summaryrefslogtreecommitdiffstats
path: root/src/corelib/text/qlocale_tools.cpp
blob: 9adb6e4a234519109ea61aa3c601776f5481f120 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
/****************************************************************************
**
** Copyright (C) 2016 The Qt Company Ltd.
** Copyright (C) 2016 Intel Corporation.
** Contact: https://www.qt.io/licensing/
**
** This file is part of the QtCore module of the Qt Toolkit.
**
** $QT_BEGIN_LICENSE:LGPL$
** Commercial License Usage
** Licensees holding valid commercial Qt licenses may use this file in
** accordance with the commercial license agreement provided with the
** Software or, alternatively, in accordance with the terms contained in
** a written agreement between you and The Qt Company. For licensing terms
** and conditions see https://www.qt.io/terms-conditions. For further
** information use the contact form at https://www.qt.io/contact-us.
**
** GNU Lesser General Public License Usage
** Alternatively, this file may be used under the terms of the GNU Lesser
** General Public License version 3 as published by the Free Software
** Foundation and appearing in the file LICENSE.LGPL3 included in the
** packaging of this file. Please review the following information to
** ensure the GNU Lesser General Public License version 3 requirements
** will be met: https://www.gnu.org/licenses/lgpl-3.0.html.
**
** GNU General Public License Usage
** Alternatively, this file may be used under the terms of the GNU
** General Public License version 2.0 or (at your option) the GNU General
** Public license version 3 or any later version approved by the KDE Free
** Qt Foundation. The licenses are as published by the Free Software
** Foundation and appearing in the file LICENSE.GPL2 and LICENSE.GPL3
** included in the packaging of this file. Please review the following
** information to ensure the GNU General Public License requirements will
** be met: https://www.gnu.org/licenses/gpl-2.0.html and
** https://www.gnu.org/licenses/gpl-3.0.html.
**
** $QT_END_LICENSE$
**
****************************************************************************/

#include "qlocale_tools_p.h"
#include "qdoublescanprint_p.h"
#include "qlocale_p.h"
#include "qstring.h"

#include <private/qnumeric_p.h>

#include <ctype.h>
#include <errno.h>
#include <float.h>
#include <limits.h>
#include <math.h>
#include <stdlib.h>
#include <time.h>

#if defined(Q_OS_LINUX) && !defined(__UCLIBC__)
#    include <fenv.h>
#endif

// Sizes as defined by the ISO C99 standard - fallback
#ifndef LLONG_MAX
#   define LLONG_MAX Q_INT64_C(0x7fffffffffffffff)
#endif
#ifndef LLONG_MIN
#   define LLONG_MIN (-LLONG_MAX - Q_INT64_C(1))
#endif
#ifndef ULLONG_MAX
#   define ULLONG_MAX Q_UINT64_C(0xffffffffffffffff)
#endif

QT_BEGIN_NAMESPACE

#include "../../3rdparty/freebsd/strtoull.c"
#include "../../3rdparty/freebsd/strtoll.c"

QT_CLOCALE_HOLDER

void qt_doubleToAscii(double d, QLocaleData::DoubleForm form, int precision, char *buf, int bufSize,
                      bool &sign, int &length, int &decpt)
{
    if (bufSize == 0) {
        decpt = 0;
        sign = d < 0;
        length = 0;
        return;
    }

    // Detect special numbers (nan, +/-inf)
    // We cannot use the high-level API of libdouble-conversion as we need to apply locale-specific
    // formatting, such as decimal points, thousands-separators, etc. Because of this, we have to
    // check for infinity and NaN before calling DoubleToAscii.
    if (qt_is_inf(d)) {
        sign = d < 0;
        if (bufSize >= 3) {
            buf[0] = 'i';
            buf[1] = 'n';
            buf[2] = 'f';
            length = 3;
        } else {
            length = 0;
        }
        return;
    } else if (qt_is_nan(d)) {
        if (bufSize >= 3) {
            buf[0] = 'n';
            buf[1] = 'a';
            buf[2] = 'n';
            length = 3;
        } else {
            length = 0;
        }
        return;
    }

    if (form == QLocaleData::DFSignificantDigits && precision == 0)
        precision = 1; // 0 significant digits is silently converted to 1

#if !defined(QT_NO_DOUBLECONVERSION) && !defined(QT_BOOTSTRAPPED)
    // one digit before the decimal dot, counts as significant digit for DoubleToStringConverter
    if (form == QLocaleData::DFExponent && precision >= 0)
        ++precision;

    double_conversion::DoubleToStringConverter::DtoaMode mode;
    if (precision == QLocale::FloatingPointShortest) {
        mode = double_conversion::DoubleToStringConverter::SHORTEST;
    } else if (form == QLocaleData::DFSignificantDigits || form == QLocaleData::DFExponent) {
        mode = double_conversion::DoubleToStringConverter::PRECISION;
    } else {
        mode = double_conversion::DoubleToStringConverter::FIXED;
    }
    double_conversion::DoubleToStringConverter::DoubleToAscii(d, mode, precision, buf, bufSize,
                                                              &sign, &length, &decpt);
#else // QT_NO_DOUBLECONVERSION || QT_BOOTSTRAPPED

    // Cut the precision at 999, to fit it into the format string. We can't get more than 17
    // significant digits, so anything after that is mostly noise. You do get closer to the "middle"
    // of the range covered by the given double with more digits, so to a degree it does make sense
    // to honor higher precisions. We define that at more than 999 digits that is not the case.
    if (precision > 999)
        precision = 999;
    else if (precision == QLocale::FloatingPointShortest)
        precision = QLocaleData::DoubleMaxSignificant; // "shortest" mode not supported by snprintf

    if (isZero(d)) {
        // Negative zero is expected as simple "0", not "-0". We cannot do d < 0, though.
        sign = false;
        buf[0] = '0';
        length = 1;
        decpt = 1;
        return;
    } else if (d < 0) {
        sign = true;
        d = -d;
    } else {
        sign = false;
    }

    const int formatLength = 7; // '%', '.', 3 digits precision, 'f', '\0'
    char format[formatLength];
    format[formatLength - 1] = '\0';
    format[0] = '%';
    format[1] = '.';
    format[2] = char((precision / 100) % 10) + '0';
    format[3] = char((precision / 10) % 10)  + '0';
    format[4] = char(precision % 10)  + '0';
    int extraChars;
    switch (form) {
    case QLocaleData::DFDecimal:
        format[formatLength - 2] = 'f';
        // <anything> '.' <precision> '\0' - optimize for numbers smaller than 512k
        extraChars = (d > (1 << 19) ? QLocaleData::DoubleMaxDigitsBeforeDecimal : 6) + 2;
        break;
    case QLocaleData::DFExponent:
        format[formatLength - 2] = 'e';
        // '.', 'e', '-', <exponent> '\0'
        extraChars = 7;
        break;
    case QLocaleData::DFSignificantDigits:
        format[formatLength - 2] = 'g';

        // either the same as in the 'e' case, or '.' and '\0'
        // precision covers part before '.'
        extraChars = 7;
        break;
    default:
        Q_UNREACHABLE();
    }

    QVarLengthArray<char> target(precision + extraChars);

    length = qDoubleSnprintf(target.data(), target.size(), QT_CLOCALE, format, d);
    int firstSignificant = 0;
    int decptInTarget = length;

    // Find the first significant digit (not 0), and note any '.' we encounter.
    // There is no '-' at the front of target because we made sure d > 0 above.
    while (firstSignificant < length) {
        if (target[firstSignificant] == '.')
            decptInTarget = firstSignificant;
        else if (target[firstSignificant] != '0')
            break;
        ++firstSignificant;
    }

    // If no '.' found so far, search the rest of the target buffer for it.
    if (decptInTarget == length)
        decptInTarget = std::find(target.data() + firstSignificant, target.data() + length, '.') -
                target.data();

    int eSign = length;
    if (form != QLocaleData::DFDecimal) {
        // In 'e' or 'g' form, look for the 'e'.
        eSign = std::find(target.data() + firstSignificant, target.data() + length, 'e') -
                target.data();

        if (eSign < length) {
            // If 'e' is found, the final decimal point is determined by the number after 'e'.
            // Mind that the final decimal point, decpt, is the offset of the decimal point from the
            // start of the resulting string in buf. It may be negative or larger than bufSize, in
            // which case the missing digits are zeroes. In the 'e' case decptInTarget is always 1,
            // as variants of snprintf always generate numbers with one digit before the '.' then.
            // This is why the final decimal point is offset by 1, relative to the number after 'e'.
            bool ok;
            const char *endptr;
            decpt = qstrtoll(target.data() + eSign + 1, &endptr, 10, &ok) + 1;
            Q_ASSERT(ok);
            Q_ASSERT(endptr - target.data() <= length);
        } else {
            // No 'e' found, so it's the 'f' form. Variants of snprintf generate numbers with
            // potentially multiple digits before the '.', but without decimal exponent then. So we
            // get the final decimal point from the position of the '.'. The '.' itself takes up one
            // character. We adjust by 1 below if that gets in the way.
            decpt = decptInTarget - firstSignificant;
        }
    } else {
        // In 'f' form, there can not be an 'e', so it's enough to look for the '.'
        // (and possibly adjust by 1 below)
        decpt = decptInTarget - firstSignificant;
    }

    // Move the actual digits from the snprintf target to the actual buffer.
    if (decptInTarget > firstSignificant) {
        // First move the digits before the '.', if any
        int lengthBeforeDecpt = decptInTarget - firstSignificant;
        memcpy(buf, target.data() + firstSignificant, qMin(lengthBeforeDecpt, bufSize));
        if (eSign > decptInTarget && lengthBeforeDecpt < bufSize) {
            // Then move any remaining digits, until 'e'
            memcpy(buf + lengthBeforeDecpt, target.data() + decptInTarget + 1,
                   qMin(eSign - decptInTarget - 1, bufSize - lengthBeforeDecpt));
            // The final length of the output is the distance between the first significant digit
            // and 'e' minus 1, for the '.', except if the buffer is smaller.
            length = qMin(eSign - firstSignificant - 1, bufSize);
        } else {
            // 'e' was before the decpt or things didn't fit. Don't subtract the '.' from the length.
            length = qMin(eSign - firstSignificant, bufSize);
        }
    } else {
        if (eSign > firstSignificant) {
            // If there are any significant digits at all, they are all after the '.' now.
            // Just copy them straight away.
            memcpy(buf, target.data() + firstSignificant, qMin(eSign - firstSignificant, bufSize));

            // The decimal point was before the first significant digit, so we were one off above.
            // Consider 0.1 - buf will be just '1', and decpt should be 0. But
            // "decptInTarget - firstSignificant" will yield -1.
            ++decpt;
            length = qMin(eSign - firstSignificant, bufSize);
        } else {
            // No significant digits means the number is just 0.
            buf[0] = '0';
            length = 1;
            decpt = 1;
        }
    }
#endif // QT_NO_DOUBLECONVERSION || QT_BOOTSTRAPPED
    while (length > 1 && buf[length - 1] == '0') // drop trailing zeroes
        --length;
}

double qt_asciiToDouble(const char *num, int numLen, bool &ok, int &processed,
                        StrayCharacterMode strayCharMode)
{
    if (*num == '\0') {
        ok = false;
        processed = 0;
        return 0.0;
    }

    ok = true;

    // We have to catch NaN before because we need NaN as marker for "garbage" in the
    // libdouble-conversion case and, in contrast to libdouble-conversion or sscanf, we don't allow
    // "-nan" or "+nan"
    if (qstrcmp(num, "nan") == 0) {
        processed = 3;
        return qt_qnan();
    } else if ((num[0] == '-' || num[0] == '+') && qstrcmp(num + 1, "nan") == 0) {
        processed = 0;
        ok = false;
        return 0.0;
    }

    // Infinity values are implementation defined in the sscanf case. In the libdouble-conversion
    // case we need infinity as overflow marker.
    if (qstrcmp(num, "+inf") == 0) {
        processed = 4;
        return qt_inf();
    } else if (qstrcmp(num, "inf") == 0) {
        processed = 3;
        return qt_inf();
    } else if (qstrcmp(num, "-inf") == 0) {
        processed = 4;
        return -qt_inf();
    }

    double d = 0.0;
#if !defined(QT_NO_DOUBLECONVERSION) && !defined(QT_BOOTSTRAPPED)
    int conv_flags = double_conversion::StringToDoubleConverter::NO_FLAGS;
    if (strayCharMode == TrailingJunkAllowed) {
        conv_flags = double_conversion::StringToDoubleConverter::ALLOW_TRAILING_JUNK;
    } else if (strayCharMode == WhitespacesAllowed) {
        conv_flags = double_conversion::StringToDoubleConverter::ALLOW_LEADING_SPACES
                | double_conversion::StringToDoubleConverter::ALLOW_TRAILING_SPACES;
    }
    double_conversion::StringToDoubleConverter conv(conv_flags, 0.0, qt_qnan(), nullptr, nullptr);
    d = conv.StringToDouble(num, numLen, &processed);

    if (!qIsFinite(d)) {
        ok = false;
        if (qIsNaN(d)) {
            // Garbage found. We don't accept it and return 0.
            processed = 0;
            return 0.0;
        } else {
            // Overflow. That's not OK, but we still return infinity.
            return d;
        }
    }
#else
    if (qDoubleSscanf(num, QT_CLOCALE, "%lf%n", &d, &processed) < 1)
        processed = 0;

    if ((strayCharMode == TrailingJunkProhibited && processed != numLen) || qIsNaN(d)) {
        // Implementation defined nan symbol or garbage found. We don't accept it.
        processed = 0;
        ok = false;
        return 0.0;
    }

    if (!qIsFinite(d)) {
        // Overflow. Check for implementation-defined infinity symbols and reject them.
        // We assume that any infinity symbol has to contain a character that cannot be part of a
        // "normal" number (that is 0-9, ., -, +, e).
        ok = false;
        for (int i = 0; i < processed; ++i) {
            char c = num[i];
            if ((c < '0' || c > '9') && c != '.' && c != '-' && c != '+' && c != 'e' && c != 'E') {
                // Garbage found
                processed = 0;
                return 0.0;
            }
        }
        return d;
    }
#endif // !defined(QT_NO_DOUBLECONVERSION) && !defined(QT_BOOTSTRAPPED)

    // Otherwise we would have gotten NaN or sorted it out above.
    Q_ASSERT(strayCharMode == TrailingJunkAllowed || processed == numLen);

    // Check if underflow has occurred.
    if (isZero(d)) {
        for (int i = 0; i < processed; ++i) {
            if (num[i] >= '1' && num[i] <= '9') {
                // if a digit before any 'e' is not 0, then a non-zero number was intended.
                ok = false;
                return 0.0;
            } else if (num[i] == 'e' || num[i] == 'E') {
                break;
            }
        }
    }
    return d;
}

unsigned long long
qstrtoull(const char * nptr, const char **endptr, int base, bool *ok)
{
    // strtoull accepts negative numbers. We don't.
    // Use a different variable so we pass the original nptr to strtoul
    // (we need that so endptr may be nptr in case of failure)
    const char *begin = nptr;
    while (ascii_isspace(*begin))
        ++begin;
    if (*begin == '-') {
        *ok = false;
        return 0;
    }

    *ok = true;
    errno = 0;
    char *endptr2 = nullptr;
    unsigned long long result = qt_strtoull(nptr, &endptr2, base);
    if (endptr)
        *endptr = endptr2;
    if ((result == 0 || result == std::numeric_limits<unsigned long long>::max())
            && (errno || endptr2 == nptr)) {
        *ok = false;
        return 0;
    }
    return result;
}

long long
qstrtoll(const char * nptr, const char **endptr, int base, bool *ok)
{
    *ok = true;
    errno = 0;
    char *endptr2 = nullptr;
    long long result = qt_strtoll(nptr, &endptr2, base);
    if (endptr)
        *endptr = endptr2;
    if ((result == 0 || result == std::numeric_limits<long long>::min()
         || result == std::numeric_limits<long long>::max())
            && (errno || nptr == endptr2)) {
        *ok = false;
        return 0;
    }
    return result;
}

QString qulltoa(qulonglong number, int base, const QStringView zero)
{
    // Length of MAX_ULLONG in base 2 is 64; and we may need a surrogate pair
    // per digit. We do not need a terminator.
    const unsigned maxlen = 128;
    Q_STATIC_ASSERT(CHAR_BIT * sizeof(number) <= maxlen);
    ushort buff[maxlen];
    ushort *const end = buff + maxlen, *p = end;

    if (base != 10 || zero == u"0") {
        while (number != 0) {
            int c = number % base;
            *--p = c < 10 ? '0' + c : c - 10 + 'a';
            number /= base;
        }
    } else if (zero.size() && !zero.at(0).isSurrogate()) {
        const ushort zeroUcs4 = zero.at(0).unicode();
        while (number != 0) {
            *(--p) = zeroUcs4 + number % base;

            number /= base;
        }
    } else if (zero.size() == 2 && zero.at(0).isHighSurrogate()) {
        const uint zeroUcs4 = QChar::surrogateToUcs4(zero.at(0), zero.at(1));
        while (number != 0) {
            const uint digit = zeroUcs4 + number % base;

            *(--p) = QChar::lowSurrogate(digit);
            *(--p) = QChar::highSurrogate(digit);

            number /= base;
        }
    } else {
        return QString();
    }

    return QString(reinterpret_cast<QChar *>(p), end - p);
}

QString &decimalForm(const QString &zero, const QString &decimal, const QString &group,
                     QString &digits, int decpt, int precision,
                     PrecisionMode pm,
                     bool always_show_decpt,
                     bool thousands_group)
{
    const auto digitWidth = zero.size();
    Q_ASSERT(digitWidth == 1 || digitWidth == 2);
    Q_ASSERT(digits.size() % digitWidth == 0);

    if (decpt < 0) {
        for (int i = 0; i < -decpt; ++i)
            digits.prepend(zero);
        decpt = 0;
    } else {
        for (int i = digits.length() / digitWidth; i < decpt; ++i)
            digits.append(zero);
    }

    switch (pm) {
    case PMDecimalDigits:
        for (int i = digits.length() / digitWidth - decpt; i < precision; ++i)
            digits.append(zero);
        break;
    case  PMSignificantDigits:
        for (int i = digits.length() / digitWidth; i < precision; ++i)
            digits.append(zero);
        break;
    case PMChopTrailingZeros:
        break;
    }

    if (always_show_decpt || decpt < digits.length() / digitWidth)
        digits.insert(decpt * digitWidth, decimal);

    // FIXME: they're not simply thousands separators !
    // Need to mirror IndianNumberGrouping code in QLocaleData::longLongToString()
    if (thousands_group) {
        for (int i = decpt - 3; i > 0; i -= 3)
            digits.insert(i * digitWidth, group);
    }

    if (decpt == 0)
        digits.prepend(zero);

    return digits;
}

QString &exponentForm(const QString &zero, const QString &decimal, const QString &exponential,
                      const QString &group, const QString &plus, const QString &minus,
                      QString &digits, int decpt, int precision,
                      PrecisionMode pm,
                      bool always_show_decpt,
                      bool leading_zero_in_exponent)
{
    const auto digitWidth = zero.size();
    Q_ASSERT(digitWidth == 1 || digitWidth == 2);
    Q_ASSERT(digits.size() % digitWidth == 0);

    switch (pm) {
    case PMDecimalDigits:
        for (int i = digits.length() / digitWidth; i < precision + 1; ++i)
            digits.append(zero);
        break;
    case PMSignificantDigits:
        for (int i = digits.length() / digitWidth; i < precision; ++i)
            digits.append(zero);
        break;
    case PMChopTrailingZeros:
        break;
    }

    if (always_show_decpt || digits.length() > digitWidth)
        digits.insert(digitWidth, decimal);

    digits.append(exponential);
    digits.append(QLocaleData::longLongToString(zero, group, plus, minus, decpt - 1,
                                                leading_zero_in_exponent ? 2 : 1,
                                                10, -1, QLocaleData::AlwaysShowSign));

    return digits;
}

double qstrtod(const char *s00, const char **se, bool *ok)
{
    const int len = static_cast<int>(strlen(s00));
    Q_ASSERT(len >= 0);
    return qstrntod(s00, len, se, ok);
}

/*!
  \internal

  Converts the initial portion of the string pointed to by \a s00 to a double, using the 'C' locale.
 */
double qstrntod(const char *s00, int len, const char **se, bool *ok)
{
    int processed = 0;
    bool nonNullOk = false;
    double d = qt_asciiToDouble(s00, len, nonNullOk, processed, TrailingJunkAllowed);
    if (se)
        *se = s00 + processed;
    if (ok)
        *ok = nonNullOk;
    return d;
}

QString qdtoa(qreal d, int *decpt, int *sign)
{
    bool nonNullSign = false;
    int nonNullDecpt = 0;
    int length = 0;

    // Some versions of libdouble-conversion like an extra digit, probably for '\0'
    char result[QLocaleData::DoubleMaxSignificant + 1];
    qt_doubleToAscii(d, QLocaleData::DFSignificantDigits, QLocale::FloatingPointShortest, result,
                     QLocaleData::DoubleMaxSignificant + 1, nonNullSign, length, nonNullDecpt);

    if (sign)
        *sign = nonNullSign ? 1 : 0;
    if (decpt)
        *decpt = nonNullDecpt;

    return QLatin1String(result, length);
}

QT_END_NAMESPACE