GCC Code Coverage Report
Directory: ../ Exec Total Coverage
File: /home/iojs/build/workspace/node-test-commit-linux-coverage-daily/nodes/benchmark/out/../src/node_i18n.cc Lines: 318 356 89.3 %
Date: 2019-09-25 22:36:03 Branches: 139 208 66.8 %

Line Branch Exec Source
1
// Copyright Joyent, Inc. and other Node contributors.
2
//
3
// Permission is hereby granted, free of charge, to any person obtaining a
4
// copy of this software and associated documentation files (the
5
// "Software"), to deal in the Software without restriction, including
6
// without limitation the rights to use, copy, modify, merge, publish,
7
// distribute, sublicense, and/or sell copies of the Software, and to permit
8
// persons to whom the Software is furnished to do so, subject to the
9
// following conditions:
10
//
11
// The above copyright notice and this permission notice shall be included
12
// in all copies or substantial portions of the Software.
13
//
14
// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
15
// OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
16
// MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN
17
// NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM,
18
// DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
19
// OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
20
// USE OR OTHER DEALINGS IN THE SOFTWARE.
21
22
/*
23
 * notes: by srl295
24
 *  - When in NODE_HAVE_SMALL_ICU mode, ICU is linked against "stub" (null) data
25
 *     ( stubdata/libicudata.a ) containing nothing, no data, and it's also
26
 *    linked against a "small" data file which the SMALL_ICUDATA_ENTRY_POINT
27
 *    macro names. That's the "english+root" data.
28
 *
29
 *    If icu_data_path is non-null, the user has provided a path and we assume
30
 *    it goes somewhere useful. We set that path in ICU, and exit.
31
 *    If icu_data_path is null, they haven't set a path and we want the
32
 *    "english+root" data.  We call
33
 *       udata_setCommonData(SMALL_ICUDATA_ENTRY_POINT,...)
34
 *    to load up the english+root data.
35
 *
36
 *  - when NOT in NODE_HAVE_SMALL_ICU mode, ICU is linked directly with its full
37
 *    data. All of the variables and command line options for changing data at
38
 *    runtime are disabled, as they wouldn't fully override the internal data.
39
 *    See:  http://bugs.icu-project.org/trac/ticket/10924
40
 */
41
42
43
#include "node_i18n.h"
44
45
#if defined(NODE_HAVE_I18N_SUPPORT)
46
47
#include "base_object-inl.h"
48
#include "node.h"
49
#include "node_buffer.h"
50
#include "node_errors.h"
51
#include "node_internals.h"
52
#include "util-inl.h"
53
#include "v8.h"
54
55
#include <unicode/utypes.h>
56
#include <unicode/putil.h>
57
#include <unicode/uchar.h>
58
#include <unicode/uclean.h>
59
#include <unicode/udata.h>
60
#include <unicode/uidna.h>
61
#include <unicode/ucnv.h>
62
#include <unicode/utf8.h>
63
#include <unicode/utf16.h>
64
#include <unicode/timezone.h>
65
#include <unicode/ulocdata.h>
66
#include <unicode/uvernum.h>
67
#include <unicode/uversion.h>
68
#include <unicode/ustring.h>
69
70
#ifdef NODE_HAVE_SMALL_ICU
71
/* if this is defined, we have a 'secondary' entry point.
72
   compare following to utypes.h defs for U_ICUDATA_ENTRY_POINT */
73
#define SMALL_ICUDATA_ENTRY_POINT \
74
  SMALL_DEF2(U_ICU_VERSION_MAJOR_NUM, U_LIB_SUFFIX_C_NAME)
75
#define SMALL_DEF2(major, suff) SMALL_DEF(major, suff)
76
#ifndef U_LIB_SUFFIX_C_NAME
77
#define SMALL_DEF(major, suff) icusmdt##major##_dat
78
#else
79
#define SMALL_DEF(major, suff) icusmdt##suff##major##_dat
80
#endif
81
82
extern "C" const char U_DATA_API SMALL_ICUDATA_ENTRY_POINT[];
83
#endif
84
85
namespace node {
86
87
using v8::Context;
88
using v8::FunctionCallbackInfo;
89
using v8::HandleScope;
90
using v8::Int32;
91
using v8::Isolate;
92
using v8::Local;
93
using v8::MaybeLocal;
94
using v8::NewStringType;
95
using v8::Object;
96
using v8::ObjectTemplate;
97
using v8::String;
98
using v8::Value;
99
100
namespace i18n {
101
namespace {
102
103
template <typename T>
104
534
MaybeLocal<Object> ToBufferEndian(Environment* env, MaybeStackBuffer<T>* buf) {
105
534
  MaybeLocal<Object> ret = Buffer::New(env, buf);
106

534
  if (ret.IsEmpty())
107
    return ret;
108
109
  static_assert(sizeof(T) == 1 || sizeof(T) == 2,
110
                "Currently only one- or two-byte buffers are supported");
111

534
  if (sizeof(T) > 1 && IsBigEndian()) {
112
    SPREAD_BUFFER_ARG(ret.ToLocalChecked(), retbuf);
113
    SwapBytes16(retbuf_data, retbuf_length);
114
  }
115
116
534
  return ret;
117
}
118
119
struct Converter {
120
8
  explicit Converter(const char* name, const char* sub = nullptr)
121
8
      : conv(nullptr) {
122
8
    UErrorCode status = U_ZERO_ERROR;
123
8
    conv = ucnv_open(name, &status);
124
8
    CHECK(U_SUCCESS(status));
125
8
    if (sub != nullptr) {
126
2
      ucnv_setSubstChars(conv, sub, strlen(sub), &status);
127
    }
128
8
  }
129
130
205
  explicit Converter(UConverter* converter,
131
205
                     const char* sub = nullptr) : conv(converter) {
132
205
    CHECK_NOT_NULL(conv);
133
205
    UErrorCode status = U_ZERO_ERROR;
134
205
    if (sub != nullptr) {
135
      ucnv_setSubstChars(conv, sub, strlen(sub), &status);
136
    }
137
205
  }
138
139
213
  ~Converter() {
140
213
    ucnv_close(conv);
141
213
  }
142
143
  UConverter* conv;
144
};
145
146
class ConverterObject : public BaseObject, Converter {
147
 public:
148
  enum ConverterFlags {
149
    CONVERTER_FLAGS_FLUSH      = 0x1,
150
    CONVERTER_FLAGS_FATAL      = 0x2,
151
    CONVERTER_FLAGS_IGNORE_BOM = 0x4
152
  };
153
154
410
  ~ConverterObject() override = default;
155
156
2
  static void Has(const FunctionCallbackInfo<Value>& args) {
157
2
    Environment* env = Environment::GetCurrent(args);
158
2
    HandleScope scope(env->isolate());
159
160
2
    CHECK_GE(args.Length(), 1);
161
4
    Utf8Value label(env->isolate(), args[0]);
162
163
2
    UErrorCode status = U_ZERO_ERROR;
164
2
    UConverter* conv = ucnv_open(*label, &status);
165
6
    args.GetReturnValue().Set(!!U_SUCCESS(status));
166
4
    ucnv_close(conv);
167
2
  }
168
169
205
  static void Create(const FunctionCallbackInfo<Value>& args) {
170
205
    Environment* env = Environment::GetCurrent(args);
171
205
    HandleScope scope(env->isolate());
172
173
205
    Local<ObjectTemplate> t = ObjectTemplate::New(env->isolate());
174
205
    t->SetInternalFieldCount(1);
175
    Local<Object> obj;
176
615
    if (!t->NewInstance(env->context()).ToLocal(&obj)) return;
177
178
205
    CHECK_GE(args.Length(), 2);
179
410
    Utf8Value label(env->isolate(), args[0]);
180
820
    int flags = args[1]->Uint32Value(env->context()).ToChecked();
181
    bool fatal =
182
205
        (flags & CONVERTER_FLAGS_FATAL) == CONVERTER_FLAGS_FATAL;
183
    bool ignoreBOM =
184
205
        (flags & CONVERTER_FLAGS_IGNORE_BOM) == CONVERTER_FLAGS_IGNORE_BOM;
185
186
205
    UErrorCode status = U_ZERO_ERROR;
187
205
    UConverter* conv = ucnv_open(*label, &status);
188
205
    if (U_FAILURE(status))
189
      return;
190
191
205
    if (fatal) {
192
94
      status = U_ZERO_ERROR;
193
      ucnv_setToUCallBack(conv, UCNV_TO_U_CALLBACK_STOP,
194
94
                          nullptr, nullptr, nullptr, &status);
195
    }
196
197
205
    new ConverterObject(env, obj, conv, ignoreBOM);
198
615
    args.GetReturnValue().Set(obj);
199
  }
200
201
616
  static void Decode(const FunctionCallbackInfo<Value>& args) {
202
616
    Environment* env = Environment::GetCurrent(args);
203
204
616
    CHECK_GE(args.Length(), 3);  // Converter, Buffer, Flags
205
206
    ConverterObject* converter;
207
1756
    ASSIGN_OR_RETURN_UNWRAP(&converter, args[0].As<Object>());
208
616
    ArrayBufferViewContents<char> input(args[1]);
209
2464
    int flags = args[2]->Uint32Value(env->context()).ToChecked();
210
211
616
    UErrorCode status = U_ZERO_ERROR;
212
616
    MaybeStackBuffer<UChar> result;
213
    MaybeLocal<Object> ret;
214
616
    size_t limit = ucnv_getMinCharSize(converter->conv) * input.length();
215
616
    if (limit > 0)
216
590
      result.AllocateSufficientStorage(limit);
217
218
616
    UBool flush = (flags & CONVERTER_FLAGS_FLUSH) == CONVERTER_FLAGS_FLUSH;
219
616
    OnScopeLeave cleanup([&]() {
220
616
      if (flush) {
221
        // Reset the converter state.
222
198
        converter->bomSeen_ = false;
223
198
        ucnv_reset(converter->conv);
224
      }
225
1324
    });
226
227
616
    const char* source = input.data();
228
616
    size_t source_length = input.length();
229
230

616
    if (converter->unicode_ && !converter->ignoreBOM_ && !converter->bomSeen_) {
231
187
      int32_t bomOffset = 0;
232
187
      ucnv_detectUnicodeSignature(source, source_length, &bomOffset, &status);
233
187
      source += bomOffset;
234
187
      source_length -= bomOffset;
235
187
      converter->bomSeen_ = true;
236
    }
237
238
616
    UChar* target = *result;
239
    ucnv_toUnicode(converter->conv,
240
616
                   &target, target + (limit * sizeof(UChar)),
241
                   &source, source + source_length,
242
1232
                   nullptr, flush, &status);
243
244
616
    if (U_SUCCESS(status)) {
245
524
      if (limit > 0)
246
498
        result.SetLength(target - &result[0]);
247
524
      ret = ToBufferEndian(env, &result);
248
1048
      args.GetReturnValue().Set(ret.ToLocalChecked());
249
524
      return;
250
    }
251
252
368
    args.GetReturnValue().Set(status);
253
  }
254
255
  SET_NO_MEMORY_INFO()
256
  SET_MEMORY_INFO_NAME(ConverterObject)
257
  SET_SELF_SIZE(ConverterObject)
258
259
 protected:
260
205
  ConverterObject(Environment* env,
261
                  Local<Object> wrap,
262
                  UConverter* converter,
263
                  bool ignoreBOM,
264
                  const char* sub = nullptr) :
265
                  BaseObject(env, wrap),
266
                  Converter(converter, sub),
267
205
                  ignoreBOM_(ignoreBOM) {
268
205
    MakeWeak();
269
270
205
    switch (ucnv_getType(converter)) {
271
      case UCNV_UTF8:
272
      case UCNV_UTF16_BigEndian:
273
      case UCNV_UTF16_LittleEndian:
274
205
        unicode_ = true;
275
205
        break;
276
      default:
277
        unicode_ = false;
278
    }
279
205
  }
280
281
 private:
282
  bool unicode_ = false;     // True if this is a Unicode converter
283
  bool ignoreBOM_ = false;   // True if the BOM should be ignored on Unicode
284
  bool bomSeen_ = false;     // True if the BOM has been seen
285
};
286
287
// One-Shot Converters
288
289
2
void CopySourceBuffer(MaybeStackBuffer<UChar>* dest,
290
                      const char* data,
291
                      const size_t length,
292
                      const size_t length_in_chars) {
293
2
  dest->AllocateSufficientStorage(length_in_chars);
294
2
  char* dst = reinterpret_cast<char*>(**dest);
295
2
  memcpy(dst, data, length);
296
2
  if (IsBigEndian()) {
297
    SwapBytes16(dst, length);
298
  }
299
2
}
300
301
typedef MaybeLocal<Object> (*TranscodeFunc)(Environment* env,
302
                                            const char* fromEncoding,
303
                                            const char* toEncoding,
304
                                            const char* source,
305
                                            const size_t source_length,
306
                                            UErrorCode* status);
307
308
2
MaybeLocal<Object> Transcode(Environment* env,
309
                             const char* fromEncoding,
310
                             const char* toEncoding,
311
                             const char* source,
312
                             const size_t source_length,
313
                             UErrorCode* status) {
314
2
  *status = U_ZERO_ERROR;
315
  MaybeLocal<Object> ret;
316
2
  MaybeStackBuffer<char> result;
317
4
  Converter to(toEncoding, "?");
318
4
  Converter from(fromEncoding);
319
2
  const uint32_t limit = source_length * ucnv_getMaxCharSize(to.conv);
320
2
  result.AllocateSufficientStorage(limit);
321
2
  char* target = *result;
322
  ucnv_convertEx(to.conv, from.conv, &target, target + limit,
323
                 &source, source + source_length, nullptr, nullptr,
324
2
                 nullptr, nullptr, true, true, status);
325
2
  if (U_SUCCESS(*status)) {
326
2
    result.SetLength(target - &result[0]);
327
2
    ret = ToBufferEndian(env, &result);
328
  }
329
4
  return ret;
330
}
331
332
4
MaybeLocal<Object> TranscodeToUcs2(Environment* env,
333
                                   const char* fromEncoding,
334
                                   const char* toEncoding,
335
                                   const char* source,
336
                                   const size_t source_length,
337
                                   UErrorCode* status) {
338
4
  *status = U_ZERO_ERROR;
339
  MaybeLocal<Object> ret;
340
  MaybeStackBuffer<UChar> destbuf(source_length);
341
8
  Converter from(fromEncoding);
342
4
  const size_t length_in_chars = source_length * sizeof(UChar);
343
  ucnv_toUChars(from.conv, *destbuf, length_in_chars,
344
4
                source, source_length, status);
345
4
  if (U_SUCCESS(*status))
346
4
    ret = ToBufferEndian(env, &destbuf);
347
8
  return ret;
348
}
349
350
MaybeLocal<Object> TranscodeFromUcs2(Environment* env,
351
                                     const char* fromEncoding,
352
                                     const char* toEncoding,
353
                                     const char* source,
354
                                     const size_t source_length,
355
                                     UErrorCode* status) {
356
  *status = U_ZERO_ERROR;
357
  MaybeStackBuffer<UChar> sourcebuf;
358
  MaybeLocal<Object> ret;
359
  Converter to(toEncoding, "?");
360
  const size_t length_in_chars = source_length / sizeof(UChar);
361
  CopySourceBuffer(&sourcebuf, source, source_length, length_in_chars);
362
  MaybeStackBuffer<char> destbuf(length_in_chars);
363
  const uint32_t len = ucnv_fromUChars(to.conv, *destbuf, length_in_chars,
364
                                       *sourcebuf, length_in_chars, status);
365
  if (U_SUCCESS(*status)) {
366
    destbuf.SetLength(len);
367
    ret = ToBufferEndian(env, &destbuf);
368
  }
369
  return ret;
370
}
371
372
2
MaybeLocal<Object> TranscodeUcs2FromUtf8(Environment* env,
373
                                         const char* fromEncoding,
374
                                         const char* toEncoding,
375
                                         const char* source,
376
                                         const size_t source_length,
377
                                         UErrorCode* status) {
378
2
  *status = U_ZERO_ERROR;
379
2
  MaybeStackBuffer<UChar> destbuf;
380
  int32_t result_length;
381
2
  u_strFromUTF8(*destbuf, destbuf.capacity(), &result_length,
382
4
                source, source_length, status);
383
  MaybeLocal<Object> ret;
384
2
  if (U_SUCCESS(*status)) {
385
1
    destbuf.SetLength(result_length);
386
1
    ret = ToBufferEndian(env, &destbuf);
387
1
  } else if (*status == U_BUFFER_OVERFLOW_ERROR) {
388
1
    *status = U_ZERO_ERROR;
389
1
    destbuf.AllocateSufficientStorage(result_length);
390
    u_strFromUTF8(*destbuf, result_length, &result_length,
391
1
                  source, source_length, status);
392
1
    if (U_SUCCESS(*status)) {
393
1
      destbuf.SetLength(result_length);
394
1
      ret = ToBufferEndian(env, &destbuf);
395
    }
396
  }
397
2
  return ret;
398
}
399
400
2
MaybeLocal<Object> TranscodeUtf8FromUcs2(Environment* env,
401
                                         const char* fromEncoding,
402
                                         const char* toEncoding,
403
                                         const char* source,
404
                                         const size_t source_length,
405
                                         UErrorCode* status) {
406
2
  *status = U_ZERO_ERROR;
407
  MaybeLocal<Object> ret;
408
2
  const size_t length_in_chars = source_length / sizeof(UChar);
409
  int32_t result_length;
410
2
  MaybeStackBuffer<UChar> sourcebuf;
411
4
  MaybeStackBuffer<char> destbuf;
412
2
  CopySourceBuffer(&sourcebuf, source, source_length, length_in_chars);
413
2
  u_strToUTF8(*destbuf, destbuf.capacity(), &result_length,
414
4
              *sourcebuf, length_in_chars, status);
415
2
  if (U_SUCCESS(*status)) {
416
1
    destbuf.SetLength(result_length);
417
1
    ret = ToBufferEndian(env, &destbuf);
418
1
  } else if (*status == U_BUFFER_OVERFLOW_ERROR) {
419
1
    *status = U_ZERO_ERROR;
420
1
    destbuf.AllocateSufficientStorage(result_length);
421
1
    u_strToUTF8(*destbuf, result_length, &result_length, *sourcebuf,
422
2
                length_in_chars, status);
423
1
    if (U_SUCCESS(*status)) {
424
1
      destbuf.SetLength(result_length);
425
1
      ret = ToBufferEndian(env, &destbuf);
426
    }
427
  }
428
4
  return ret;
429
}
430
431
20
const char* EncodingName(const enum encoding encoding) {
432

20
  switch (encoding) {
433
2
    case ASCII: return "us-ascii";
434
4
    case LATIN1: return "iso8859-1";
435
8
    case UCS2: return "utf16le";
436
6
    case UTF8: return "utf-8";
437
    default: return nullptr;
438
  }
439
}
440
441
22
bool SupportedEncoding(const enum encoding encoding) {
442
22
  switch (encoding) {
443
    case ASCII:
444
    case LATIN1:
445
    case UCS2:
446
20
    case UTF8: return true;
447
2
    default: return false;
448
  }
449
}
450
451
12
void Transcode(const FunctionCallbackInfo<Value>&args) {
452
12
  Environment* env = Environment::GetCurrent(args);
453
12
  Isolate* isolate = env->isolate();
454
12
  UErrorCode status = U_ZERO_ERROR;
455
  MaybeLocal<Object> result;
456
457
12
  ArrayBufferViewContents<char> input(args[0]);
458
12
  const enum encoding fromEncoding = ParseEncoding(isolate, args[1], BUFFER);
459
12
  const enum encoding toEncoding = ParseEncoding(isolate, args[2], BUFFER);
460
461

12
  if (SupportedEncoding(fromEncoding) && SupportedEncoding(toEncoding)) {
462
10
    TranscodeFunc tfn = &Transcode;
463

10
    switch (fromEncoding) {
464
      case ASCII:
465
      case LATIN1:
466
4
        if (toEncoding == UCS2)
467
4
          tfn = &TranscodeToUcs2;
468
4
        break;
469
      case UTF8:
470
4
        if (toEncoding == UCS2)
471
2
          tfn = &TranscodeUcs2FromUtf8;
472
4
        break;
473
      case UCS2:
474
2
        switch (toEncoding) {
475
          case UCS2:
476
            tfn = &Transcode;
477
            break;
478
          case UTF8:
479
2
            tfn = &TranscodeUtf8FromUcs2;
480
2
            break;
481
          default:
482
            tfn = &TranscodeFromUcs2;
483
        }
484
2
        break;
485
      default:
486
        // This should not happen because of the SupportedEncoding checks
487
        ABORT();
488
    }
489
490
    result = tfn(env, EncodingName(fromEncoding), EncodingName(toEncoding),
491
10
                 input.data(), input.length(), &status);
492
  } else {
493
2
    status = U_ILLEGAL_ARGUMENT_ERROR;
494
  }
495
496
12
  if (result.IsEmpty())
497
6
    return args.GetReturnValue().Set(status);
498
499
20
  return args.GetReturnValue().Set(result.ToLocalChecked());
500
}
501
502
2
void ICUErrorName(const FunctionCallbackInfo<Value>& args) {
503
2
  Environment* env = Environment::GetCurrent(args);
504
4
  CHECK(args[0]->IsInt32());
505
6
  UErrorCode status = static_cast<UErrorCode>(args[0].As<Int32>()->Value());
506
  args.GetReturnValue().Set(
507
      String::NewFromUtf8(env->isolate(),
508
                          u_errorName(status),
509
6
                          NewStringType::kNormal).ToLocalChecked());
510
2
}
511
512
}  // anonymous namespace
513
514
4958
bool InitializeICUDirectory(const std::string& path) {
515
4958
  UErrorCode status = U_ZERO_ERROR;
516
4958
  if (path.empty()) {
517
#ifdef NODE_HAVE_SMALL_ICU
518
    // install the 'small' data.
519
4956
    udata_setCommonData(&SMALL_ICUDATA_ENTRY_POINT, &status);
520
#else  // !NODE_HAVE_SMALL_ICU
521
    // no small data, so nothing to do.
522
#endif  // !NODE_HAVE_SMALL_ICU
523
  } else {
524
2
    u_setDataDirectory(path.c_str());
525
2
    u_init(&status);
526
  }
527
4958
  return status == U_ZERO_ERROR;
528
}
529
530
382
int32_t ToUnicode(MaybeStackBuffer<char>* buf,
531
                  const char* input,
532
                  size_t length) {
533
382
  UErrorCode status = U_ZERO_ERROR;
534
382
  uint32_t options = UIDNA_NONTRANSITIONAL_TO_UNICODE;
535
382
  UIDNA* uidna = uidna_openUTS46(options, &status);
536
382
  if (U_FAILURE(status))
537
    return -1;
538
382
  UIDNAInfo info = UIDNA_INFO_INITIALIZER;
539
540
  int32_t len = uidna_nameToUnicodeUTF8(uidna,
541
                                        input, length,
542
382
                                        **buf, buf->capacity(),
543
                                        &info,
544
382
                                        &status);
545
546
  // Do not check info.errors like we do with ToASCII since ToUnicode always
547
  // returns a string, despite any possible errors that may have occurred.
548
549
382
  if (status == U_BUFFER_OVERFLOW_ERROR) {
550
    status = U_ZERO_ERROR;
551
    buf->AllocateSufficientStorage(len);
552
    len = uidna_nameToUnicodeUTF8(uidna,
553
                                  input, length,
554
                                  **buf, buf->capacity(),
555
                                  &info,
556
                                  &status);
557
  }
558
559
  // info.errors is ignored as UTS #46 ToUnicode always produces a Unicode
560
  // string, regardless of whether an error occurred.
561
562
382
  if (U_FAILURE(status)) {
563
    len = -1;
564
    buf->SetLength(0);
565
  } else {
566
382
    buf->SetLength(len);
567
  }
568
569
382
  uidna_close(uidna);
570
382
  return len;
571
}
572
573
13777
int32_t ToASCII(MaybeStackBuffer<char>* buf,
574
                const char* input,
575
                size_t length,
576
                enum idna_mode mode) {
577
13777
  UErrorCode status = U_ZERO_ERROR;
578
  uint32_t options =                  // CheckHyphens = false; handled later
579
    UIDNA_CHECK_BIDI |                // CheckBidi = true
580
    UIDNA_CHECK_CONTEXTJ |            // CheckJoiners = true
581
13777
    UIDNA_NONTRANSITIONAL_TO_ASCII;   // Nontransitional_Processing
582
13777
  if (mode == IDNA_STRICT) {
583
    options |= UIDNA_USE_STD3_RULES;  // UseSTD3ASCIIRules = beStrict
584
                                      // VerifyDnsLength = beStrict;
585
                                      //   handled later
586
  }
587
588
13777
  UIDNA* uidna = uidna_openUTS46(options, &status);
589
13777
  if (U_FAILURE(status))
590
    return -1;
591
13777
  UIDNAInfo info = UIDNA_INFO_INITIALIZER;
592
593
  int32_t len = uidna_nameToASCII_UTF8(uidna,
594
                                       input, length,
595
13777
                                       **buf, buf->capacity(),
596
                                       &info,
597
13777
                                       &status);
598
599
13777
  if (status == U_BUFFER_OVERFLOW_ERROR) {
600
2
    status = U_ZERO_ERROR;
601
2
    buf->AllocateSufficientStorage(len);
602
    len = uidna_nameToASCII_UTF8(uidna,
603
                                 input, length,
604
2
                                 **buf, buf->capacity(),
605
                                 &info,
606
2
                                 &status);
607
  }
608
609
  // In UTS #46 which specifies ToASCII, certain error conditions are
610
  // configurable through options, and the WHATWG URL Standard promptly elects
611
  // to disable some of them to accommodate for real-world use cases.
612
  // Unfortunately, ICU4C's IDNA module does not support disabling some of
613
  // these options through `options` above, and thus continues throwing
614
  // unnecessary errors. To counter this situation, we just filter out the
615
  // errors that may have happened afterwards, before deciding whether to
616
  // return an error from this function.
617
618
  // CheckHyphens = false
619
  // (Specified in the current UTS #46 draft rev. 18.)
620
  // Refs:
621
  // - https://github.com/whatwg/url/issues/53
622
  // - https://github.com/whatwg/url/pull/309
623
  // - http://www.unicode.org/review/pri317/
624
  // - http://www.unicode.org/reports/tr46/tr46-18.html
625
  // - https://www.icann.org/news/announcement-2000-01-07-en
626
13777
  info.errors &= ~UIDNA_ERROR_HYPHEN_3_4;
627
13777
  info.errors &= ~UIDNA_ERROR_LEADING_HYPHEN;
628
13777
  info.errors &= ~UIDNA_ERROR_TRAILING_HYPHEN;
629
630
13777
  if (mode != IDNA_STRICT) {
631
    // VerifyDnsLength = beStrict
632
13777
    info.errors &= ~UIDNA_ERROR_EMPTY_LABEL;
633
13777
    info.errors &= ~UIDNA_ERROR_LABEL_TOO_LONG;
634
13777
    info.errors &= ~UIDNA_ERROR_DOMAIN_NAME_TOO_LONG;
635
  }
636
637


13777
  if (U_FAILURE(status) || (mode != IDNA_LENIENT && info.errors != 0)) {
638
78
    len = -1;
639
78
    buf->SetLength(0);
640
  } else {
641
13699
    buf->SetLength(len);
642
  }
643
644
13777
  uidna_close(uidna);
645
13777
  return len;
646
}
647
648
189
static void ToUnicode(const FunctionCallbackInfo<Value>& args) {
649
189
  Environment* env = Environment::GetCurrent(args);
650
189
  CHECK_GE(args.Length(), 1);
651
567
  CHECK(args[0]->IsString());
652
189
  Utf8Value val(env->isolate(), args[0]);
653
654
378
  MaybeStackBuffer<char> buf;
655
189
  int32_t len = ToUnicode(&buf, *val, val.length());
656
657
189
  if (len < 0) {
658
189
    return THROW_ERR_INVALID_ARG_VALUE(env, "Cannot convert name to Unicode");
659
  }
660
661
  args.GetReturnValue().Set(
662
      String::NewFromUtf8(env->isolate(),
663
189
                          *buf,
664
                          NewStringType::kNormal,
665
756
                          len).ToLocalChecked());
666
}
667
668
11096
static void ToASCII(const FunctionCallbackInfo<Value>& args) {
669
11096
  Environment* env = Environment::GetCurrent(args);
670
11096
  CHECK_GE(args.Length(), 1);
671
33288
  CHECK(args[0]->IsString());
672
11096
  Utf8Value val(env->isolate(), args[0]);
673
  // optional arg
674
33288
  bool lenient = args[1]->BooleanValue(env->isolate());
675
11096
  enum idna_mode mode = lenient ? IDNA_LENIENT : IDNA_DEFAULT;
676
677
22183
  MaybeStackBuffer<char> buf;
678
11096
  int32_t len = ToASCII(&buf, *val, val.length(), mode);
679
680
11096
  if (len < 0) {
681
11105
    return THROW_ERR_INVALID_ARG_VALUE(env, "Cannot convert name to ASCII");
682
  }
683
684
  args.GetReturnValue().Set(
685
      String::NewFromUtf8(env->isolate(),
686
11087
                          *buf,
687
                          NewStringType::kNormal,
688
44348
                          len).ToLocalChecked());
689
}
690
691
// This is similar to wcwidth except that it takes the current unicode
692
// character properties database into consideration, allowing it to
693
// correctly calculate the column widths of things like emoji's and
694
// newer wide characters. wcwidth, on the other hand, uses a fixed
695
// algorithm that does not take things like emoji into proper
696
// consideration.
697
//
698
// TODO(TimothyGu): Investigate Cc (C0/C1 control codes). Both VTE (used by
699
// GNOME Terminal) and Konsole don't consider them to be zero-width (see refs
700
// below), and when printed in VTE it is Narrow. However GNOME Terminal doesn't
701
// allow it to be input. Linux's PTY terminal prints control characters as
702
// Narrow rhombi.
703
//
704
// TODO(TimothyGu): Investigate Hangul jamo characters. Medial vowels and final
705
// consonants are 0-width when combined with initial consonants; otherwise they
706
// are technically Wide. But many terminals (including Konsole and
707
// VTE/GLib-based) implement all medials and finals as 0-width.
708
//
709
// Refs: https://eev.ee/blog/2015/09/12/dark-corners-of-unicode/#combining-characters-and-character-width
710
// Refs: https://github.com/GNOME/glib/blob/79e4d4c6be/glib/guniprop.c#L388-L420
711
// Refs: https://github.com/KDE/konsole/blob/8c6a5d13c0/src/konsole_wcwidth.cpp#L101-L223
712
2134516
static int GetColumnWidth(UChar32 codepoint,
713
                          bool ambiguous_as_full_width = false) {
714
  const auto zero_width_mask = U_GC_CC_MASK |  // C0/C1 control code
715
                               U_GC_CF_MASK |  // Format control character
716
                               U_GC_ME_MASK |  // Enclosing mark
717
2134516
                               U_GC_MN_MASK;   // Nonspacing mark
718

4269536
  if (codepoint != 0x00AD &&  // SOFT HYPHEN is Cf but not zero-width
719
4268502
      ((U_MASK(u_charType(codepoint)) & zero_width_mask) ||
720
2133995
       u_hasBinaryProperty(codepoint, UCHAR_EMOJI_MODIFIER))) {
721
513
    return 0;
722
  }
723
724
  // UCHAR_EAST_ASIAN_WIDTH is the Unicode property that identifies a
725
  // codepoint as being full width, wide, ambiguous, neutral, narrow,
726
  // or halfwidth.
727
2134003
  const int eaw = u_getIntPropertyValue(codepoint, UCHAR_EAST_ASIAN_WIDTH);
728

2134003
  switch (eaw) {
729
    case U_EA_FULLWIDTH:
730
    case U_EA_WIDE:
731
642
      return 2;
732
    case U_EA_AMBIGUOUS:
733
      // See: http://www.unicode.org/reports/tr11/#Ambiguous for details
734
356
      if (ambiguous_as_full_width) {
735
177
        return 2;
736
      }
737
      // If ambiguous_as_full_width is false:
738
      // Fall through
739
    case U_EA_NEUTRAL:
740
604
      if (u_hasBinaryProperty(codepoint, UCHAR_EMOJI_PRESENTATION)) {
741
        return 2;
742
      }
743
      // Fall through
744
    case U_EA_HALFWIDTH:
745
    case U_EA_NARROW:
746
    default:
747
2133184
      return 1;
748
  }
749
}
750
751
// Returns the column width for the given String.
752
2134207
static void GetStringWidth(const FunctionCallbackInfo<Value>& args) {
753
2134207
  Environment* env = Environment::GetCurrent(args);
754
2134207
  if (args.Length() < 1)
755
2133147
    return;
756
757
4268414
  bool ambiguous_as_full_width = args[1]->IsTrue();
758
4268414
  bool expand_emoji_sequence = args[2]->IsTrue();
759
760
4268414
  if (args[0]->IsNumber()) {
761
    uint32_t val;
762
8532588
    if (!args[0]->Uint32Value(env->context()).To(&val)) return;
763
6399441
    args.GetReturnValue().Set(GetColumnWidth(val, ambiguous_as_full_width));
764
2133147
    return;
765
  }
766
767
1060
  TwoByteValue value(env->isolate(), args[0]);
768
  // reinterpret_cast is required by windows to compile
769
1060
  UChar* str = reinterpret_cast<UChar*>(*value);
770
  static_assert(sizeof(*str) == sizeof(**value),
771
                "sizeof(*str) == sizeof(**value)");
772
1060
  UChar32 c = 0;
773
  UChar32 p;
774
1060
  size_t n = 0;
775
1060
  uint32_t width = 0;
776
777
3492
  while (n < value.length()) {
778
1372
    p = c;
779


1372
    U16_NEXT(str, n, value.length(), c);
780
    // Don't count individual emoji codepoints that occur within an
781
    // emoji sequence. This is not necessarily foolproof. Some
782
    // environments display emoji sequences in the appropriate
783
    // condensed form (as a single emoji glyph), other environments
784
    // may not understand an emoji sequence and will display each
785
    // individual emoji separately. When this happens, the width
786
    // calculated will be off, and there's no reliable way of knowing
787
    // in advance if a particular sequence is going to be supported.
788
    // The expand_emoji_sequence option allows the caller to skip this
789
    // check and count each code within an emoji sequence separately.
790

4109
    if (!expand_emoji_sequence &&
791

2743
        n > 0 && p == 0x200d &&  // 0x200d == ZWJ (zero width joiner)
792
3
        (u_hasBinaryProperty(c, UCHAR_EMOJI_PRESENTATION) ||
793
         u_hasBinaryProperty(c, UCHAR_EMOJI_MODIFIER))) {
794
3
      continue;
795
    }
796
1369
    width += GetColumnWidth(c, ambiguous_as_full_width);
797
  }
798
2120
  args.GetReturnValue().Set(width);
799
}
800
801
5156
void Initialize(Local<Object> target,
802
                Local<Value> unused,
803
                Local<Context> context,
804
                void* priv) {
805
5156
  Environment* env = Environment::GetCurrent(context);
806
5156
  env->SetMethod(target, "toUnicode", ToUnicode);
807
5156
  env->SetMethod(target, "toASCII", ToASCII);
808
5156
  env->SetMethod(target, "getStringWidth", GetStringWidth);
809
810
  // One-shot converters
811
5156
  env->SetMethod(target, "icuErrName", ICUErrorName);
812
5156
  env->SetMethod(target, "transcode", Transcode);
813
814
  // ConverterObject
815
5156
  env->SetMethod(target, "getConverter", ConverterObject::Create);
816
5156
  env->SetMethod(target, "decode", ConverterObject::Decode);
817
5156
  env->SetMethod(target, "hasConverter", ConverterObject::Has);
818
5156
}
819
820
}  // namespace i18n
821
}  // namespace node
822
823
5010
NODE_MODULE_CONTEXT_AWARE_INTERNAL(icu, node::i18n::Initialize)
824
825
#endif  // NODE_HAVE_I18N_SUPPORT