| 949 | #define PEEK_LENGTH 1024 |
| 950 | |
| 951 | QString TeXDocumentWindow::readFile(const QFileInfo & fileInfo, |
| 952 | QTextCodec **codecUsed, |
| 953 | int *lineEndings, |
| 954 | QTextCodec * forceCodec) |
| 955 | // reads the text from a file, after checking for %!TEX encoding.... metadata |
| 956 | // sets codecUsed to the QTextCodec used to read the text |
| 957 | // returns a null (not just empty) QString on failure |
| 958 | { |
| 959 | if (lineEndings) { |
| 960 | // initialize to default for the platform |
| 961 | #if defined(Q_OS_WIN) |
| 962 | *lineEndings = kLineEnd_CRLF; |
| 963 | #else |
| 964 | *lineEndings = kLineEnd_LF; |
| 965 | #endif |
| 966 | } |
| 967 | |
| 968 | utf8BOM = false; |
| 969 | QFile file(fileInfo.absoluteFilePath()); |
| 970 | // Not using QFile::Text because this prevents us reading "classic" Mac files |
| 971 | // with CR-only line endings. See issue #242. |
| 972 | if (!file.open(QFile::ReadOnly)) { |
| 973 | QMessageBox::warning(this, QCoreApplication::applicationName(), |
| 974 | tr("Cannot read file \"%1\":\n%2") |
| 975 | .arg(fileInfo.absoluteFilePath(), file.errorString())); |
| 976 | return QString(); |
| 977 | } |
| 978 | |
| 979 | QByteArray peekBytes(file.peek(PEEK_LENGTH)); |
| 980 | |
| 981 | QString reqName; |
| 982 | if (forceCodec) |
| 983 | *codecUsed = forceCodec; |
| 984 | else { |
| 985 | bool hasMetadata{false}; |
| 986 | *codecUsed = scanForEncoding(QString::fromUtf8(peekBytes.constData()), hasMetadata, reqName); |
| 987 | if (!(*codecUsed)) { |
| 988 | *codecUsed = TWApp::instance()->getDefaultCodec(); |
| 989 | if (hasMetadata) { |
| 990 | if (QMessageBox::warning(this, tr("Unrecognized encoding"), |
| 991 | tr("The text encoding %1 used in %2 is not supported.\n\n" |
| 992 | "It will be interpreted as %3 instead, which may result in incorrect text.") |
| 993 | .arg(reqName, fileInfo.absoluteFilePath(), QString::fromUtf8((*codecUsed)->name().constData())), |
| 994 | QMessageBox::Ok | QMessageBox::Cancel, QMessageBox::Ok) == QMessageBox::Cancel) |
| 995 | return QString(); |
| 996 | } |
| 997 | } |
| 998 | } |
| 999 | |
| 1000 | // When using the UTF-8 codec (mib = 106), byte order marks (BOMs) are |
| 1001 | // ignored during reading and not produced when writing. To keep them in |
| 1002 | // files that have them, we need to check for them ourselves. |
| 1003 | if ((*codecUsed)->mibEnum() == 106 && peekBytes.size() >= 3 && peekBytes[0] == '\xEF' && peekBytes[1] == '\xBB' && peekBytes[2] == '\xBF') |
| 1004 | utf8BOM = true; |
| 1005 | |
| 1006 | // If the file is empty (we're already at the end), don't try to read |
| 1007 | // anything using QTextStream below as that would return a Null-String |
| 1008 | // (QString()) rather than an empty string (QString("")). Instead, return |
no test coverage detected