author | zecke <zecke> | 2003-04-20 17:24:50 (UTC) |
---|---|---|
committer | zecke <zecke> | 2003-04-20 17:24:50 (UTC) |
commit | 92c687d281f69085436a77efb8cd1d4d2d9333f8 (patch) (side-by-side diff) | |
tree | 7bbfac3ea310ed140e78cb3c17f8219294cb61d0 /development/translation/opie-lupdate | |
parent | f1f4e6794507d9b8dafb46ce05968a0647a41777 (diff) | |
download | opie-92c687d281f69085436a77efb8cd1d4d2d9333f8.zip opie-92c687d281f69085436a77efb8cd1d4d2d9333f8.tar.gz opie-92c687d281f69085436a77efb8cd1d4d2d9333f8.tar.bz2 |
Initial revision
Diffstat (limited to 'development/translation/opie-lupdate') (more/less context) (ignore whitespace changes)
-rw-r--r-- | development/translation/opie-lupdate/fetchtr.cpp | 784 | ||||
-rwxr-xr-x | development/translation/opie-lupdate/lupdate | bin | 0 -> 128093 bytes | |||
-rw-r--r-- | development/translation/opie-lupdate/main.cpp | 215 | ||||
-rw-r--r-- | development/translation/opie-lupdate/merge.cpp | 115 | ||||
-rw-r--r-- | development/translation/opie-lupdate/numberh.cpp | 235 | ||||
-rw-r--r-- | development/translation/opie-lupdate/opie-lupdate.pro | 21 | ||||
-rw-r--r-- | development/translation/opie-lupdate/sametexth.cpp | 84 |
7 files changed, 1454 insertions, 0 deletions
diff --git a/development/translation/opie-lupdate/fetchtr.cpp b/development/translation/opie-lupdate/fetchtr.cpp new file mode 100644 index 0000000..eb25555 --- a/dev/null +++ b/development/translation/opie-lupdate/fetchtr.cpp @@ -0,0 +1,784 @@ +/********************************************************************** +** Copyright (C) 2000 Trolltech AS. All rights reserved. +** +** This file is part of Qt Linguist. +** +** This file may be distributed and/or modified under the terms of the +** GNU General Public License version 2 as published by the Free Software +** Foundation and appearing in the file LICENSE.GPL included in the +** packaging of this file. +** +** This file is provided AS IS with NO WARRANTY OF ANY KIND, INCLUDING THE +** WARRANTY OF DESIGN, MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. +** +** See http://www.trolltech.com/gpl/ for GPL licensing information. +** +** Contact info@trolltech.com if any conditions of this licensing are +** not clear to you. +** +**********************************************************************/ + +#include <metatranslator.h> + +#include <qfile.h> +#include <qregexp.h> +#include <qstring.h> +#include <qtextstream.h> +#include <qvaluestack.h> +#include <qxml.h> + +#include <ctype.h> +#include <errno.h> +#include <stdio.h> +#include <string.h> + +/* qmake ignore Q_OBJECT */ + +static const char MagicComment[] = "TRANSLATOR "; + +static QMap<QCString, int> needs_Q_OBJECT; +static QMap<QCString, int> lacks_Q_OBJECT; + +/* + The first part of this source file is the C++ tokenizer. We skip + most of C++; the only tokens that interest us are defined here. + Thus, the code fragment + + int main() + { + printf( "Hello, world!\n" ); + return 0; + } + + is broken down into the following tokens (Tok_ omitted): + + Ident Ident LeftParen RightParen + LeftBrace + Ident LeftParen String RightParen Semicolon + return Semicolon + RightBrace. + + The 0 doesn't produce any token. +*/ + +enum { Tok_Eof, Tok_class, Tok_namespace, Tok_return, Tok_tr, + Tok_trUtf8, Tok_translate, Tok_Q_OBJECT, Tok_Ident, + Tok_Comment, Tok_String, Tok_Arrow, Tok_Colon, + Tok_Gulbrandsen, Tok_LeftBrace, Tok_RightBrace, Tok_LeftParen, + Tok_RightParen, Tok_Comma, Tok_Semicolon }; + +/* + The tokenizer maintains the following global variables. The names + should be self-explanatory. +*/ +static QCString yyFileName; +static int yyCh; +static char yyIdent[128]; +static size_t yyIdentLen; +static char yyComment[65536]; +static size_t yyCommentLen; +static char yyString[16384]; +static size_t yyStringLen; +static QValueStack<int> yySavedBraceDepth; +static int yyBraceDepth; +static int yyParenDepth; +static int yyLineNo; +static int yyCurLineNo; + +// the file to read from (if reading from a file) +static FILE *yyInFile; + +// the string to read from and current position in the string (otherwise) +static QString yyInStr; +static int yyInPos; + +static int (*getChar)(); + +static int getCharFromFile() +{ + int c = getc( yyInFile ); + if ( c == '\n' ) + yyCurLineNo++; + return c; +} + +static int getCharFromString() +{ + if ( yyInPos == (int) yyInStr.length() ) { + return EOF; + } else { + return yyInStr[yyInPos++].latin1(); + } +} + +static void startTokenizer( const char *fileName, int (*getCharFunc)() ) +{ + yyInPos = 0; + getChar = getCharFunc; + + yyFileName = fileName; + yyCh = getChar(); + yySavedBraceDepth.clear(); + yyBraceDepth = 0; + yyParenDepth = 0; + yyCurLineNo = 1; +} + +static int getToken() +{ + const char tab[] = "abfnrtv"; + const char backTab[] = "\a\b\f\n\r\t\v"; + uint n; + + yyIdentLen = 0; + yyCommentLen = 0; + yyStringLen = 0; + + while ( yyCh != EOF ) { + yyLineNo = yyCurLineNo; + + if ( isalpha(yyCh) || yyCh == '_' ) { + do { + if ( yyIdentLen < sizeof(yyIdent) - 1 ) + yyIdent[yyIdentLen++] = (char) yyCh; + yyCh = getChar(); + } while ( isalnum(yyCh) || yyCh == '_' ); + yyIdent[yyIdentLen] = '\0'; + + switch ( yyIdent[0] ) { + case 'Q': + if ( strcmp(yyIdent + 1, "_OBJECT") == 0 ) { + return Tok_Q_OBJECT; + } else if ( strcmp(yyIdent + 1, "T_TR_NOOP") == 0 ) { + return Tok_tr; + } else if ( strcmp(yyIdent + 1, "T_TRANSLATE_NOOP") == 0 ) { + return Tok_translate; + } + break; + case 'T': + // TR() for when all else fails + if ( qstricmp(yyIdent + 1, "R") == 0 ) + return Tok_tr; + break; + case 'c': + if ( strcmp(yyIdent + 1, "lass") == 0 ) + return Tok_class; + break; + case 'n': + if ( strcmp(yyIdent + 1, "amespace") == 0 ) + return Tok_namespace; + break; + case 'r': + if ( strcmp(yyIdent + 1, "eturn") == 0 ) + return Tok_return; + break; + case 's': + if ( strcmp(yyIdent + 1, "truct") == 0 ) + return Tok_class; + break; + case 't': + if ( strcmp(yyIdent + 1, "r") == 0 ) { + return Tok_tr; + } else if ( qstrcmp(yyIdent + 1, "rUtf8") == 0 ) { + return Tok_trUtf8; + } else if ( qstrcmp(yyIdent + 1, "ranslate") == 0 ) { + return Tok_translate; + } + } + return Tok_Ident; + } else { + switch ( yyCh ) { + case '#': + /* + Early versions of lupdate complained about + unbalanced braces in the following code: + + #ifdef ALPHA + while ( beta ) { + #else + while ( gamma ) { + #endif + delta; + } + + The code contains, indeed, two opening braces for + one closing brace; yet there's no reason to panic. + + The solution is to remember yyBraceDepth as it was + when #if, #ifdef or #ifndef was met, and to set + yyBraceDepth to that value when meeting #elif or + #else. + */ + do { + yyCh = getChar(); + } while ( isspace(yyCh) && yyCh != '\n' ); + + switch ( yyCh ) { + case 'i': + yyCh = getChar(); + if ( yyCh == 'f' ) { + // if, ifdef, ifndef + yySavedBraceDepth.push( yyBraceDepth ); + } + break; + case 'e': + yyCh = getChar(); + if ( yyCh == 'l' ) { + // elif, else + if ( !yySavedBraceDepth.isEmpty() ) + yyBraceDepth = yySavedBraceDepth.top(); + } else if ( yyCh == 'n' ) { + // endif + if ( !yySavedBraceDepth.isEmpty() ) + yySavedBraceDepth.pop(); + } + } + while ( isalnum(yyCh) || yyCh == '_' ) + yyCh = getChar(); + break; + case '/': + yyCh = getChar(); + if ( yyCh == '/' ) { + do { + yyCh = getChar(); + } while ( yyCh != EOF && yyCh != '\n' ); + } else if ( yyCh == '*' ) { + bool metAster = FALSE; + bool metAsterSlash = FALSE; + + while ( !metAsterSlash ) { + yyCh = getChar(); + if ( yyCh == EOF ) { + fprintf( stderr, + "%s: Unterminated C++ comment starting at" + " line %d\n", + (const char *) yyFileName, yyLineNo ); + yyComment[yyCommentLen] = '\0'; + return Tok_Comment; + } + if ( yyCommentLen < sizeof(yyComment) - 1 ) + yyComment[yyCommentLen++] = (char) yyCh; + + if ( yyCh == '*' ) + metAster = TRUE; + else if ( metAster && yyCh == '/' ) + metAsterSlash = TRUE; + else + metAster = FALSE; + } + yyCh = getChar(); + yyCommentLen -= 2; + yyComment[yyCommentLen] = '\0'; + return Tok_Comment; + } + break; + case '"': + yyCh = getChar(); + + while ( yyCh != EOF && yyCh != '\n' && yyCh != '"' ) { + if ( yyCh == '\\' ) { + yyCh = getChar(); + + if ( yyCh == '\n' ) { + yyCh = getChar(); + } else if ( yyCh == 'x' ) { + QCString hex = "0"; + + yyCh = getChar(); + while ( isxdigit(yyCh) ) { + hex += (char) yyCh; + yyCh = getChar(); + } + sscanf( hex, "%x", &n ); + if ( yyStringLen < sizeof(yyString) - 1 ) + yyString[yyStringLen++] = (char) n; + } else if ( yyCh >= '0' && yyCh < '8' ) { + QCString oct = ""; + + do { + oct += (char) yyCh; + yyCh = getChar(); + } while ( yyCh >= '0' && yyCh < '8' ); + sscanf( oct, "%o", &n ); + if ( yyStringLen < sizeof(yyString) - 1 ) + yyString[yyStringLen++] = (char) n; + } else { + const char *p = strchr( tab, yyCh ); + if ( yyStringLen < sizeof(yyString) - 1 ) + yyString[yyStringLen++] = ( p == 0 ) ? + (char) yyCh : backTab[p - tab]; + yyCh = getChar(); + } + } else { + if ( yyStringLen < sizeof(yyString) - 1 ) + yyString[yyStringLen++] = (char) yyCh; + yyCh = getChar(); + } + } + yyString[yyStringLen] = '\0'; + + if ( yyCh != '"' ) + qWarning( "%s:%d: Unterminated C++ string", + (const char *) yyFileName, yyLineNo ); + + if ( yyCh == EOF ) { + return Tok_Eof; + } else { + yyCh = getChar(); + return Tok_String; + } + break; + case '-': + yyCh = getChar(); + if ( yyCh == '>' ) { + yyCh = getChar(); + return Tok_Arrow; + } + break; + case ':': + yyCh = getChar(); + if ( yyCh == ':' ) { + yyCh = getChar(); + return Tok_Gulbrandsen; + } + return Tok_Colon; + case '\'': + yyCh = getChar(); + if ( yyCh == '\\' ) + yyCh = getChar(); + + do { + yyCh = getChar(); + } while ( yyCh != EOF && yyCh != '\'' ); + yyCh = getChar(); + break; + case '{': + yyBraceDepth++; + yyCh = getChar(); + return Tok_LeftBrace; + case '}': + yyBraceDepth--; + yyCh = getChar(); + return Tok_RightBrace; + case '(': + yyParenDepth++; + yyCh = getChar(); + return Tok_LeftParen; + case ')': + yyParenDepth--; + yyCh = getChar(); + return Tok_RightParen; + case ',': + yyCh = getChar(); + return Tok_Comma; + case ';': + yyCh = getChar(); + return Tok_Semicolon; + default: + yyCh = getChar(); + } + } + } + return Tok_Eof; +} + +/* + The second part of this source file is the parser. It accomplishes + a very easy task: It finds all strings inside a tr() or translate() + call, and possibly finds out the context of the call. It supports + three cases: (1) the context is specified, as in + FunnyDialog::tr("Hello") or translate("FunnyDialog", "Hello"); + (2) the call appears within an inlined function; (3) the call + appears within a function defined outside the class definition. +*/ + +static int yyTok; + +static bool match( int t ) +{ + bool matches = ( yyTok == t ); + if ( matches ) + yyTok = getToken(); + return matches; +} + +static bool matchString( QCString *s ) +{ + bool matches = ( yyTok == Tok_String ); + *s = ""; + while ( yyTok == Tok_String ) { + *s += yyString; + yyTok = getToken(); + } + return matches; +} + +static bool matchEncoding( bool *utf8 ) +{ + if ( yyTok == Tok_Ident ) { + if ( strcmp(yyIdent, "QApplication") == 0 ) { + yyTok = getToken(); + if ( yyTok == Tok_Gulbrandsen ) + yyTok = getToken(); + } + *utf8 = QString( yyIdent ).endsWith( QString("UTF8") ); + yyTok = getToken(); + return TRUE; + } else { + return FALSE; + } +} + +static void parse( MetaTranslator *tor, const char *initialContext, + const char *defaultContext ) +{ + QMap<QCString, QCString> qualifiedContexts; + QStringList namespaces; + QCString context; + QCString text; + QCString com; + QCString functionContext = initialContext; + QCString prefix; + bool utf8 = FALSE; + bool missing_Q_OBJECT = FALSE; + + yyTok = getToken(); + while ( yyTok != Tok_Eof ) { + switch ( yyTok ) { + case Tok_class: + /* + Partial support for inlined functions. + */ + yyTok = getToken(); + if ( yyBraceDepth == (int) namespaces.count() && + yyParenDepth == 0 ) { + do { + /* + This code should execute only once, but we play + safe with impure definitions such as + 'class Q_EXPORT QMessageBox', in which case + 'QMessageBox' is the class name, not 'Q_EXPORT'. + */ + functionContext = yyIdent; + yyTok = getToken(); + } while ( yyTok == Tok_Ident ); + + while ( yyTok == Tok_Gulbrandsen ) { + yyTok = getToken(); + functionContext += "::"; + functionContext += yyIdent; + yyTok = getToken(); + } + + if ( yyTok == Tok_Colon ) { + missing_Q_OBJECT = TRUE; + } else { + functionContext = defaultContext; + } + } + break; + case Tok_namespace: + yyTok = getToken(); + if ( yyTok == Tok_Ident ) { + QCString ns = yyIdent; + yyTok = getToken(); + if ( yyTok == Tok_LeftBrace && + yyBraceDepth == (int) namespaces.count() + 1 ) + namespaces.append( QString(ns) ); + } + break; + case Tok_tr: + case Tok_trUtf8: + utf8 = ( yyTok == Tok_trUtf8 ); + yyTok = getToken(); + if ( match(Tok_LeftParen) && matchString(&text) ) { + com = ""; + if ( match(Tok_RightParen) || (match(Tok_Comma) && + matchString(&com) && match(Tok_RightParen)) ) { + if ( prefix.isNull() ) { + context = functionContext; + if ( !namespaces.isEmpty() ) + context.prepend( (namespaces.join(QString("::")) + + QString("::")).latin1() ); + } else { + context = prefix; + } + prefix = (const char *) 0; + + if ( qualifiedContexts.contains(context) ) + context = qualifiedContexts[context]; + tor->insert( MetaTranslatorMessage(context, text, com, + QString::null, utf8) ); + + if ( lacks_Q_OBJECT.contains(context) ) { + qWarning( "%s:%d: Class '%s' lacks Q_OBJECT macro", + (const char *) yyFileName, yyLineNo, + (const char *) context ); + lacks_Q_OBJECT.remove( context ); + } else { + needs_Q_OBJECT.insert( context, 0 ); + } + } + } + break; + case Tok_translate: + utf8 = FALSE; + yyTok = getToken(); + if ( match(Tok_LeftParen) && + matchString(&context) && + match(Tok_Comma) && + matchString(&text) ) { + com = ""; + if ( match(Tok_RightParen) || + (match(Tok_Comma) && + matchString(&com) && + (match(Tok_RightParen) || + match(Tok_Comma) && + matchEncoding(&utf8) && + match(Tok_RightParen))) ) + tor->insert( MetaTranslatorMessage(context, text, com, + QString::null, utf8) ); + } + break; + case Tok_Q_OBJECT: + missing_Q_OBJECT = FALSE; + yyTok = getToken(); + break; + case Tok_Ident: + if ( !prefix.isNull() ) + prefix += "::"; + prefix += yyIdent; + yyTok = getToken(); + if ( yyTok != Tok_Gulbrandsen ) + prefix = (const char *) 0; + break; + case Tok_Comment: + com = yyComment; + com = com.simplifyWhiteSpace(); + if ( com.left(sizeof(MagicComment) - 1) == MagicComment ) { + com.remove( 0, sizeof(MagicComment) - 1 ); + int k = com.find( ' ' ); + if ( k == -1 ) { + context = com; + } else { + context = com.left( k ); + com.remove( 0, k + 1 ); + tor->insert( MetaTranslatorMessage(context, "", com, + QString::null, FALSE) ); + } + + /* + Provide a backdoor for people using "using + namespace". See the manual for details. + */ + k = 0; + while ( (k = context.find("::", k)) != -1 ) { + qualifiedContexts.insert( context.mid(k + 2), context ); + k++; + } + } + yyTok = getToken(); + break; + case Tok_Arrow: + yyTok = getToken(); + if ( yyTok == Tok_tr || yyTok == Tok_trUtf8 ) + qWarning( "%s:%d: Cannot invoke tr() like this", + (const char *) yyFileName, yyLineNo ); + break; + case Tok_Gulbrandsen: + // at top level? + if ( yyBraceDepth == (int) namespaces.count() && yyParenDepth == 0 ) + functionContext = prefix; + yyTok = getToken(); + break; + case Tok_RightBrace: + case Tok_Semicolon: + if ( yyBraceDepth >= 0 && + yyBraceDepth + 1 == (int) namespaces.count() ) + namespaces.remove( namespaces.fromLast() ); + if ( yyBraceDepth == (int) namespaces.count() ) { + if ( missing_Q_OBJECT ) { + if ( needs_Q_OBJECT.contains(functionContext) ) { + qWarning( "%s:%d: Class '%s' lacks Q_OBJECT macro", + (const char *) yyFileName, yyLineNo, + (const char *) functionContext ); + } else { + lacks_Q_OBJECT.insert( functionContext, 0 ); + } + } + functionContext = defaultContext; + missing_Q_OBJECT = FALSE; + } + yyTok = getToken(); + break; + default: + yyTok = getToken(); + } + } + + if ( yyBraceDepth != 0 ) + fprintf( stderr, + "%s: Unbalanced braces in C++ code (or abuse of the C++" + " preprocessor)\n", + (const char *) yyFileName ); + if ( yyParenDepth != 0 ) + fprintf( stderr, + "%s: Unbalanced parentheses in C++ code (or abuse of the C++" + " preprocessor)\n", + (const char *) yyFileName ); +} + +void fetchtr_cpp( const char *fileName, MetaTranslator *tor, + const char *defaultContext, bool mustExist ) +{ + yyInFile = fopen( fileName, "r" ); + if ( yyInFile == 0 ) { + if ( mustExist ) + fprintf( stderr, + "lupdate error: Cannot open C++ source file '%s': %s\n", + fileName, strerror(errno) ); + return; + } + + startTokenizer( fileName, getCharFromFile ); + parse( tor, 0, defaultContext ); + fclose( yyInFile ); +} + +/* + In addition to C++, we support Qt Designer UI files. +*/ + +/* + Fetches tr() calls in C++ code in UI files (inside "<function>" + tag). This mechanism is obsolete. +*/ +void fetchtr_inlined_cpp( const char *fileName, const QString& in, + MetaTranslator *tor, const char *context ) +{ + yyInStr = in; + startTokenizer( fileName, getCharFromString ); + parse( tor, context, 0 ); + yyInStr = QString::null; +} + +class UiHandler : public QXmlDefaultHandler +{ +public: + UiHandler( MetaTranslator *translator, const char *fileName ) + : tor( translator ), fname( fileName ), comment( "" ) { } + + virtual bool startElement( const QString& namespaceURI, + const QString& localName, const QString& qName, + const QXmlAttributes& atts ); + virtual bool endElement( const QString& namespaceURI, + const QString& localName, const QString& qName ); + virtual bool characters( const QString& ch ); + virtual bool fatalError( const QXmlParseException& exception ); + +private: + void flush(); + + MetaTranslator *tor; + QCString fname; + QString context; + QString source; + QString comment; + + QString accum; +}; + +bool UiHandler::startElement( const QString& /* namespaceURI */, + const QString& /* localName */, + const QString& qName, + const QXmlAttributes& atts ) +{ + if ( qName == QString("item") ) { + flush(); + if ( !atts.value(QString("text")).isEmpty() ) + source = atts.value( QString("text") ); + } else if ( qName == QString("string") ) { + flush(); + } + accum.truncate( 0 ); + return TRUE; +} + +bool UiHandler::endElement( const QString& /* namespaceURI */, + const QString& /* localName */, + const QString& qName ) +{ + accum.replace( QRegExp(QString("\r\n")), "\n" ); + + if ( qName == QString("class") ) { + if ( context.isEmpty() ) + context = accum; + } else if ( qName == QString("string") ) { + source = accum; + } else if ( qName == QString("comment") ) { + comment = accum; + flush(); + } else if ( qName == QString("function") ) { + fetchtr_inlined_cpp( (const char *) fname, accum, tor, + context.latin1() ); + } else { + flush(); + } + return TRUE; +} + +bool UiHandler::characters( const QString& ch ) +{ + accum += ch; + return TRUE; +} + +bool UiHandler::fatalError( const QXmlParseException& exception ) +{ + QString msg; + msg.sprintf( "Parse error at line %d, column %d (%s).", + exception.lineNumber(), exception.columnNumber(), + exception.message().latin1() ); + fprintf( stderr, "XML error: %s\n", msg.latin1() ); + return FALSE; +} + +void UiHandler::flush() +{ + if ( !context.isEmpty() && !source.isEmpty() ) + tor->insert( MetaTranslatorMessage(context.utf8(), source.utf8(), + comment.utf8(), QString::null, + TRUE) ); + source.truncate( 0 ); + comment.truncate( 0 ); +} + +void fetchtr_ui( const char *fileName, MetaTranslator *tor, + const char * /* defaultContext */, bool mustExist ) +{ + QFile f( fileName ); + if ( !f.open(IO_ReadOnly) ) { + if ( mustExist ) + fprintf( stderr, "lupdate error: cannot open UI file '%s': %s\n", + fileName, strerror(errno) ); + return; + } + + QTextStream t( &f ); + QXmlInputSource in( t ); + QXmlSimpleReader reader; + reader.setFeature( "http://xml.org/sax/features/namespaces", FALSE ); + reader.setFeature( "http://xml.org/sax/features/namespace-prefixes", TRUE ); + reader.setFeature( "http://trolltech.com/xml/features/report-whitespace" + "-only-CharData", FALSE ); + QXmlDefaultHandler *hand = new UiHandler( tor, fileName ); + reader.setContentHandler( hand ); + reader.setErrorHandler( hand ); + + if ( !reader.parse(in) ) + fprintf( stderr, "%s: Parse error in UI file\n", fileName ); + reader.setContentHandler( 0 ); + reader.setErrorHandler( 0 ); + delete hand; + f.close(); +} diff --git a/development/translation/opie-lupdate/lupdate b/development/translation/opie-lupdate/lupdate Binary files differnew file mode 100755 index 0000000..0e1604d --- a/dev/null +++ b/development/translation/opie-lupdate/lupdate diff --git a/development/translation/opie-lupdate/main.cpp b/development/translation/opie-lupdate/main.cpp new file mode 100644 index 0000000..ce65e7a --- a/dev/null +++ b/development/translation/opie-lupdate/main.cpp @@ -0,0 +1,215 @@ +/********************************************************************** +** Copyright (C) 2000-2002 Trolltech AS. All rights reserved. +** Copyright (C) 2003 zecke +** +** This file is part of Qt Linguist. +** +** This file may be distributed and/or modified under the terms of the +** GNU General Public License version 2 as published by the Free Software +** Foundation and appearing in the file LICENSE.GPL included in the +** packaging of this file. +** +** This file is provided AS IS with NO WARRANTY OF ANY KIND, INCLUDING THE +** WARRANTY OF DESIGN, MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. +** +** See http://www.trolltech.com/gpl/ for GPL licensing information. +** +** Contact info@trolltech.com if any conditions of this licensing are +** not clear to you. +** +**********************************************************************/ + +#include <metatranslator.h> +#include <proparser.h> +#include <opie.h> + +#include <qfile.h> +#include <qfileinfo.h> +#include <qstring.h> +#include <qstringlist.h> +#include <qtextstream.h> + +#include <errno.h> +#include <string.h> + +// defined in fetchtr.cpp +extern void fetchtr_cpp( const char *fileName, MetaTranslator *tor, + const char *defaultContext, bool mustExist ); +extern void fetchtr_ui( const char *fileName, MetaTranslator *tor, + const char *defaultContext, bool mustExist ); + +// defined in merge.cpp +extern void merge( MetaTranslator *tor, const MetaTranslator *virginTor, + bool verbose ); + +typedef QValueList<MetaTranslatorMessage> TML; + +static const char* LUPDATE_VERSION = "0.1"; + +static void printUsage() +{ + fprintf( stderr, "Usage:\n" + " opie-lupdate [options] project-file\n" + " opie-lupdate [options] source-files -ts ts-files\n" + "Options:\n" + " -opie The OPIE base dir if not supplied $OPIEDIR will be taken\n" + " -help Display this information and exit\n" + " -noobsolete\n" + " Drop all obsolete strings\n" + " -verbose\n" + " Explain what is being done\n" + " -version\n" + " Display the version of lupdate and exit\n" ); +} + +static void updateTsFiles( const MetaTranslator& fetchedTor, + const QString& opiedir, + const QStringList& languages, + const QString& basename, + const QString& codec, + bool noObsolete, bool verbose ) +{ + QStringList::ConstIterator it = languages.begin(); + for ( ; it != languages.end(); ++it ) { + QString fileName = opiedir + "/i18n/" + (*it) + "/" + basename; + MetaTranslator tor; + tor.load( fileName ); + if ( !codec.isEmpty() ) + tor.setCodec( codec ); + if ( verbose ) + fprintf( stderr, "Updating '%s'...\n", fileName.latin1() ); + merge( &tor, &fetchedTor, verbose ); + if ( noObsolete ) + tor.stripObsoleteMessages(); + tor.stripEmptyContexts(); + if ( !tor.save(fileName) ) + fprintf( stderr, "lupdate error: Cannot save '%s': %s\n", + fileName.latin1(), strerror(errno) ); + } +} + +int main( int argc, char **argv ) +{ + QString defaultContext = "@default"; + MetaTranslator fetchedTor; + QCString codec; + QStringList tsFileNames; + QString opiedir; + QString translationBase; + QString target; + + bool verbose = FALSE; + bool noObsolete = FALSE; + bool metSomething = FALSE; + bool isLib = FALSE; + int numFiles = 0; + + int i; + + QStringList languageList = OPIE::self()->languageList(opiedir); + + for ( i = 1; i < argc; i++ ) { + if ( qstrcmp(argv[i], "-help") == 0 ) { + printUsage(); + return 0; + } else if ( qstrcmp(argv[i], "-noobsolete") == 0 ) { + noObsolete = TRUE; + continue; + } else if ( qstrcmp(argv[i], "-verbose") == 0 ) { + verbose = TRUE; + continue; + } else if ( qstrcmp(argv[i], "-version") == 0 ) { + fprintf( stderr, "lupdate version %s\n", LUPDATE_VERSION ); + return 0; + } else if ( qstrcmp(argv[i], "-opie") == 0 ) { + if( i+1 < argc ) { + opiedir = argv[i+1]; + languageList = OPIE::self()->languageList(opiedir); + } + i++; // UGLY but we want to skip the next argument + continue; + } + + numFiles++; + + QString fullText; + + QFile f( argv[i] ); + if ( !f.open(IO_ReadOnly) ) { + fprintf( stderr, "lupdate error: Cannot open file '%s': %s\n", + argv[i], strerror(errno) ); + return 1; + } + + QTextStream t( &f ); + fullText = t.read(); + f.close(); + + fetchedTor = MetaTranslator(); + codec.truncate( 0 ); + tsFileNames.clear(); + isLib = FALSE; + + QMap<QString, QString> tagMap = proFileTagMap( fullText ); + QMap<QString, QString>::Iterator it; + + for ( it = tagMap.begin(); it != tagMap.end(); ++it ) { + QStringList toks = QStringList::split( ' ', it.data() ); + QStringList::Iterator t; + + for ( t = toks.begin(); t != toks.end(); ++t ) { + if ( it.key() == "HEADERS" || it.key() == "SOURCES" ) { + fetchtr_cpp( *t, &fetchedTor, defaultContext, TRUE ); + metSomething = TRUE; + } else if ( it.key() == "INTERFACES" || + it.key() == "FORMS" ) { + fetchtr_ui( *t, &fetchedTor, defaultContext, TRUE ); + fetchtr_cpp( *t + ".h", &fetchedTor, defaultContext, + FALSE ); + metSomething = TRUE; + } else if ( it.key() == "TRANSLATIONS" ) { + // we do not care for that attribute anymore + //tsFileNames.append( *t ); + metSomething = TRUE; + } else if ( it.key() == "CODEC" ) { + codec = (*t).latin1(); + } else if ( it.key() == "TARGET" ) { + target = *t; + metSomething = TRUE; + } else if ( it.key() == "TEMPLATE" ) { + if ( (*t).stripWhiteSpace().lower() == "lib" ) + isLib = true; + } + } + } + /** + * We know the $OPIEDIR or have opiedir + * we've a list of languages (de,en,gb,foo,bar) + * we've got the TARGET and we no it's the lib + * so let's do that + * $OPIEDIR/language[i]/ifLibAppendLib$TARGET.ts + */ + qWarning("TARGET %s IsLib:%d", target.latin1(), isLib ); + qWarning("LANGS %s", languageList.join(";").latin1() ); + qWarning("OPIEDIR %s", OPIE::self()->opieDir(opiedir).latin1() ); + if (isLib ) + target.prepend("lib"); + target += ".ts"; + updateTsFiles( fetchedTor, OPIE::self()->opieDir(opiedir), + languageList, target, codec, noObsolete, verbose ); + + if ( !metSomething ) { + fprintf( stderr, + "lupdate warning: File '%s' does not look like a" + " project file\n", + argv[i] ); + } + + } + + if ( numFiles == 0 ) { + printUsage(); + return 1; + } + return 0; +} diff --git a/development/translation/opie-lupdate/merge.cpp b/development/translation/opie-lupdate/merge.cpp new file mode 100644 index 0000000..a96104e --- a/dev/null +++ b/development/translation/opie-lupdate/merge.cpp @@ -0,0 +1,115 @@ +/********************************************************************** +** Copyright (C) 2000 Trolltech AS. All rights reserved. +** +** This file is part of Qt Linguist. +** +** This file may be distributed and/or modified under the terms of the +** GNU General Public License version 2 as published by the Free Software +** Foundation and appearing in the file LICENSE.GPL included in the +** packaging of this file. +** +** This file is provided AS IS with NO WARRANTY OF ANY KIND, INCLUDING THE +** WARRANTY OF DESIGN, MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. +** +** See http://www.trolltech.com/gpl/ for GPL licensing information. +** +** Contact info@trolltech.com if any conditions of this licensing are +** not clear to you. +** +**********************************************************************/ + +#include <metatranslator.h> + +// defined in numberh.cpp +extern void applyNumberHeuristic( MetaTranslator *tor, bool verbose ); +// defined in sametexth.cpp +extern void applySameTextHeuristic( MetaTranslator *tor, bool verbose ); + +typedef QValueList<MetaTranslatorMessage> TML; + +/* + Merges two MetaTranslator objects into the first one. The first one + is a set of source texts and translations for a previous version of + the internationalized program; the second one is a set of fresh + source texts newly extracted from the source code, without any + translation yet. +*/ + +void merge( MetaTranslator *tor, const MetaTranslator *virginTor, bool verbose ) +{ + int known = 0; + int neww = 0; + int obsoleted = 0; + TML all = tor->messages(); + TML::Iterator it; + + /* + The types of all the messages from the vernacular translator + are updated according to the virgin translator. + */ + for ( it = all.begin(); it != all.end(); ++it ) { + MetaTranslatorMessage::Type newType; + MetaTranslatorMessage m = *it; + + // skip context comment + if ( !QCString((*it).sourceText()).isEmpty() ) { + if ( !virginTor->contains((*it).context(), (*it).sourceText(), + (*it).comment()) ) { + newType = MetaTranslatorMessage::Obsolete; + if ( m.type() != MetaTranslatorMessage::Obsolete ) + obsoleted++; + } else { + switch ( m.type() ) { + case MetaTranslatorMessage::Finished: + newType = MetaTranslatorMessage::Finished; + known++; + break; + case MetaTranslatorMessage::Unfinished: + newType = MetaTranslatorMessage::Unfinished; + known++; + break; + case MetaTranslatorMessage::Obsolete: + newType = MetaTranslatorMessage::Unfinished; + neww++; + } + } + + if ( newType != m.type() ) { + m.setType( newType ); + tor->insert( m ); + } + } + } + + /* + Messages found only in the virgin translator are added to the + vernacular translator. Among these are all the context comments. + */ + all = virginTor->messages(); + + for ( it = all.begin(); it != all.end(); ++it ) { + if ( !tor->contains((*it).context(), (*it).sourceText(), + (*it).comment()) ) { + tor->insert( *it ); + if ( !QCString((*it).sourceText()).isEmpty() ) + neww++; + } + } + + /* + The same-text heuristic handles cases where a message has an + obsolete counterpart with a different context or comment. + */ + applySameTextHeuristic( tor, verbose ); + + /* + The number heuristic handles cases where a message has an + obsolete counterpart with mostly numbers differing in the + source text. + */ + applyNumberHeuristic( tor, verbose ); + + if ( verbose ) + fprintf( stderr, " %d known, %d new and %d obsoleted messages\n", known, + neww, obsoleted ); +} diff --git a/development/translation/opie-lupdate/numberh.cpp b/development/translation/opie-lupdate/numberh.cpp new file mode 100644 index 0000000..f7b7bf8 --- a/dev/null +++ b/development/translation/opie-lupdate/numberh.cpp @@ -0,0 +1,235 @@ +/********************************************************************** +** Copyright (C) 2000-2002 Trolltech AS. All rights reserved. +** +** This file is part of Qt Linguist. +** +** This file may be distributed and/or modified under the terms of the +** GNU General Public License version 2 as published by the Free Software +** Foundation and appearing in the file LICENSE.GPL included in the +** packaging of this file. +** +** This file is provided AS IS with NO WARRANTY OF ANY KIND, INCLUDING THE +** WARRANTY OF DESIGN, MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. +** +** See http://www.trolltech.com/gpl/ for GPL licensing information. +** +** Contact info@trolltech.com if any conditions of this licensing are +** not clear to you. +** +**********************************************************************/ + +#include <metatranslator.h> + +#include <qmemarray.h> +#include <qcstring.h> +#include <qmap.h> +#include <qstringlist.h> + +#include <ctype.h> + +typedef QMap<QCString, MetaTranslatorMessage> TMM; +typedef QValueList<MetaTranslatorMessage> TML; + +static bool isDigitFriendly( int c ) +{ + return ispunct( c ) || isspace( c ); +} + +static int numberLength( const char *s ) +{ + int i = 0; + + if ( isdigit(s[0]) ) { + do { + i++; + } while ( isdigit(s[i]) || + (isDigitFriendly(s[i]) && + (isdigit(s[i + 1]) || + (isDigitFriendly(s[i + 1]) && isdigit(s[i + 2])))) ); + } + return i; +} + +/* + Returns a version of 'key' where all numbers have been replaced by zeroes. If + there were none, returns "". +*/ +static QCString zeroKey( const char *key ) +{ + QCString zeroed( strlen(key) + 1 ); + char *z = zeroed.data(); + int i = 0, j = 0; + int len; + bool metSomething = FALSE; + + while ( key[i] != '\0' ) { + len = numberLength( key + i ); + if ( len > 0 ) { + i += len; + z[j++] = '0'; + metSomething = TRUE; + } else { + z[j++] = key[i++]; + } + } + z[j] = '\0'; + + if ( metSomething ) + return zeroed; + else + return ""; +} + +static QString translationAttempt( const QString& oldTranslation, + const char *oldSource, + const char *newSource ) +{ + int p = zeroKey( oldSource ).contains( '0' ); + int oldSourceLen = qstrlen( oldSource ); + QString attempt; + QStringList oldNumbers; + QStringList newNumbers; + QMemArray<bool> met( p ); + QMemArray<int> matchedYet( p ); + int i, j; + int k = 0, ell, best; + int m, n; + int pass; + + /* + This algorithm is hard to follow, so we'll consider an example + all along: oldTranslation is "XeT 3.0", oldSource is "TeX 3.0" + and newSource is "XeT 3.1". + + First, we set up two tables: oldNumbers and newNumbers. In our + example, oldNumber[0] is "3.0" and newNumber[0] is "3.1". + */ + for ( i = 0, j = 0; i < oldSourceLen; i++, j++ ) { + m = numberLength( oldSource + i ); + n = numberLength( newSource + j ); + if ( m > 0 ) { + oldNumbers.append( QCString(oldSource + i, m + 1) ); + newNumbers.append( QCString(newSource + j, n + 1) ); + i += m; + j += n; + met[k] = FALSE; + matchedYet[k] = 0; + k++; + } + } + + /* + We now go over the old translation, "XeT 3.0", one letter at a + time, looking for numbers found in oldNumbers. Whenever such a + number is met, it is replaced with its newNumber equivalent. In + our example, the "3.0" of "XeT 3.0" becomes "3.1". + */ + for ( i = 0; i < (int) oldTranslation.length(); i++ ) { + attempt += oldTranslation[i]; + for ( k = 0; k < p; k++ ) { + if ( oldTranslation[i] == oldNumbers[k][matchedYet[k]] ) + matchedYet[k]++; + else + matchedYet[k] = 0; + } + + /* + Let's find out if the last character ended a match. We make + two passes over the data. In the first pass, we try to + match only numbers that weren't matched yet; if that fails, + the second pass does the trick. This is useful in some + suspicious cases, flagged below. + */ + for ( pass = 0; pass < 2; pass++ ) { + best = p; // an impossible value + for ( k = 0; k < p; k++ ) { + if ( (!met[k] || pass > 0) && + matchedYet[k] == (int) oldNumbers[k].length() && + numberLength(oldTranslation.latin1() + (i + 1) - + matchedYet[k]) == matchedYet[k] ) { + // the longer the better + if ( best == p || matchedYet[k] > matchedYet[best] ) + best = k; + } + } + if ( best != p ) { + attempt.truncate( attempt.length() - matchedYet[best] ); + attempt += newNumbers[best]; + met[best] = TRUE; + for ( k = 0; k < p; k++ ) + matchedYet[k] = 0; + break; + } + } + } + + /* + We flag two kinds of suspicious cases. They are identified as + such with comments such as "{2000?}" at the end. + + Example of the first kind: old source text "TeX 3.0" translated + as "XeT 2.0" is flagged "TeX 2.0 {3.0?}", no matter what the + new text is. + */ + for ( k = 0; k < p; k++ ) { + if ( !met[k] ) + attempt += QString( " {" ) + newNumbers[k] + QString( "?}" ); + } + + /* + Example of the second kind: "1 of 1" translated as "1 af 1", + with new source text "1 of 2", generates "1 af 2 {1 or 2?}" + because it's not clear which of "1 af 2" and "2 af 1" is right. + */ + for ( k = 0; k < p; k++ ) { + for ( ell = 0; ell < p; ell++ ) { + if ( k != ell && oldNumbers[k] == oldNumbers[ell] && + newNumbers[k] < newNumbers[ell] ) + attempt += QString( " {" ) + newNumbers[k] + QString( " or " ) + + newNumbers[ell] + QString( "?}" ); + } + } + return attempt; +} + +/* + Augments a MetaTranslator with translations easily derived from + similar existing (probably obsolete) translations. + + For example, if "TeX 3.0" is translated as "XeT 3.0" and "TeX 3.1" + has no translation, "XeT 3.1" is added to the translator and is + marked Unfinished. +*/ +void applyNumberHeuristic( MetaTranslator *tor, bool verbose ) +{ + TMM translated, untranslated; + TMM::Iterator t, u; + TML all = tor->messages(); + TML::Iterator it; + int inserted = 0; + + for ( it = all.begin(); it != all.end(); ++it ) { + if ( (*it).type() == MetaTranslatorMessage::Unfinished ) { + if ( (*it).translation().isEmpty() ) + untranslated.insert( zeroKey((*it).sourceText()), *it ); + } else if ( !(*it).translation().isEmpty() ) { + translated.insert( zeroKey((*it).sourceText()), *it ); + } + } + + for ( u = untranslated.begin(); u != untranslated.end(); ++u ) { + t = translated.find( u.key() ); + if ( t != translated.end() && !t.key().isEmpty() && + qstrcmp((*t).sourceText(), (*u).sourceText()) != 0 ) { + MetaTranslatorMessage m( *u ); + m.setTranslation( translationAttempt((*t).translation(), + (*t).sourceText(), + (*u).sourceText()) ); + tor->insert( m ); + inserted++; + } + } + if ( verbose && inserted != 0 ) + fprintf( stderr, " number heuristic provided %d translation%s\n", + inserted, inserted == 1 ? "" : "s" ); +} diff --git a/development/translation/opie-lupdate/opie-lupdate.pro b/development/translation/opie-lupdate/opie-lupdate.pro new file mode 100644 index 0000000..ea51c8b --- a/dev/null +++ b/development/translation/opie-lupdate/opie-lupdate.pro @@ -0,0 +1,21 @@ +TEMPLATE = app +CONFIG += qt warn_on console +HEADERS = ../shared/metatranslator.h \ + ../shared/proparser.h \ + ../shared/opie.h + +SOURCES = fetchtr.cpp \ + main.cpp \ + merge.cpp \ + numberh.cpp \ + sametexth.cpp \ + ../shared/metatranslator.cpp \ + ../shared/proparser.cpp \ + ../shared/opie.cpp + +DEFINES += QT_INTERNAL_XML + +TARGET = opie-lupdate +INCLUDEPATH += ../shared +#DESTDIR = + diff --git a/development/translation/opie-lupdate/sametexth.cpp b/development/translation/opie-lupdate/sametexth.cpp new file mode 100644 index 0000000..574cfd5 --- a/dev/null +++ b/development/translation/opie-lupdate/sametexth.cpp @@ -0,0 +1,84 @@ +/********************************************************************** +** Copyright (C) 2000-2002 Trolltech AS. All rights reserved. +** +** This file is part of Qt Linguist. +** +** This file may be distributed and/or modified under the terms of the +** GNU General Public License version 2 as published by the Free Software +** Foundation and appearing in the file LICENSE.GPL included in the +** packaging of this file. +** +** This file is provided AS IS with NO WARRANTY OF ANY KIND, INCLUDING THE +** WARRANTY OF DESIGN, MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. +** +** See http://www.trolltech.com/gpl/ for GPL licensing information. +** +** Contact info@trolltech.com if any conditions of this licensing are +** not clear to you. +** +**********************************************************************/ + +#include <metatranslator.h> + +#include <qcstring.h> +#include <qmap.h> + +typedef QMap<QCString, MetaTranslatorMessage> TMM; +typedef QValueList<MetaTranslatorMessage> TML; + +/* + Augments a MetaTranslator with trivially derived translations. + + For example, if "Enabled:" is consistendly translated as "Eingeschaltet:" no + matter the context or the comment, "Eingeschaltet:" is added as the + translation of any untranslated "Enabled:" text and is marked Unfinished. +*/ + +void applySameTextHeuristic( MetaTranslator *tor, bool verbose ) +{ + TMM translated; + TMM avoid; + TMM::Iterator t; + TML untranslated; + TML::Iterator u; + TML all = tor->messages(); + TML::Iterator it; + int inserted = 0; + + for ( it = all.begin(); it != all.end(); ++it ) { + if ( (*it).type() == MetaTranslatorMessage::Unfinished ) { + if ( (*it).translation().isEmpty() ) + untranslated.append( *it ); + } else { + QCString key = (*it).sourceText(); + t = translated.find( key ); + if ( t != translated.end() ) { + /* + The same source text is translated at least two + different ways. Do nothing then. + */ + if ( (*t).translation() != (*it).translation() ) { + translated.remove( key ); + avoid.insert( key, *it ); + } + } else if ( !avoid.contains(key) && + !(*it).translation().isEmpty() ) { + translated.insert( key, *it ); + } + } + } + + for ( u = untranslated.begin(); u != untranslated.end(); ++u ) { + QCString key = (*u).sourceText(); + t = translated.find( key ); + if ( t != translated.end() ) { + MetaTranslatorMessage m( *u ); + m.setTranslation( (*t).translation() ); + tor->insert( m ); + inserted++; + } + } + if ( verbose && inserted != 0 ) + fprintf( stderr, " same-text heuristic provided %d translation%s\n", + inserted, inserted == 1 ? "" : "s" ); +} |