diff options
Diffstat (limited to 'poxml/po2xml.cpp')
| -rw-r--r-- | poxml/po2xml.cpp | 261 |
1 files changed, 261 insertions, 0 deletions
diff --git a/poxml/po2xml.cpp b/poxml/po2xml.cpp new file mode 100644 index 00000000..9e8bc1a5 --- /dev/null +++ b/poxml/po2xml.cpp @@ -0,0 +1,261 @@ + // #define POXML_DEBUG + +#include "parser.h" +#include <stdlib.h> +#include <iostream> +#include <assert.h> +#include <qregexp.h> + +#include <fstream> +#include "GettextLexer.hpp" +#include "GettextParser.hpp" +#include "antlr/AST.hpp" +#include "antlr/CommonAST.hpp" + +using namespace std; + +QString translate(QString xml, QString orig, QString translation) +{ + QString prefix; + while (xml.at(0) == '<' && orig.at(0) != '<') { + // a XML tag as prefix + int index = xml.find('>'); + assert(index != -1); + index++; + while (xml.at(index) == ' ') + index++; + prefix = prefix + xml.left(index); + xml = xml.mid(index, xml.length()); + } + + int index = xml.find(orig); + if (index == -1) { + qWarning("can't find\n%s\nin\n%s", orig.latin1(), xml.latin1()); + exit(1); + } + if (!translation.isEmpty()) + xml.replace(index, orig.length(), translation); + return prefix + xml; +} + +int main( int argc, char **argv ) +{ + if (argc != 3) { + qWarning("usage: %s english-XML translated-PO", argv[0]); + ::exit(1); + } + + MsgList english = parseXML(argv[1]); + MsgList translated; + + try { + ifstream s(argv[2]); + GettextLexer lexer(s); + GettextParser parser(lexer); + translated = parser.file(); + + } catch(exception& e) { + cerr << "exception: " << e.what() << endl; + return 1; + } + + QMap<QString, QString> translations; + for (MsgList::ConstIterator it = translated.begin(); + it != translated.end(); ++it) + { + QString msgstr; + QString msgid = escapePO((*it).msgid); + if ((*it).comment.find("fuzzy") < 0) + msgstr = escapePO((*it).msgstr); + +#ifdef POXML_DEBUG + qDebug("inserting translations '%s' -> '%s'", msgid.latin1(),msgstr.latin1()); +#endif + translations.insert(msgid, msgstr); + } + + QFile xml(argv[1]); + xml.open(IO_ReadOnly); + QTextStream ds(&xml); + ds.setEncoding(QTextStream::UnicodeUTF8); + QString xml_text = ds.read(); + xml.close(); + QString output; + QTextStream ts(&output, IO_WriteOnly); + StructureParser::cleanupTags(xml_text); + + QValueList<int> line_offsets; + line_offsets.append(0); + int index = 0; + while (true) { + index = xml_text.find('\n', index) + 1; + if (index <= 0) + break; + line_offsets.append(index); + } + + int old_start_line = -1, old_start_col = -1; + QString old_text; + MsgList::Iterator old_it = english.end(); + + for (MsgList::Iterator it = english.begin(); + it != english.end(); ++it) + { + BlockInfo bi = (*it).lines.first(); + int start_pos = line_offsets[bi.start_line - 1] + bi.start_col; + if (!bi.end_line) + continue; + int end_pos = line_offsets[bi.end_line - 1] + bi.end_col - 1; + + (*it).start = start_pos; + if (old_start_line == bi.start_line && + old_start_col == bi.start_col) + { + (*old_it).end = bi.offset; + (*it).end = end_pos; + } else { + (*it).lines.first().offset = 0; + (*it).end = 0; + } + + old_start_line = bi.start_line; + old_start_col = bi.start_col; + old_it = it; + } + + int old_pos = 0; + + for (MsgList::Iterator it = english.begin(); + it != english.end(); ++it) + { + BlockInfo bi = (*it).lines.first(); + int start_pos = line_offsets[bi.start_line - 1] + bi.start_col; + if (!bi.end_line) + continue; + int end_pos = line_offsets[bi.end_line - 1] + bi.end_col - 1; + + QString xml = xml_text.mid(start_pos, end_pos - start_pos); + int index = 0; + while (true) { + index = xml.find("<!--"); + if (index == -1) + break; + int end_index = index + 4; + while (xml.at(end_index) != '>' || + xml.at(end_index-1) != '-' || + xml.at(end_index-2) != '-') + { + end_index++; + } + xml.replace(index, end_index + 1 - index, " "); + index = end_index; + } + StructureParser::descape(xml); + + QString descaped = StructureParser::descapeLiterals((*it).msgid); + if (translations.contains(descaped)) + descaped = translations[descaped]; + +#ifdef POXML_DEBUG + // assert(!descaped.isEmpty()); +#endif + + if ((*it).msgid.at(0) == '<' && StructureParser::isClosure((*it).msgid)) { + // if the id starts with a tag, then we remembered the + // correct line information and need to strip the target + // now, so it fits + int index = 0; + while ((*it).msgid.at(index) != '>') + index++; + index++; + while ((*it).msgid.at(index) == ' ') + index++; + QString omsgid = (*it).msgid; + (*it).msgid = (*it).msgid.mid(index); + + index = (*it).msgid.length() - 1; + while ((*it).msgid.at(index) != '<') + index--; + + (*it).msgid = (*it).msgid.left(index); + + if (!descaped.isEmpty()) { + if (descaped.at(0) != '<') { + qWarning("the translation of '%s' doesn't start with a tag.", omsgid.latin1()); + exit(1); + } + index = 0; + while (index <= (int)descaped.length() && descaped.at(index) != '>') + index++; + index++; + while (descaped.at(index) == ' ') + index++; + descaped = descaped.mid(index); + + index = descaped.length() - 1; + while (index >= 0 && descaped.at(index) != '<') + index--; + + descaped = descaped.left(index); + } + } + +#ifdef POXML_DEBUG + qDebug("english \"%s\" ORIG \"%s\" %d(%d-%d) %d(%d-%d) %d %d TRANS \"%s\" %d '%s'", xml.latin1(), (*it).msgid.latin1(), + start_pos, bi.start_line, bi.start_col, + end_pos, bi.end_line, bi.end_col, + (*it).lines.first().offset, + (*it).end, + translations[(*it).msgid].latin1(), (*it).end, + descaped.latin1() + ); +#endif + + if ((*it).end) { + if (!(*it).lines.first().offset && end_pos != old_pos) { + assert(start_pos >= old_pos); + ts << xml_text.mid(old_pos, start_pos - old_pos); + } + assert((*it).end >= bi.offset); + ts << translate(xml.mid(bi.offset, (*it).end - bi.offset), + (*it).msgid, descaped); + old_pos = end_pos; + } else { + if (start_pos != old_pos) { + if (start_pos < old_pos) { + qDebug("so far: '%s'", output.latin1()); + } + assert(start_pos > old_pos); + ts << xml_text.mid(old_pos, start_pos - old_pos); + } + old_pos = end_pos; + ts << translate(xml, + (*it).msgid, descaped); + } + } + + ts << xml_text.mid(old_pos); + + output.replace(QRegExp("<trans_comment\\s*>"), ""); + output.replace(QRegExp("</trans_comment\\s*>"), ""); + + StructureParser::removeEmptyTags(output); + + index = 0; + while (true) { + index = output.find(QRegExp(">[^\n]"), index ); + if ( index == -1 ) + break; + if ( output.at( index - 1 ) == '/' || output.at( index - 1 ) == '-' || + output.at( index - 1 ) == ']' || output.at( index - 1 ) == '?' ) + index = index + 1; + else { + output.replace( index, 1, "\n>" ); + index = index + 2; + } + } + output = StructureParser::descapeLiterals(output); + + cout << output.utf8().data(); + return 0; +} |