You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
tellico/src/translators/tellico_xml.cpp

99 lines
4.0 KiB

/***************************************************************************
copyright : (C) 2003-2006 by Robby Stephenson
email : robby@periapsis.org
***************************************************************************/
/***************************************************************************
* *
* This program is free software; you can redistribute it and/or modify *
* it under the terms of version 2 of the GNU General Public License as *
* published by the Free Software Foundation; *
* *
***************************************************************************/
#include "tellico_xml.h"
#include <libxml/parserInternals.h> // needed for IS_LETTER
#include <libxml/parser.h> // has to be before valid.h
#include <libxml/valid.h>
#include <tqregexp.h>
const TQString Tellico::XML::nsXSL = TQString::fromLatin1("http://www.w3.org/1999/XSL/Transform");
const TQString Tellico::XML::nsBibtexml = TQString::fromLatin1("http://bibtexml.sf.net/");
const TQString Tellico::XML::dtdBibtexml = TQString::fromLatin1("bibtexml.dtd");
/*
* VERSION 2 added namespaces, changed to multiple elements,
* and changed the "keywords" field to "keyword"
*
* VERSION 3 broke out the formatFlag, and changed NoComplete to AllowCompletion
*
* VERSION 4 added a bibtex-field name for Bibtex collections, element name was
* changed to 'entry', field elements changed to 'field', and boolean fields are now "true"
*
* VERSION 5 moved the bibtex-field and any other extended field property to property elements
* inside the field element, and added the image element.
*
* VERSION 6 added id, i18n attributes, and year, month, day elements in date fields with a calendar name
* attribute.
*
* VERSION 7 changed the application name to Tellico, renamed unitTitle to entryTitle, and made the id permanent.
*
* VERSION 8 added loans and saved filters.
*
* VERSION 9 changed music collections to always have three columns by default, with title/artist/length and
* added file catalog collection.
*
* VERSION 10 added the game board collection.
*/
const uint Tellico::XML::syntaxVersion = 10;
const TQString Tellico::XML::nsTellico = TQString::fromLatin1("http://periapsis.org/tellico/");
const TQString Tellico::XML::nsBookcase = TQString::fromLatin1("http://periapsis.org/bookcase/");
const TQString Tellico::XML::nsDublinCore = TQString::fromLatin1("http://purl.org/dc/elements/1.1/");
const TQString Tellico::XML::nsZing = TQString::fromLatin1("http://www.loc.gov/zing/srw/");
const TQString Tellico::XML::nsZingDiag = TQString::fromLatin1("http://www.loc.gov/zing/srw/diagnostic/");
TQString Tellico::XML::pubTellico(int version) {
return TQString::fromLatin1("-//Robby Stephenson/DTD Tellico V%1.0//EN").arg(version);
}
TQString Tellico::XML::dtdTellico(int version) {
return TQString::fromLatin1("http://periapsis.org/tellico/dtd/v%1/tellico.dtd").arg(version);
}
bool Tellico::XML::validXMLElementName(const TQString& name_) {
return xmlValidateNameValue((xmlChar *)name_.utf8().data());
}
TQString Tellico::XML::elementName(const TQString& name_) {
TQString name = name_;
// change white space to dashes
name.replace(TQRegExp(TQString::fromLatin1("\\s+")), TQString::fromLatin1("-"));
// first cut, if it passes, we're done
if(XML::validXMLElementName(name)) {
return name;
}
// next check first characters IS_DIGIT is defined in libxml/vali.d
for(uint i = 0; i < name.length() && (!IS_LETTER(name[i].unicode()) || name[i] == '_'); ++i) {
name = name.mid(1);
}
if(name.isEmpty() || XML::validXMLElementName(name)) {
return name; // empty names are handled later
}
// now brute-force it, one character at a time
uint i = 0;
while(i < name.length()) {
if(!XML::validXMLElementName(name.left(i+1))) {
name.remove(i, 1); // remember it's zero-indexed
} else {
// character is ok, increment i
++i;
}
}
return name;
}