diff options
Diffstat (limited to 'src/corelib/io/qurl.cpp')
| -rw-r--r-- | src/corelib/io/qurl.cpp | 318 |
1 files changed, 217 insertions, 101 deletions
diff --git a/src/corelib/io/qurl.cpp b/src/corelib/io/qurl.cpp index 180e9ec04..79cd2f03e 100644 --- a/src/corelib/io/qurl.cpp +++ b/src/corelib/io/qurl.cpp @@ -2336,12 +2336,12 @@ static const NameprepCaseFoldingEntry NameprepCaseFolding[] = { { 0x1D7BB, { 0x03C3, 0x0000, 0x0000, 0x0000 } } }; -static void mapToLowerCase(QString *str) +static void mapToLowerCase(QString *str, int from) { int N = sizeof(NameprepCaseFolding) / sizeof(NameprepCaseFolding[0]); QChar *d = 0; - for (int i = 0; i < str->size(); ++i) { + for (int i = from; i < str->size(); ++i) { int uc = str->at(i).unicode(); if (uc < 0x80) { if (uc <= 'Z' && uc >= 'A') { @@ -2388,11 +2388,11 @@ static bool isMappedToNothing(const QChar &ch) } -static void stripProhibitedOutput(QString *str) +static void stripProhibitedOutput(QString *str, int from) { - ushort *out = (ushort *)str->data(); + ushort *out = (ushort *)str->data() + from; const ushort *in = out; - const ushort *end = out + str->size(); + const ushort *end = (ushort *)str->data() + str->size(); while (in < end) { ushort uc = *in; if (uc < 0x80 || @@ -2901,66 +2901,99 @@ static bool isBidirectionalL(const QChar &ch) return false; } +#ifdef QT_BUILD_INTERNAL +// export for tst_qurl.cpp +Q_AUTOTEST_EXPORT void qt_nameprep(QString *source, int from); +Q_AUTOTEST_EXPORT bool qt_check_std3rules(const QChar *uc, int len); +#else +// non-test build, keep the symbols for ourselves +static void qt_nameprep(QString *source, int from); +static bool qt_check_std3rules(const QChar *uc, int len); +#endif -Q_AUTOTEST_EXPORT QString qt_nameprep(const QString &source) +void qt_nameprep(QString *source, int from) { - QString mapped = source; - - bool simple = true; - for (int i = 0; i < mapped.size(); ++i) { - ushort uc = mapped.at(i).unicode(); + QChar *src = source->data(); // causes a detach, so we're sure the only one using it + QChar *out = src + from; + const QChar *e = src + source->size(); + + for ( ; out < e; ++out) { + register ushort uc = out->unicode(); if (uc > 0x80) { - simple = false; break; } else if (uc >= 'A' && uc <= 'Z') { - mapped[i] = QChar(uc | 0x20); + *out = QChar(uc | 0x20); } } - if (simple) - return mapped; - + if (out == e) + return; // everything was mapped easily (lowercased, actually) + int firstNonAscii = out - src; + // Characters commonly mapped to nothing are simply removed // (Table B.1) - QChar *out = mapped.data(); const QChar *in = out; - const QChar *e = in + mapped.size(); while (in < e) { if (!isMappedToNothing(*in)) *out++ = *in; ++in; } if (out != in) - mapped.truncate(out - mapped.constData()); + source->truncate(out - src); // Map to lowercase (Table B.2) - mapToLowerCase(&mapped); + mapToLowerCase(source, firstNonAscii); // Normalize to Unicode 3.2 form KC - mapped = mapped.normalized(QString::NormalizationForm_KC, QChar::Unicode_3_2); + extern void qt_string_normalize(QString *data, QString::NormalizationForm mode, + QChar::UnicodeVersion version, int from); + qt_string_normalize(source, QString::NormalizationForm_KC, QChar::Unicode_3_2, firstNonAscii); // Strip prohibited output - stripProhibitedOutput(&mapped); + stripProhibitedOutput(source, firstNonAscii); // Check for valid bidirectional characters bool containsLCat = false; bool containsRandALCat = false; - for (int j = 0; j < mapped.size() && (!containsLCat || !containsRandALCat); ++j) { - if (isBidirectionalL(mapped.at(j))) + src = source->data(); + e = src + source->size(); + for (in = src + from; in < e && (!containsLCat || !containsRandALCat); ++in) { + if (isBidirectionalL(*in)) containsLCat = true; - else if (isBidirectionalRorAL(mapped.at(j))) + else if (isBidirectionalRorAL(*in)) containsRandALCat = true; } if (containsRandALCat) { - if (containsLCat || (!isBidirectionalRorAL(mapped.at(0)) - || !isBidirectionalRorAL(mapped.at(mapped.size() - 1)))) - mapped.clear(); + if (containsLCat || (!isBidirectionalRorAL(src[from]) + || !isBidirectionalRorAL(e[-1]))) + source->resize(from); // not allowed, clear the label } +} + +bool qt_check_std3rules(const QChar *uc, int len) +{ + if (len > 63) + return false; - return mapped; + for (int i = 0; i < len; ++i) { + register ushort c = uc[i].unicode(); + if (c == '-' && (i == 0 || i == len - 1)) + return false; + + // verifying the absence of LDH is the same as verifying that + // only LDH is present + if (c == '-' || (c >= '0' && c <= '9') + || (c >= 'A' && c <= 'Z') + || (c >= 'a' && c <= 'z')) + continue; + + return false; + } + + return true; } -static inline char encodeDigit(uint digit) +static inline uint encodeDigit(uint digit) { return digit + 22 + 75 * (digit < 26); } @@ -2977,7 +3010,7 @@ static inline uint adapt(uint delta, uint numpoints, bool firsttime) return k + (((base - tmin + 1) * delta) / (delta + skew)); } -static inline void appendEncode(QByteArray* output, uint& delta, uint& bias, uint& b, uint& h) +static inline void appendEncode(QString* output, uint& delta, uint& bias, uint& b, uint& h) { uint qq; uint k; @@ -2991,17 +3024,17 @@ static inline void appendEncode(QByteArray* output, uint& delta, uint& bias, uin t = (k <= bias) ? tmin : (k >= bias + tmax) ? tmax : k - bias; if (qq < t) break; - *output += encodeDigit(t + (qq - t) % (base - t)); + *output += QChar(encodeDigit(t + (qq - t) % (base - t))); qq = (qq - t) / (base - t); } - *output += encodeDigit(qq); + *output += QChar(encodeDigit(qq)); bias = adapt(delta, h + 1, h == b); delta = 0; ++h; } -static void toPunycodeHelper(const QChar *s, int ucLength, QByteArray *output) +static void toPunycodeHelper(const QChar *s, int ucLength, QString *output) { uint n = initial_n; uint delta = 0; @@ -3010,7 +3043,7 @@ static void toPunycodeHelper(const QChar *s, int ucLength, QByteArray *output) int outLen = output->length(); output->resize(outLen + ucLength); - char *d = output->data() + outLen; + QChar *d = output->data() + outLen; bool skipped = false; // copy all basic code points verbatim to output. for (uint j = 0; j < (uint) ucLength; ++j) { @@ -3035,7 +3068,7 @@ static void toPunycodeHelper(const QChar *s, int ucLength, QByteArray *output) // if basic code points were copied, add the delimiter character. if (h > 0) - *output += 0x2d; + *output += QChar(0x2d); // while there are still unprocessed non-basic code points left in // the input string... @@ -3083,7 +3116,7 @@ static void toPunycodeHelper(const QChar *s, int ucLength, QByteArray *output) } // prepend ACE prefix - output->insert(outLen, "xn--"); + output->insert(outLen, QLatin1String("xn--")); return; } @@ -3144,11 +3177,15 @@ static bool qt_is_idn_enabled(const QString &domain) int idx = domain.lastIndexOf(QLatin1Char('.')); if (idx == -1) return false; - const QChar *tld = domain.constData() + idx + 1; + int len = domain.size() - idx - 1; + QString tldString(domain.constData() + idx + 1, len); + qt_nameprep(&tldString, 0); + + const QChar *tld = tldString.constData(); if (user_idn_whitelist) - return user_idn_whitelist->contains(QString(tld, len)); + return user_idn_whitelist->contains(tldString); int l = 0; int r = sizeof(idn_whitelist)/sizeof(const char *) - 1; @@ -3164,46 +3201,127 @@ static bool qt_is_idn_enabled(const QString &domain) return equal(tld, len, idn_whitelist[i]); } -static QString qt_from_ACE(const QString &domainMC) +static inline bool isDotDelimiter(ushort uc) { - QString domain = domainMC.toLower(); - int idx = domain.indexOf(QLatin1Char('.')); - if (idx != -1) { - if (!domain.contains(QLatin1String("xn--"))) { - bool simple = true; - for (int i = 0; i < domain.size(); ++i) { - ushort ch = domain.at(i).unicode(); - if (ch > 'z' || ch < '-' || ch == '/' || (ch > '9' && ch < 'A') || (ch > 'Z' && ch < 'a')) { + // IDNA / rfc3490 describes these four delimiters used for + // separating labels in unicode international domain + // names. + return uc == 0x2e || uc == 0x3002 || uc == 0xff0e || uc == 0xff61; +} + +static int nextDotDelimiter(const QString &domain, int from = 0) +{ + const QChar *b = domain.unicode(); + const QChar *ch = b + from; + const QChar *e = b + domain.length(); + while (ch < e) { + if (isDotDelimiter(ch->unicode())) + break; + else + ++ch; + } + return ch - b; +} + +enum AceOperation { ToAceOnly, NormalizeAce }; +static QString qt_ACE_do(const QString &domain, AceOperation op) +{ + if (domain.isEmpty()) + return domain; + + QString result; + result.reserve(domain.length()); + + const bool isIdnEnabled = op == NormalizeAce ? qt_is_idn_enabled(domain) : false; + int lastIdx = 0; + QString aceForm; // this variable is here for caching + + while (1) { + int idx = nextDotDelimiter(domain, lastIdx); + int labelLength = idx - lastIdx; + if (labelLength == 0) + return QString(); // two delimiters in a row -- empty label not allowed + + // RFC 3490 says, about the ToASCII operation: + // 3. If the UseSTD3ASCIIRules flag is set, then perform these checks: + // + // (a) Verify the absence of non-LDH ASCII code points; that is, the + // absence of 0..2C, 2E..2F, 3A..40, 5B..60, and 7B..7F. + // + // (b) Verify the absence of leading and trailing hyphen-minus; that + // is, the absence of U+002D at the beginning and end of the + // sequence. + // and: + // 8. Verify that the number of code points is in the range 1 to 63 + // inclusive. + + // copy the label to the destination, which also serves as our scratch area, lowercasing it + int prevLen = result.size(); + bool simple = true; + result.resize(prevLen + labelLength); + { + QChar *out = result.data() + prevLen; + const QChar *in = domain.constData() + lastIdx; + const QChar *e = in + labelLength; + for (; in < e; ++in, ++out) { + register ushort uc = in->unicode(); + if (uc > 0x7f) simple = false; - break; - } + if (uc >= 'A' && uc <= 'Z') + *out = QChar(uc | 0x20); + else + *out = *in; } - if (simple) - return domain; } - - const bool isIdnEnabled = qt_is_idn_enabled(domain); - int lastIdx = 0; - QString result; - while (1) { - // Nameprep the host. If the labels in the hostname are Punycode - // encoded, we decode them immediately, then nameprep them. - QByteArray label; - toPunycodeHelper(domain.constData() + lastIdx, idx - lastIdx, &label); - result += qt_nameprep(isIdnEnabled ? QUrl::fromPunycode(label) : QString::fromLatin1(label)); - lastIdx = idx + 1; - if (lastIdx < domain.size() + 1) - result += QLatin1Char('.'); - else - break; - idx = domain.indexOf(QLatin1Char('.'), lastIdx); - if (idx == -1) - idx = domain.size(); + + if (simple && labelLength > 6) { + // ACE form domains contain only ASCII characters, but we can't consider them simple + // is this an ACE form? + // the shortest valid ACE domain is 6 characters long (U+0080 would be 1, but it's not allowed) + static const ushort acePrefixUtf16[] = { 'x', 'n', '-', '-' }; + if (memcmp(result.constData() + prevLen, acePrefixUtf16, sizeof acePrefixUtf16) == 0) + simple = false; } - return result; - } else { - return qt_nameprep(domain); + + if (simple) { + // fastest case: this is the common case (non IDN-domains) + // so we're done + if (!qt_check_std3rules(result.constData() + prevLen, labelLength)) + return QString(); + } else { + // Punycode encoding and decoding cannot be done in-place + // That means we need one or two temporaries + qt_nameprep(&result, prevLen); + labelLength = result.length() - prevLen; + register int toReserve = labelLength + 4 + 6; // "xn--" plus some extra bytes + if (toReserve > aceForm.capacity()) + aceForm.reserve(toReserve); + toPunycodeHelper(result.constData() + prevLen, result.size() - prevLen, &aceForm); + + // We use resize()+memcpy() here because we're overwriting the data we've copied + if (isIdnEnabled) { + QString tmp = QUrl::fromPunycode(aceForm.toLatin1()); + if (tmp.isEmpty()) + return QString(); // shouldn't happen, since we've just punycode-encoded it + result.resize(prevLen + tmp.size()); + memcpy(result.data() + prevLen, tmp.constData(), tmp.size() * sizeof(QChar)); + } else { + result.resize(prevLen + aceForm.size()); + memcpy(result.data() + prevLen, aceForm.constData(), aceForm.size() * sizeof(QChar)); + } + + if (!qt_check_std3rules(aceForm.constData(), aceForm.size())) + return QString(); + } + + + lastIdx = idx + 1; + if (lastIdx < domain.size() + 1) + result += QLatin1Char('.'); + else + break; } + return result; } @@ -3246,12 +3364,27 @@ QUrlPrivate::QUrlPrivate(const QUrlPrivate ©) QString QUrlPrivate::canonicalHost() const { - if (QURL_HASFLAG(stateFlags, HostCanonicalized)) + if (QURL_HASFLAG(stateFlags, HostCanonicalized) || host.isEmpty()) return host; QUrlPrivate *that = const_cast<QUrlPrivate *>(this); QURL_SETFLAG(that->stateFlags, HostCanonicalized); - that->host = qt_from_ACE(host); + if (host.contains(QLatin1Char(':'))) { + // This is an IP Literal, use _IPLiteral to validate + QByteArray ba = host.toLatin1(); + if (!ba.startsWith('[')) { + // surround the IP Literal with [ ] if it's not already done so + ba.reserve(ba.length() + 2); + ba.prepend('['); + ba.append(']'); + } + + const char *ptr = ba.constData(); + if (!_IPLiteral(&ptr)) + that->host.clear(); + } else { + that->host = qt_ACE_do(host, NormalizeAce); + } return that->host; } @@ -3737,7 +3870,10 @@ QByteArray QUrlPrivate::toEncoded(QUrl::FormattingOptions options) const } } - url += QUrl::toAce(host); + if (host.startsWith(QLatin1Char('['))) + url += host.toLatin1(); + else + url += QUrl::toAce(host); if (!(options & QUrl::RemovePort) && port != -1) { url += ':'; url += QString::number(port).toAscii(); @@ -4412,8 +4548,6 @@ void QUrl::setHost(const QString &host) QURL_UNSETFLAG(d->stateFlags, QUrlPrivate::Validated | QUrlPrivate::Normalized | QUrlPrivate::HostCanonicalized); d->host = host; - if (d->host.contains(QLatin1Char(':'))) - d->host = QLatin1Char('[') + d->host + QLatin1Char(']'); } /*! @@ -5425,9 +5559,9 @@ QByteArray QUrl::toPercentEncoding(const QString &input, const QByteArray &exclu */ QByteArray QUrl::toPunycode(const QString &uc) { - QByteArray output; + QString output; toPunycodeHelper(uc.constData(), uc.size(), &output); - return output; + return output.toLatin1(); } /*! @@ -5528,7 +5662,7 @@ QString QUrl::fromPunycode(const QByteArray &pc) */ QString QUrl::fromAce(const QByteArray &domain) { - return qt_from_ACE(QString::fromLatin1(domain)); + return qt_ACE_do(QString::fromLatin1(domain), NormalizeAce); } /*! @@ -5545,26 +5679,8 @@ QString QUrl::fromAce(const QByteArray &domain) */ QByteArray QUrl::toAce(const QString &domain) { - // IDNA / rfc3490 describes these four delimiters used for - // separating labels in unicode international domain - // names. - QString nameprepped = qt_nameprep(domain); - int lastIdx = 0; - QByteArray result; - for (int i = 0; i < nameprepped.size(); ++i) { - ushort uc = nameprepped.at(i).unicode(); - if (uc == 0x2e || uc == 0x3002 || uc == 0xff0e || uc == 0xff61) { - if (lastIdx) - result += '.'; - toPunycodeHelper(nameprepped.constData() + lastIdx, i - lastIdx, &result); - lastIdx = i + 1; - } - } - if (lastIdx) - result += '.'; - toPunycodeHelper(nameprepped.constData() + lastIdx, nameprepped.size() - lastIdx, &result); - - return result; + QString result = qt_ACE_do(domain, ToAceOnly); + return result.toLatin1(); } /*! |