb_glyph_info_reset_continuation (&dottedcircle);')
170 print ('static void')
171 print ('_output_with_dotted_circle (hb_buffer_t *buffer)')
173 print (' _output_dotted_circle (buffer);')
174 print (' buffer->next_glyph ();')
179 print ('_hb_preprocess_text_vowel_constraints (const hb_ot_shape_plan_t *plan HB_UNUSED,')
180 print ('\t\t\t\t hb_buffer_t *buffer,')
181 print ('\t\t\t\t hb_font_t *font HB_UNUSED)')
183 print (' if (buffer->flags & HB_BUFFER_FLAG_DO_NOT_INSERT_DOTTED_CIRCLE)')
186 print (' /* UGLY UGLY UGLY business of adding dotted-circle in the middle of')
187 print (' * vowel-sequences that look like another vowel. Data for each script')
188 print (' * collected from the USE script development spec.')
190 print (' * https://github.com/harfbuzz/harfbuzz/issues/1019')
192 print (' bool processed = false;')
193 print (' buffer->clear_output ();')
194 print (' unsigned int count = buffer->len;')
195 print (' switch ((unsigned) buffer->props.script)')
198 for script, constraints in sorted (constraints.items (), key=lambda s_c: script_order[s_c[0]]):
199 print (' case HB_SCRIPT_{}:'.format (script.upper ()))
200 print (' for (buffer->idx = 0; buffer->idx + 1 < count && buffer->successful;)')
202 print ('\tbool matched = false;')
203 write (str (constraints))
204 print ('\tbuffer->next_glyph ();')
205 print ('\tif (matched) _output_with_dotted_circle (buffer);')
207 print (' processed = true;')
214 print (' if (processed)')
216 print (' if (buffer->idx < count)')
217 print (' buffer->next_glyph ();')
218 print (' buffer->swap_buffers ();')
223 print ('/* == End of generated functions == */')