Extract Core::TextFile from TextEditor::BaseTextDocument.

Use it in BaseTextEditor and Designer to preserve CRLF
of the files.

Task-number: QTCREATORBUG-5901
Change-Id: I7599ce78649a3b09f2e5118d02f8cbf3db27ed31
Reviewed-on: http://codereview.qt.nokia.com/3591


Reviewed-by: Friedemann Kleint <Friedemann.Kleint@nokia.com>

src/libs/utils/textfileformat.cpp

+/**************************************************************************
+**
+** This file is part of Qt Creator
+**
+** Copyright (c) 2011 Nokia Corporation and/or its subsidiary(-ies).
+**
+** Contact: Nokia Corporation (info@qt.nokia.com)
+**
+**
+** GNU Lesser General Public License Usage
+**
+** This file may be used under the terms of the GNU Lesser General Public
+** License version 2.1 as published by the Free Software Foundation and
+** appearing in the file LICENSE.LGPL included in the packaging of this file.
+** Please review the following information to ensure the GNU Lesser General
+** Public License version 2.1 requirements will be met:
+** http://www.gnu.org/licenses/old-licenses/lgpl-2.1.html.
+**
+** In addition, as a special exception, Nokia gives you certain additional
+** rights. These rights are described in the Nokia Qt LGPL Exception
+** version 1.1, included in the file LGPL_EXCEPTION.txt in this package.
+**
+** Other Usage
+**
+** Alternatively, this file may be used in accordance with the terms and
+** conditions contained in a signed written agreement between you and Nokia.
+**
+** If you have questions regarding the use of this file, please contact
+** Nokia at info@qt.nokia.com.
+**
+**************************************************************************/
+
+#include "textfileformat.h"
+#include "fileutils.h"
+#include "qtcassert.h"
+
+#include <QtCore/QTextCodec>
+#include <QtCore/QStringList>
+#include <QtCore/QCoreApplication>
+#include <QtCore/QDebug>
+
+enum { debug = 0 };
+
+#include <exception>
+
+namespace Utils {
+
+QDebug operator<<(QDebug d, const TextFileFormat &format)
+{
+    QDebug nsp = d.nospace();
+    nsp << "TextFileFormat: ";
+    if (format.codec) {
+        nsp << format.codec->name();
+        foreach (const QByteArray &alias, format.codec->aliases())
+            nsp << ' ' << alias;
+    } else {
+        nsp << "NULL";
+    }
+    nsp << " hasUtf8Bom=" << format.hasUtf8Bom
+        << (format.lineTerminationMode == TextFileFormat::LFLineTerminator ? " LF" : " CRLF");
+    return d;
+}
+
+/*!
+    \class Utils::TextFileFormat
+
+    \brief Describes the format of a text file and provides autodetection.
+
+    The format comprises
+    \list
+    \o Encoding represented by a pointer to a QTextCodec
+    \o Presence of an UTF8 Byte Order Marker (BOM)
+    \o Line feed storage convention
+    \endlist
+
+    The class also provides convenience functions to read text files and return them
+    as strings or string lists and to write out files.
+*/
+
+TextFileFormat::TextFileFormat() :
+    lineTerminationMode(NativeLineTerminator), hasUtf8Bom(false), codec(0)
+{
+}
+
+/*!
+    \brief Detect the format of text data.
+*/
+
+TextFileFormat TextFileFormat::detect(const QByteArray &data)
+{
+    TextFileFormat result;
+    if (data.isEmpty())
+        return result;
+    const int bytesRead = data.size();
+    const unsigned char *buf = reinterpret_cast<const unsigned char *>(data.constData());
+    // code taken from qtextstream
+    if (bytesRead >= 4 && ((buf[0] == 0xff && buf[1] == 0xfe && buf[2] == 0 && buf[3] == 0)
+                           || (buf[0] == 0 && buf[1] == 0 && buf[2] == 0xfe && buf[3] == 0xff))) {
+        result.codec = QTextCodec::codecForName("UTF-32");
+    } else if (bytesRead >= 2 && ((buf[0] == 0xff && buf[1] == 0xfe)
+                                  || (buf[0] == 0xfe && buf[1] == 0xff))) {
+        result.codec = QTextCodec::codecForName("UTF-16");
+    } else if (bytesRead >= 3 && ((buf[0] == 0xef && buf[1] == 0xbb) && buf[2] == 0xbf)) {
+        result.codec = QTextCodec::codecForName("UTF-8");
+        result.hasUtf8Bom = true;
+    }
+    // end code taken from qtextstream
+    const int newLinePos = data.indexOf('\n');
+    if (newLinePos == -1) {
+        result.lineTerminationMode = NativeLineTerminator;
+    } else if (newLinePos == 0) {
+        result.lineTerminationMode = LFLineTerminator;
+    } else {
+        result.lineTerminationMode = data.at(newLinePos - 1) == '\r' ? CRLFLineTerminator : LFLineTerminator;
+    }
+    return result;
+}
+
+/*!
+    \brief Returns a piece of text suitable as display for a encoding error.
+*/
+
+QByteArray TextFileFormat::decodingErrorSample(const QByteArray &data)
+{
+    const int p = data.indexOf('\n', 16384);
+    return p < 0 ? data : data.left(p);
+}
+
+enum { textChunkSize = 65536 };
+
+static bool verifyDecodingError(const QString &text, const QTextCodec *codec,
+                                const char *data, const int dataSize,
+                                const bool possibleHeader)
+{
+    const QByteArray verifyBuf = codec->fromUnicode(text); // slow
+    // the minSize trick lets us ignore unicode headers
+    const int minSize = qMin(verifyBuf.size(), dataSize);
+    return (minSize < dataSize - (possibleHeader? 4 : 0)
+            || memcmp(verifyBuf.constData() + verifyBuf.size() - minSize,
+                      data + dataSize - minSize,
+                      minSize));
+}
+
+// Decode a potentially large file in chunks and append it to target
+// using the append function passed on (fits QStringList and QString).
+
+template <class Target>
+bool decodeTextFileContent(const QByteArray &dataBA,
+                           const TextFileFormat &format,
+                           Target *target,
+                           void (Target::*appendFunction)(const QString &))
+{
+    QTC_ASSERT(format.codec, return  false; )
+
+    QTextCodec::ConverterState state;
+    bool hasDecodingError = false;
+
+    const char *start = dataBA.constData();
+    const char *data = start;
+    const char *end  = data + dataBA.size();
+    // Process chunkwise as QTextCodec allocates too much memory when doing it in one
+    // go. An alternative to the code below would be creating a decoder from the codec,
+    // but its failure detection does not seem be working reliably.
+    for (const char *data = start; data < end; ) {
+        const char *chunkStart = data;
+        const int chunkSize = qMin(int(textChunkSize), int(end - chunkStart));
+        QString text = format.codec->toUnicode(chunkStart, chunkSize, &state);
+        data += chunkSize;
+        // Process until the end of the current multi-byte character. Remaining might
+        // actually contain more than needed so try one-be-one. If EOF is reached with
+        // and characters remain->encoding error.
+        for ( ; state.remainingChars && data < end ; ++data)
+            text.append(format.codec->toUnicode(data, 1, &state));
+        if (state.remainingChars)
+            hasDecodingError = true;
+        if (!hasDecodingError)
+            hasDecodingError =
+                verifyDecodingError(text, format.codec, chunkStart, data - chunkStart,
+                                    chunkStart == start);
+        if (format.lineTerminationMode == TextFileFormat::CRLFLineTerminator)
+            text.remove(QLatin1Char('\r'));
+        (target->*appendFunction)(text);
+    }
+    return !hasDecodingError;
+}
+
+/*!
+    \brief Decode data to a plain string.
+*/
+
+bool TextFileFormat::decode(const QByteArray &data, QString *target) const
+{
+    target->clear();
+    return decodeTextFileContent<QString>(data, *this, target, &QString::push_back);
+}
+
+/*!
+    \brief Decode data to a list of strings.
+
+    Intended for use with progress bars loading large files.
+*/
+
+bool TextFileFormat::decode(const QByteArray &data, QStringList *target) const
+{
+    target->clear();
+    if (data.size() > textChunkSize)
+        target->reserve(5 + data.size() / textChunkSize);
+    return decodeTextFileContent<QStringList>(data, *this, target, &QStringList::append);
+}
+
+// Read text file contents to string or stringlist.
+template <class Target>
+TextFileFormat::ReadResult readTextFile(const QString &fileName, const QTextCodec *defaultCodec,
+                                        Target *target, TextFileFormat *format, QString *errorString,
+                                        QByteArray *decodingErrorSampleIn = 0)
+{
+    if (decodingErrorSampleIn)
+        decodingErrorSampleIn->clear();
+
+    QByteArray data;
+    try {
+        Utils::FileReader reader;
+        if (!reader.fetch(fileName, errorString))
+            return TextFileFormat::ReadIOError;
+        data = reader.data();
+    } catch (const std::bad_alloc &) {
+        *errorString = QCoreApplication::translate("Utils::TextFileFormat", "Out of memory.");
+        return TextFileFormat::ReadMemoryAllocationError;
+    }
+
+    *format = TextFileFormat::detect(data);
+    if (!format->codec)
+        format->codec = defaultCodec ? defaultCodec : QTextCodec::codecForLocale();
+
+    if (!format->decode(data, target)) {
+        *errorString = QCoreApplication::translate("Utils::TextFileFormat", "An encoding error was encountered.");
+        if (decodingErrorSampleIn)
+            *decodingErrorSampleIn = TextFileFormat::decodingErrorSample(data);
+        return TextFileFormat::ReadEncodingError;
+    }
+    return TextFileFormat::ReadSuccess;
+}
+
+/*!
+    \brief Read text file into a list of strings.
+*/
+
+TextFileFormat::ReadResult
+    TextFileFormat::readFile(const QString &fileName, const QTextCodec *defaultCodec,
+                             QStringList *plainTextList, TextFileFormat *format, QString *errorString,
+                             QByteArray *decodingErrorSample /* = 0 */)
+{
+    const TextFileFormat::ReadResult result =
+        readTextFile(fileName, defaultCodec,
+                     plainTextList, format, errorString, decodingErrorSample);
+    if (debug)
+        qDebug().nospace() << Q_FUNC_INFO << fileName << ' ' << *format
+                           << " returns " << result << '/' << plainTextList->size() << " chunks";
+    return result;
+}
+
+/*!
+    \brief Read text file into a string.
+*/
+
+TextFileFormat::ReadResult
+    TextFileFormat::readFile(const QString &fileName, const QTextCodec *defaultCodec,
+                             QString *plainText, TextFileFormat *format, QString *errorString,
+                             QByteArray *decodingErrorSample /* = 0 */)
+{
+    const TextFileFormat::ReadResult result =
+        readTextFile(fileName, defaultCodec,
+                     plainText, format, errorString, decodingErrorSample);
+    if (debug)
+        qDebug().nospace() << Q_FUNC_INFO << fileName << ' ' << *format
+                           << " returns " << result << '/' << plainText->size() << " characters";
+    return result;
+}
+
+/*!
+    \brief Write out a text file.
+*/
+
+bool TextFileFormat::writeFile(const QString &fileName, QString plainText, QString *errorString) const
+{
+    QTC_ASSERT(codec, return false;)
+
+    // Does the user want CRLF? If that is native,
+    // let QFile do the work, else manually add.
+    QIODevice::OpenMode fileMode = QIODevice::NotOpen;
+    if (lineTerminationMode == CRLFLineTerminator) {
+        if (NativeLineTerminator == CRLFLineTerminator) {
+            fileMode |= QIODevice::Text;
+        } else {
+            plainText.replace(QLatin1Char('\n'), QLatin1String("\r\n"));
+        }
+    }
+
+    Utils::FileSaver saver(fileName, fileMode);
+    if (saver.hasError()) {
+        *errorString = saver.errorString();
+        return false;
+    }
+    if (hasUtf8Bom && codec->name() == "UTF-8")
+        saver.write("\xef\xbb\xbf", 3);
+    saver.write(codec->fromUnicode(plainText));
+    const bool ok = saver.finalize(errorString);
+    if (debug)
+        qDebug().nospace() << Q_FUNC_INFO << fileName << ' ' << *this <<  ' ' << plainText.size()
+                           << " bytes, returns " << ok;
+    return ok;
+}
+
+} // namespace Utils

src/libs/utils/textfileformat.h

+/**************************************************************************
+**
+** This file is part of Qt Creator
+**
+** Copyright (c) 2011 Nokia Corporation and/or its subsidiary(-ies).
+**
+** Contact: Nokia Corporation (info@qt.nokia.com)
+**
+**
+** GNU Lesser General Public License Usage
+**
+** This file may be used under the terms of the GNU Lesser General Public
+** License version 2.1 as published by the Free Software Foundation and
+** appearing in the file LICENSE.LGPL included in the packaging of this file.
+** Please review the following information to ensure the GNU Lesser General
+** Public License version 2.1 requirements will be met:
+** http://www.gnu.org/licenses/old-licenses/lgpl-2.1.html.
+**
+** In addition, as a special exception, Nokia gives you certain additional
+** rights. These rights are described in the Nokia Qt LGPL Exception
+** version 1.1, included in the file LGPL_EXCEPTION.txt in this package.
+**
+** Other Usage
+**
+** Alternatively, this file may be used in accordance with the terms and
+** conditions contained in a signed written agreement between you and Nokia.
+**
+** If you have questions regarding the use of this file, please contact
+** Nokia at info@qt.nokia.com.
+**
+**************************************************************************/
+
+#ifndef TEXTFILEUTILS_H
+#define TEXTFILEUTILS_H
+
+#include "utils_global.h"
+
+#include <QtCore/QtGlobal>
+
+QT_BEGIN_NAMESPACE
+class QTextCodec;
+class QStringList;
+class QString;
+class QByteArray;
+QT_END_NAMESPACE
+
+namespace Utils {
+
+class QTCREATOR_UTILS_EXPORT TextFileFormat  {
+public:
+    enum LineTerminationMode
+    {
+        LFLineTerminator,
+        CRLFLineTerminator,
+        NativeLineTerminator =
+#if defined (Q_OS_WIN)
+        CRLFLineTerminator,
+#else
+        LFLineTerminator
+#endif
+    };
+
+    enum ReadResult
+    {
+        ReadSuccess,
+        ReadEncodingError,
+        ReadMemoryAllocationError,
+        ReadIOError
+    };
+
+    TextFileFormat();
+
+    static TextFileFormat detect(const QByteArray &data);
+
+    bool decode(const QByteArray &data, QString *target) const;
+    bool decode(const QByteArray &data, QStringList *target) const;
+
+    static ReadResult readFile(const QString &fileName, const QTextCodec *defaultCodec,
+                               QStringList *plainText, TextFileFormat *format, QString *errorString,
+                               QByteArray *decodingErrorSample = 0);
+    static ReadResult readFile(const QString &fileName, const QTextCodec *defaultCodec,
+                               QString *plainText, TextFileFormat *format, QString *errorString,
+                               QByteArray *decodingErrorSample = 0);
+
+    bool writeFile(const QString &fileName, QString plainText, QString *errorString) const;
+
+    static QByteArray decodingErrorSample(const QByteArray &data);
+
+    LineTerminationMode lineTerminationMode;
+    bool hasUtf8Bom;
+    const QTextCodec *codec;
+};
+
+} // namespace Utils
+
+#endif // TEXTFILEUTILS_H

src/libs/utils/utils-lib.pri

@@ -40,6 +40,7 @@ SOURCES += $$PWD/environment.cpp \
     $$PWD/synchronousprocess.cpp \
     $$PWD/savefile.cpp \
     $$PWD/fileutils.cpp \
+    $$PWD/textfileformat.cpp \
     $$PWD/submitfieldwidget.cpp \
     $$PWD/consoleprocess.cpp \
     $$PWD/uncommentselection.cpp \
@@ -135,6 +136,7 @@ HEADERS += \
     $$PWD/synchronousprocess.h \
     $$PWD/savefile.h \
     $$PWD/fileutils.h \
+    $$PWD/textfileformat.h \
     $$PWD/submitfieldwidget.h \
     $$PWD/uncommentselection.h \
     $$PWD/parameteraction.h \

src/plugins/coreplugin/coreplugin.pro

@@ -91,7 +91,8 @@ SOURCES += mainwindow.cpp \
     mimetypemagicdialog.cpp \
     mimetypesettings.cpp \
     dialogs/promptoverwritedialog.cpp \
-    fileutils.cpp
+    fileutils.cpp \
+    textfile.cpp
 
 HEADERS += mainwindow.h \
     editmode.h \
@@ -185,7 +186,8 @@ HEADERS += mainwindow.h \
     mimetypesettings.h \
     dialogs/promptoverwritedialog.h \
     fileutils.h \
-    externaltoolmanager.h
+    externaltoolmanager.h \
+    textfile.h
 
 FORMS += dialogs/newdialog.ui \
     actionmanager/commandmappings.ui \

src/plugins/coreplugin/textfile.cpp

+/**************************************************************************
+**
+** This file is part of Qt Creator
+**
+** Copyright (c) 2011 Nokia Corporation and/or its subsidiary(-ies).
+**
+** Contact: Nokia Corporation (info@qt.nokia.com)
+**
+**
+** GNU Lesser General Public License Usage
+**
+** This file may be used under the terms of the GNU Lesser General Public
+** License version 2.1 as published by the Free Software Foundation and
+** appearing in the file LICENSE.LGPL included in the packaging of this file.
+** Please review the following information to ensure the GNU Lesser General
+** Public License version 2.1 requirements will be met:
+** http://www.gnu.org/licenses/old-licenses/lgpl-2.1.html.
+**
+** In addition, as a special exception, Nokia gives you certain additional
+** rights. These rights are described in the Nokia Qt LGPL Exception
+** version 1.1, included in the file LGPL_EXCEPTION.txt in this package.
+**
+** Other Usage
+**
+** Alternatively, this file may be used in accordance with the terms and
+** conditions contained in a signed written agreement between you and Nokia.
+**
+** If you have questions regarding the use of this file, please contact
+** Nokia at info@qt.nokia.com.
+**
+**************************************************************************/
+
+#include "textfile.h"
+#include "editormanager.h"
+
+#include <QtCore/QDebug>
+#include <QtCore/QTextCodec>
+
+/*!
+    \class Core::TextFile
+    \brief Base class for text files with encoding helpers.
+
+    Stores the format obtained from read operations and uses that when writing
+    out files, thus ensuring that CRLF/encodings are preserved.
+
+    \sa Utils::TextFileUtils
+*/
+
+enum { debug = 0 };
+
+namespace Core {
+
+namespace Internal {
+class TextFilePrivate
+{
+public:
+    TextFilePrivate() : m_readResult(Utils::TextFileFormat::ReadSuccess) {}
+
+    Utils::TextFileFormat m_format;
+    Utils::TextFileFormat::ReadResult m_readResult;
+    QByteArray m_decodingErrorSample;
+};
+
+} // namespace Internal
+
+TextFile::TextFile(QObject *parent) :
+    IFile(parent), d(new Internal::TextFilePrivate)
+{
+    setCodec(Core::EditorManager::instance()->defaultTextCodec());
+}
+
+TextFile::~TextFile()
+{
+    delete d;
+}
+
+bool TextFile::hasDecodingError() const
+{
+    return d->m_readResult == Utils::TextFileFormat::ReadEncodingError;
+}
+
+QByteArray TextFile::decodingErrorSample() const
+{
+    return d->m_decodingErrorSample;
+}
+
+/*!
+    \brief Writes out text using the format obtained from the last read.
+*/
+
+bool TextFile::write(const QString &fileName, const QString &data, QString *errorMessage) const
+{
+    return write(fileName, format(), data, errorMessage);
+}
+
+/*!
+    \brief Writes out text using a custom format obtained.
+*/
+
+bool TextFile::write(const QString &fileName, const Utils::TextFileFormat &format, const QString &data, QString *errorMessage) const
+{
+    if (debug)
+        qDebug() << Q_FUNC_INFO << this << fileName;
+    return format.writeFile(fileName, data, errorMessage);
+}
+
+/*!
+    \brief Autodetect format and read in a text file.
+*/
+
+TextFile::ReadResult TextFile::read(const QString &fileName, QStringList *plainTextList, QString *errorString)
+{
+    d->m_readResult =
+        Utils::TextFileFormat::readFile(fileName, codec(),
+                                        plainTextList, &d->m_format, errorString, &d->m_decodingErrorSample);
+    return d->m_readResult;
+}
+
+/*!
+    \brief Autodetect format and read in a text file.
+*/
+
+TextFile::ReadResult TextFile::read(const QString &fileName, QString *plainText, QString *errorString)
+{
+    d->m_readResult =
+        Utils::TextFileFormat::readFile(fileName, codec(),
+                                        plainText, &d->m_format, errorString, &d->m_decodingErrorSample);
+    return d->m_readResult;
+}
+
+const QTextCodec *TextFile::codec() const
+{
+    return d->m_format.codec;
+}
+
+void TextFile::setCodec(const QTextCodec *codec)
+{
+    if (debug)
+        qDebug() << Q_FUNC_INFO << this << (codec ? codec->name() : QByteArray());
+    d->m_format.codec = codec;
+}
+
+/*!
+    \brief Returns the format obtained from the last call to read().
+*/
+
+Utils::TextFileFormat TextFile::format() const
+{
+    return d->m_format;
+}
+
+} // namespace Core

src/plugins/coreplugin/textfile.h

+/**************************************************************************
+**
+** This file is part of Qt Creator
+**
+** Copyright (c) 2011 Nokia Corporation and/or its subsidiary(-ies).
+**
+** Contact: Nokia Corporation (info@qt.nokia.com)