diff --git a/src/corelib/io/qdebug.cpp b/src/corelib/io/qdebug.cpp index 357d63137c..8eeac26d3b 100644 --- a/src/corelib/io/qdebug.cpp +++ b/src/corelib/io/qdebug.cpp @@ -174,6 +174,16 @@ void QDebug::putUcs4(uint ucs4) maybeQuote('\''); } +// These two functions return true if the character should be printed by QDebug. +// For QByteArray, this is technically identical to US-ASCII isprint(); +// for QString, we use QChar::isPrint, which requires a full UCS-4 decode. +static inline bool isPrintable(uint ucs4) +{ return QChar::isPrint(ucs4); } +static inline bool isPrintable(ushort uc) +{ return QChar::isPrint(uc); } +static inline bool isPrintable(uchar c) +{ return c >= ' ' && c < 0x7f; } + template static inline void putEscapedString(QTextStreamPrivate *d, const Char *begin, int length, bool isUnicode = true) { @@ -194,22 +204,23 @@ static inline void putEscapedString(QTextStreamPrivate *d, const Char *begin, in } if (sizeof(Char) == sizeof(QChar)) { + // Surrogate characters are category Cs (Other_Surrogate), so isPrintable = false for them int runLength = 0; while (p + runLength != end && - p[runLength] < 0x7f && p[runLength] >= 0x20 && p[runLength] != '\\' && p[runLength] != '"') + isPrintable(p[runLength]) && p[runLength] != '\\' && p[runLength] != '"') ++runLength; if (runLength) { d->write(reinterpret_cast(p), runLength); p += runLength - 1; continue; } - } else if (*p < 0x7f && *p >= 0x20 && *p != '\\' && *p != '"') { + } else if (isPrintable(*p) && *p != '\\' && *p != '"') { QChar c = QLatin1Char(*p); d->write(&c, 1); continue; } - // print as an escape sequence + // print as an escape sequence (maybe, see below for surrogate pairs) int buflen = 2; ushort buf[sizeof "\\U12345678" - 1]; buf[0] = '\\'; @@ -248,17 +259,23 @@ static inline void putEscapedString(QTextStreamPrivate *d, const Char *begin, in if ((p + 1) != end && QChar::isLowSurrogate(p[1])) { // properly-paired surrogates uint ucs4 = QChar::surrogateToUcs4(*p, p[1]); + if (isPrintable(ucs4)) { + buf[0] = *p; + buf[1] = p[1]; + buflen = 2; + } else { + buf[1] = 'U'; + buf[2] = '0'; // toHexUpper(ucs4 >> 32); + buf[3] = '0'; // toHexUpper(ucs4 >> 28); + buf[4] = toHexUpper(ucs4 >> 20); + buf[5] = toHexUpper(ucs4 >> 16); + buf[6] = toHexUpper(ucs4 >> 12); + buf[7] = toHexUpper(ucs4 >> 8); + buf[8] = toHexUpper(ucs4 >> 4); + buf[9] = toHexUpper(ucs4); + buflen = 10; + } ++p; - buf[1] = 'U'; - buf[2] = '0'; // toHexUpper(ucs4 >> 32); - buf[3] = '0'; // toHexUpper(ucs4 >> 28); - buf[4] = toHexUpper(ucs4 >> 20); - buf[5] = toHexUpper(ucs4 >> 16); - buf[6] = toHexUpper(ucs4 >> 12); - buf[7] = toHexUpper(ucs4 >> 8); - buf[8] = toHexUpper(ucs4 >> 4); - buf[9] = toHexUpper(ucs4); - buflen = 10; break; } // improperly-paired surrogates, fall through diff --git a/tests/auto/corelib/io/qdebug/tst_qdebug.cpp b/tests/auto/corelib/io/qdebug/tst_qdebug.cpp index 4e5cfed019..3e19e816c9 100644 --- a/tests/auto/corelib/io/qdebug/tst_qdebug.cpp +++ b/tests/auto/corelib/io/qdebug/tst_qdebug.cpp @@ -379,18 +379,48 @@ void tst_QDebug::qDebugQString() const qDebug().noquote().nospace() << qSetFieldWidth(8) << string; QCOMPARE(s_msg, " " + string); - string = QLatin1String("\nSm\xF8rg\xE5sbord\\"); + string = "Sm\xc3\xb8rg\xc3\xa5sbord " // Latin script + "\xce\x91\xce\xb8\xce\xae\xce\xbd\xce\xb1 " // Greek script + "\xd0\x9c\xd0\xbe\xd1\x81\xd0\xba\xd0\xb2\xd0\xb0"; // Cyrillic script qDebug().noquote().nospace() << string; QCOMPARE(s_msg, string); + // This string only contains printable characters qDebug() << string; - QCOMPARE(s_msg, QString("\"\\nSm\\u00F8rg\\u00E5sbord\\\\\"")); + QCOMPARE(s_msg, '"' + string + '"'); - // surrogate pairs (including broken pairings) - ushort utf16[] = { 0xDC00, 0xD800, 0xDC00, 'x', 0xD800, 0xDC00, 0xD800, 0 }; + string = "\n\t\\\""; + qDebug().noquote().nospace() << string; + QCOMPARE(s_msg, string); + + // This string only contains characters that must be escaped + qDebug() << string; + QCOMPARE(s_msg, QString("\"\\n\\t\\\\\\\"\"")); + + // Unicode escapes, BMP + string = "\1" // U+0001: START OF HEADING (category Cc) + "\x7f" // U+007F: DELETE (category Cc) + "\xc2\xad" // U+00AD: SOFT HYPHEN (category Cf) + "\xef\xbb\xbf"; // U+FEFF: ZERO WIDTH NO-BREAK SPACE / BOM (category Cf) + qDebug() << string; + QCOMPARE(s_msg, QString("\"\\u0001\\u007F\\u00AD\\uFEFF\"")); + + // Unicode printable non-BMP + string = "\xf0\x90\x80\x80"; // U+10000: LINEAR B SYLLABLE B008 A (category Lo) + qDebug() << string; + QCOMPARE(s_msg, '"' + string + '"'); + + // non-BMP and non-printable + string = "\xf3\xa0\x80\x81 " // U+E0001: LANGUAGE TAG (category Cf) + "\xf4\x80\x80\x80"; // U+100000: Plane 16 Private Use (category Co) + qDebug() << string; + QCOMPARE(s_msg, QString("\"\\U000E0001 \\U00100000\"")); + + // broken surrogate pairs + ushort utf16[] = { 0xDC00, 0xD800, 'x', 0xD800, 0 }; string = QString::fromUtf16(utf16); qDebug() << string; - QCOMPARE(s_msg, QString("\"\\uDC00\\U00010000x\\U00010000\\uD800\"")); + QCOMPARE(s_msg, QString("\"\\uDC00\\uD800x\\uD800\"")); } void tst_QDebug::qDebugQStringRef() const diff --git a/tests/auto/testlib/selftests/expected_badxml.lightxml b/tests/auto/testlib/selftests/expected_badxml.lightxml index e0de44f0be..94f479a79a 100644 --- a/tests/auto/testlib/selftests/expected_badxml.lightxml +++ b/tests/auto/testlib/selftests/expected_badxml.lightxml @@ -113,7 +113,7 @@ - + diff --git a/tests/auto/testlib/selftests/expected_badxml.xml b/tests/auto/testlib/selftests/expected_badxml.xml index 8667c2d9cd..c1266bfeed 100644 --- a/tests/auto/testlib/selftests/expected_badxml.xml +++ b/tests/auto/testlib/selftests/expected_badxml.xml @@ -115,7 +115,7 @@ - + diff --git a/tests/auto/testlib/selftests/expected_badxml.xunitxml b/tests/auto/testlib/selftests/expected_badxml.xunitxml index 46f4be4391..49048fdad6 100644 --- a/tests/auto/testlib/selftests/expected_badxml.xunitxml +++ b/tests/auto/testlib/selftests/expected_badxml.xunitxml @@ -30,7 +30,7 @@ - + @@ -46,6 +46,6 @@ open < tags < text]]> " mixed ]]]> up > " in < the ]]]> hopes < of triggering "< ]]]> bugs]]> - +