2 * Copyright © 2011,2012,2014 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
32 #include "hb-open-type.hh"
37 typedef uint8_t codepoint_t
;
39 static const codepoint_t
*
40 next (const codepoint_t
*text
,
41 const codepoint_t
*end
,
42 hb_codepoint_t
*unicode
,
43 hb_codepoint_t replacement
)
45 /* Written to only accept well-formed sequences.
46 * Based on ideas from ICU's U8_NEXT.
47 * Generates one "replacement" for each ill-formed byte. */
49 hb_codepoint_t c
= *text
++;
53 if (hb_in_range
<hb_codepoint_t
> (c
, 0xC2u
, 0xDFu
)) /* Two-byte */
56 if (likely (text
< end
&&
57 (t1
= text
[0] - 0x80u
) <= 0x3Fu
))
59 c
= ((c
&0x1Fu
)<<6) | t1
;
65 else if (hb_in_range
<hb_codepoint_t
> (c
, 0xE0u
, 0xEFu
)) /* Three-byte */
68 if (likely (1 < end
- text
&&
69 (t1
= text
[0] - 0x80u
) <= 0x3Fu
&&
70 (t2
= text
[1] - 0x80u
) <= 0x3Fu
))
72 c
= ((c
&0xFu
)<<12) | (t1
<<6) | t2
;
73 if (unlikely (c
< 0x0800u
|| hb_in_range
<hb_codepoint_t
> (c
, 0xD800u
, 0xDFFFu
)))
80 else if (hb_in_range
<hb_codepoint_t
> (c
, 0xF0u
, 0xF4u
)) /* Four-byte */
82 unsigned int t1
, t2
, t3
;
83 if (likely (2 < end
- text
&&
84 (t1
= text
[0] - 0x80u
) <= 0x3Fu
&&
85 (t2
= text
[1] - 0x80u
) <= 0x3Fu
&&
86 (t3
= text
[2] - 0x80u
) <= 0x3Fu
))
88 c
= ((c
&0x7u
)<<18) | (t1
<<12) | (t2
<<6) | t3
;
89 if (unlikely (!hb_in_range
<hb_codepoint_t
> (c
, 0x10000u
, 0x10FFFFu
)))
104 *unicode
= replacement
;
108 static const codepoint_t
*
109 prev (const codepoint_t
*text
,
110 const codepoint_t
*start
,
111 hb_codepoint_t
*unicode
,
112 hb_codepoint_t replacement
)
114 const codepoint_t
*end
= text
--;
115 while (start
< text
&& (*text
& 0xc0) == 0x80 && end
- text
< 4)
118 if (likely (next (text
, end
, unicode
, replacement
) == end
))
121 *unicode
= replacement
;
126 strlen (const codepoint_t
*text
)
127 { return ::strlen ((const char *) text
); }
130 encode_len (hb_codepoint_t unicode
)
132 if (unicode
< 0x0080u
) return 1;
133 if (unicode
< 0x0800u
) return 2;
134 if (unicode
< 0x10000u
) return 3;
135 if (unicode
< 0x110000u
) return 4;
140 encode (codepoint_t
*text
,
141 const codepoint_t
*end
,
142 hb_codepoint_t unicode
)
144 if (unlikely (unicode
>= 0xD800u
&& (unicode
<= 0xDFFFu
|| unicode
> 0x10FFFFu
)))
146 if (unicode
< 0x0080u
)
148 else if (unicode
< 0x0800u
)
152 *text
++ = 0xC0u
+ (0x1Fu
& (unicode
>> 6));
153 *text
++ = 0x80u
+ (0x3Fu
& (unicode
));
156 else if (unicode
< 0x10000u
)
160 *text
++ = 0xE0u
+ (0x0Fu
& (unicode
>> 12));
161 *text
++ = 0x80u
+ (0x3Fu
& (unicode
>> 6));
162 *text
++ = 0x80u
+ (0x3Fu
& (unicode
));
169 *text
++ = 0xF0u
+ (0x07u
& (unicode
>> 18));
170 *text
++ = 0x80u
+ (0x3Fu
& (unicode
>> 12));
171 *text
++ = 0x80u
+ (0x3Fu
& (unicode
>> 6));
172 *text
++ = 0x80u
+ (0x3Fu
& (unicode
));
180 template <typename TCodepoint
>
183 static_assert (sizeof (TCodepoint
) == 2, "");
184 typedef TCodepoint codepoint_t
;
186 static const codepoint_t
*
187 next (const codepoint_t
*text
,
188 const codepoint_t
*end
,
189 hb_codepoint_t
*unicode
,
190 hb_codepoint_t replacement
)
192 hb_codepoint_t c
= *text
++;
194 if (likely (!hb_in_range
<hb_codepoint_t
> (c
, 0xD800u
, 0xDFFFu
)))
200 if (likely (c
<= 0xDBFFu
&& text
< end
))
202 /* High-surrogate in c */
203 hb_codepoint_t l
= *text
;
204 if (likely (hb_in_range
<hb_codepoint_t
> (l
, 0xDC00u
, 0xDFFFu
)))
206 /* Low-surrogate in l */
207 *unicode
= (c
<< 10) + l
- ((0xD800u
<< 10) - 0x10000u
+ 0xDC00u
);
213 /* Lonely / out-of-order surrogate. */
214 *unicode
= replacement
;
218 static const codepoint_t
*
219 prev (const codepoint_t
*text
,
220 const codepoint_t
*start
,
221 hb_codepoint_t
*unicode
,
222 hb_codepoint_t replacement
)
224 hb_codepoint_t c
= *--text
;
226 if (likely (!hb_in_range
<hb_codepoint_t
> (c
, 0xD800u
, 0xDFFFu
)))
232 if (likely (c
>= 0xDC00u
&& start
< text
))
234 /* Low-surrogate in c */
235 hb_codepoint_t h
= text
[-1];
236 if (likely (hb_in_range
<hb_codepoint_t
> (h
, 0xD800u
, 0xDBFFu
)))
238 /* High-surrogate in h */
239 *unicode
= (h
<< 10) + c
- ((0xD800u
<< 10) - 0x10000u
+ 0xDC00u
);
245 /* Lonely / out-of-order surrogate. */
246 *unicode
= replacement
;
252 strlen (const codepoint_t
*text
)
260 encode_len (hb_codepoint_t unicode
)
262 return unicode
< 0x10000 ? 1 : 2;
266 encode (codepoint_t
*text
,
267 const codepoint_t
*end
,
268 hb_codepoint_t unicode
)
270 if (unlikely (unicode
>= 0xD800u
&& (unicode
<= 0xDFFFu
|| unicode
> 0x10FFFFu
)))
272 if (unicode
< 0x10000u
)
274 else if (end
- text
>= 2)
277 *text
++ = 0xD800u
+ (unicode
>> 10);
278 *text
++ = 0xDC00u
+ (unicode
& 0x03FFu
);
284 typedef hb_utf16_xe_t
<uint16_t> hb_utf16_t
;
285 typedef hb_utf16_xe_t
<OT::HBUINT16
> hb_utf16_be_t
;
288 template <typename TCodepoint
, bool validate
=true>
291 static_assert (sizeof (TCodepoint
) == 4, "");
292 typedef TCodepoint codepoint_t
;
294 static const TCodepoint
*
295 next (const TCodepoint
*text
,
296 const TCodepoint
*end HB_UNUSED
,
297 hb_codepoint_t
*unicode
,
298 hb_codepoint_t replacement
)
300 hb_codepoint_t c
= *unicode
= *text
++;
301 if (validate
&& unlikely (c
>= 0xD800u
&& (c
<= 0xDFFFu
|| c
> 0x10FFFFu
)))
302 *unicode
= replacement
;
306 static const TCodepoint
*
307 prev (const TCodepoint
*text
,
308 const TCodepoint
*start HB_UNUSED
,
309 hb_codepoint_t
*unicode
,
310 hb_codepoint_t replacement
)
312 hb_codepoint_t c
= *unicode
= *--text
;
313 if (validate
&& unlikely (c
>= 0xD800u
&& (c
<= 0xDFFFu
|| c
> 0x10FFFFu
)))
314 *unicode
= replacement
;
319 strlen (const TCodepoint
*text
)
327 encode_len (hb_codepoint_t unicode HB_UNUSED
)
333 encode (codepoint_t
*text
,
334 const codepoint_t
*end HB_UNUSED
,
335 hb_codepoint_t unicode
)
337 if (validate
&& unlikely (unicode
>= 0xD800u
&& (unicode
<= 0xDFFFu
|| unicode
> 0x10FFFFu
)))
344 typedef hb_utf32_xe_t
<uint32_t> hb_utf32_t
;
345 typedef hb_utf32_xe_t
<uint32_t, false> hb_utf32_novalidate_t
;
350 typedef uint8_t codepoint_t
;
352 static const codepoint_t
*
353 next (const codepoint_t
*text
,
354 const codepoint_t
*end HB_UNUSED
,
355 hb_codepoint_t
*unicode
,
356 hb_codepoint_t replacement HB_UNUSED
)
362 static const codepoint_t
*
363 prev (const codepoint_t
*text
,
364 const codepoint_t
*start HB_UNUSED
,
365 hb_codepoint_t
*unicode
,
366 hb_codepoint_t replacement HB_UNUSED
)
373 strlen (const codepoint_t
*text
)
381 encode_len (hb_codepoint_t unicode HB_UNUSED
)
387 encode (codepoint_t
*text
,
388 const codepoint_t
*end HB_UNUSED
,
389 hb_codepoint_t unicode
)
391 if (unlikely (unicode
>= 0x0100u
))
401 typedef uint8_t codepoint_t
;
403 static const codepoint_t
*
404 next (const codepoint_t
*text
,
405 const codepoint_t
*end HB_UNUSED
,
406 hb_codepoint_t
*unicode
,
407 hb_codepoint_t replacement HB_UNUSED
)
410 if (*unicode
>= 0x0080u
)
411 *unicode
= replacement
;
415 static const codepoint_t
*
416 prev (const codepoint_t
*text
,
417 const codepoint_t
*start HB_UNUSED
,
418 hb_codepoint_t
*unicode
,
419 hb_codepoint_t replacement
)
422 if (*unicode
>= 0x0080u
)
423 *unicode
= replacement
;
428 strlen (const codepoint_t
*text
)
436 encode_len (hb_codepoint_t unicode HB_UNUSED
)
442 encode (codepoint_t
*text
,
443 const codepoint_t
*end HB_UNUSED
,
444 hb_codepoint_t unicode
)
446 if (unlikely (unicode
>= 0x0080u
))
453 #endif /* HB_UTF_HH */