/////////////////////////////////////////////////////////////////////////////
-// Name: intl.cpp
-// Purpose: Internationalization and localisation for wxWindows
+// Name: src/common/intl.cpp
+// Purpose: Internationalization and localisation for wxWidgets
// Author: Vadim Zeitlin
-// Modified by:
+// Modified by: Michael N. Filippov <michael@idisys.iae.nsk.su>
+// (2003/09/30 - PluralForms support)
// Created: 29/01/98
// RCS-ID: $Id$
// Copyright: (c) 1998 Vadim Zeitlin <zeitlin@dptmaths.ens-cachan.fr>
-// Licence: wxWindows license
+// Licence: wxWindows licence
/////////////////////////////////////////////////////////////////////////////
// ============================================================================
// headers
// ----------------------------------------------------------------------------
-#ifdef __GNUG__
+#if defined(__GNUG__) && !defined(NO_GCC_PRAGMA)
#pragma implementation "intl.h"
#endif
+#if defined(__BORLAND__) && !defined(__WXDEBUG__)
+ // There's a bug in Borland's compiler that breaks wxLocale with -O2,
+ // so make sure that flag is not used for this file:
+ #pragma option -O1
+#endif
+
+#ifdef __EMX__
+// The following define is needed by Innotek's libc to
+// make the definition of struct localeconv available.
+#define __INTERNAL_DEFS
+#endif
+
// For compilers that support precompilation, includes "wx.h".
#include "wx/wxprec.h"
#if wxUSE_INTL
// standard headers
-#include <locale.h>
-#include <ctype.h>
-// wxWindows
-#include "wx/defs.h"
-#include "wx/string.h"
-#include "wx/tokenzr.h"
-#include "wx/intl.h"
-#include "wx/file.h"
-#include "wx/log.h"
-#include "wx/debug.h"
-#include "wx/utils.h"
-#include "wx/dynarray.h"
-#include "wx/module.h"
+#ifndef __WXWINCE__
+#include <locale.h>
+#endif
+
+#include <ctype.h>
+#include <stdlib.h>
+#ifdef HAVE_LANGINFO_H
+ #include <langinfo.h>
+#endif
+
+// wxWidgets
+#ifndef WX_PRECOMP
+ #include "wx/string.h"
+ #include "wx/intl.h"
+ #include "wx/log.h"
+ #include "wx/debug.h"
+ #include "wx/utils.h"
+ #include "wx/dynarray.h"
+#endif // WX_PRECOMP
+
#ifdef __WIN32__
-#include "wx/msw/private.h"
+ #include "wx/msw/private.h"
+#elif defined(__UNIX_LIKE__)
+ #include "wx/fontmap.h" // for CharsetToEncoding()
#endif
+#include "wx/file.h"
+#include "wx/tokenzr.h"
+#include "wx/module.h"
+#include "wx/fontmap.h"
+#include "wx/encconv.h"
+#include "wx/hashmap.h"
+#include "wx/ptr_scpd.h"
-#include <stdlib.h>
+#if defined(__WXMAC__)
+ #include "wx/mac/private.h" // includes mac headers
+#endif
// ----------------------------------------------------------------------------
// simple types
// ----------------------------------------------------------------------------
// this should *not* be wxChar, this type must have exactly 8 bits!
-typedef unsigned char size_t8;
-
-#ifdef __WXMSW__
- #if defined(__WIN16__)
- typedef unsigned long size_t32;
- #elif defined(__WIN32__)
- typedef unsigned int size_t32;
- #else
- // Win64 will have different type sizes
- #error "Please define a 32 bit type"
- #endif
-#else // !Windows
- // SIZEOF_XXX are defined by configure
- #if defined(SIZEOF_INT) && (SIZEOF_INT == 4)
- typedef unsigned int size_t32;
- #elif defined(SIZEOF_LONG) && (SIZEOF_LONG == 4)
- typedef unsigned long size_t32;
- #else
- // assume sizeof(int) == 4 - what else can we do
- typedef unsigned int size_t32;
-
- // ... but at least check it during run time
- static class IntSizeChecker
- {
- public:
- IntSizeChecker()
- {
- // Asserting a sizeof directly causes some compilers to
- // issue a "using constant in a conditional expression" warning
- size_t intsize = sizeof(int);
-
- wxASSERT_MSG( intsize == 4,
- "size_t32 is incorrectly defined!" );
- }
- } intsizechecker;
- #endif
-#endif // Win/!Win
+typedef wxUint8 size_t8;
+typedef wxUint32 size_t32;
// ----------------------------------------------------------------------------
// constants
#endif // __UNIX__
+
+// ----------------------------------------------------------------------------
+// Plural forms parser
+// ----------------------------------------------------------------------------
+
+/*
+ Simplified Grammar
+
+Expression:
+ LogicalOrExpression '?' Expression ':' Expression
+ LogicalOrExpression
+
+LogicalOrExpression:
+ LogicalAndExpression "||" LogicalOrExpression // to (a || b) || c
+ LogicalAndExpression
+
+LogicalAndExpression:
+ EqualityExpression "&&" LogicalAndExpression // to (a && b) && c
+ EqualityExpression
+
+EqualityExpression:
+ RelationalExpression "==" RelationalExperession
+ RelationalExpression "!=" RelationalExperession
+ RelationalExpression
+
+RelationalExpression:
+ MultiplicativeExpression '>' MultiplicativeExpression
+ MultiplicativeExpression '<' MultiplicativeExpression
+ MultiplicativeExpression ">=" MultiplicativeExpression
+ MultiplicativeExpression "<=" MultiplicativeExpression
+ MultiplicativeExpression
+
+MultiplicativeExpression:
+ PmExpression '%' PmExpression
+ PmExpression
+
+PmExpression:
+ N
+ Number
+ '(' Expression ')'
+*/
+
+class wxPluralFormsToken
+{
+public:
+ enum Type
+ {
+ T_ERROR, T_EOF, T_NUMBER, T_N, T_PLURAL, T_NPLURALS, T_EQUAL, T_ASSIGN,
+ T_GREATER, T_GREATER_OR_EQUAL, T_LESS, T_LESS_OR_EQUAL,
+ T_REMINDER, T_NOT_EQUAL,
+ T_LOGICAL_AND, T_LOGICAL_OR, T_QUESTION, T_COLON, T_SEMICOLON,
+ T_LEFT_BRACKET, T_RIGHT_BRACKET
+ };
+ Type type() const { return m_type; }
+ void setType(Type type) { m_type = type; }
+ // for T_NUMBER only
+ typedef int Number;
+ Number number() const { return m_number; }
+ void setNumber(Number num) { m_number = num; }
+private:
+ Type m_type;
+ Number m_number;
+};
+
+
+class wxPluralFormsScanner
+{
+public:
+ wxPluralFormsScanner(const char* s);
+ const wxPluralFormsToken& token() const { return m_token; }
+ bool nextToken(); // returns false if error
+private:
+ const char* m_s;
+ wxPluralFormsToken m_token;
+};
+
+wxPluralFormsScanner::wxPluralFormsScanner(const char* s) : m_s(s)
+{
+ nextToken();
+}
+
+bool wxPluralFormsScanner::nextToken()
+{
+ wxPluralFormsToken::Type type = wxPluralFormsToken::T_ERROR;
+ while (isspace(*m_s))
+ {
+ ++m_s;
+ }
+ if (*m_s == 0)
+ {
+ type = wxPluralFormsToken::T_EOF;
+ }
+ else if (isdigit(*m_s))
+ {
+ wxPluralFormsToken::Number number = *m_s++ - '0';
+ while (isdigit(*m_s))
+ {
+ number = number * 10 + (*m_s++ - '0');
+ }
+ m_token.setNumber(number);
+ type = wxPluralFormsToken::T_NUMBER;
+ }
+ else if (isalpha(*m_s))
+ {
+ const char* begin = m_s++;
+ while (isalnum(*m_s))
+ {
+ ++m_s;
+ }
+ size_t size = m_s - begin;
+ if (size == 1 && memcmp(begin, "n", size) == 0)
+ {
+ type = wxPluralFormsToken::T_N;
+ }
+ else if (size == 6 && memcmp(begin, "plural", size) == 0)
+ {
+ type = wxPluralFormsToken::T_PLURAL;
+ }
+ else if (size == 8 && memcmp(begin, "nplurals", size) == 0)
+ {
+ type = wxPluralFormsToken::T_NPLURALS;
+ }
+ }
+ else if (*m_s == '=')
+ {
+ ++m_s;
+ if (*m_s == '=')
+ {
+ ++m_s;
+ type = wxPluralFormsToken::T_EQUAL;
+ }
+ else
+ {
+ type = wxPluralFormsToken::T_ASSIGN;
+ }
+ }
+ else if (*m_s == '>')
+ {
+ ++m_s;
+ if (*m_s == '=')
+ {
+ ++m_s;
+ type = wxPluralFormsToken::T_GREATER_OR_EQUAL;
+ }
+ else
+ {
+ type = wxPluralFormsToken::T_GREATER;
+ }
+ }
+ else if (*m_s == '<')
+ {
+ ++m_s;
+ if (*m_s == '=')
+ {
+ ++m_s;
+ type = wxPluralFormsToken::T_LESS_OR_EQUAL;
+ }
+ else
+ {
+ type = wxPluralFormsToken::T_LESS;
+ }
+ }
+ else if (*m_s == '%')
+ {
+ ++m_s;
+ type = wxPluralFormsToken::T_REMINDER;
+ }
+ else if (*m_s == '!' && m_s[1] == '=')
+ {
+ m_s += 2;
+ type = wxPluralFormsToken::T_NOT_EQUAL;
+ }
+ else if (*m_s == '&' && m_s[1] == '&')
+ {
+ m_s += 2;
+ type = wxPluralFormsToken::T_LOGICAL_AND;
+ }
+ else if (*m_s == '|' && m_s[1] == '|')
+ {
+ m_s += 2;
+ type = wxPluralFormsToken::T_LOGICAL_OR;
+ }
+ else if (*m_s == '?')
+ {
+ ++m_s;
+ type = wxPluralFormsToken::T_QUESTION;
+ }
+ else if (*m_s == ':')
+ {
+ ++m_s;
+ type = wxPluralFormsToken::T_COLON;
+ } else if (*m_s == ';') {
+ ++m_s;
+ type = wxPluralFormsToken::T_SEMICOLON;
+ }
+ else if (*m_s == '(')
+ {
+ ++m_s;
+ type = wxPluralFormsToken::T_LEFT_BRACKET;
+ }
+ else if (*m_s == ')')
+ {
+ ++m_s;
+ type = wxPluralFormsToken::T_RIGHT_BRACKET;
+ }
+ m_token.setType(type);
+ return type != wxPluralFormsToken::T_ERROR;
+}
+
+class wxPluralFormsNode;
+
+// NB: Can't use wxDEFINE_SCOPED_PTR_TYPE because wxPluralFormsNode is not
+// fully defined yet:
+class wxPluralFormsNodePtr
+{
+public:
+ wxPluralFormsNodePtr(wxPluralFormsNode *p = NULL) : m_p(p) {}
+ ~wxPluralFormsNodePtr();
+ wxPluralFormsNode& operator*() const { return *m_p; }
+ wxPluralFormsNode* operator->() const { return m_p; }
+ wxPluralFormsNode* get() const { return m_p; }
+ wxPluralFormsNode* release();
+ void reset(wxPluralFormsNode *p);
+
+private:
+ wxPluralFormsNode *m_p;
+};
+
+class wxPluralFormsNode
+{
+public:
+ wxPluralFormsNode(const wxPluralFormsToken& token) : m_token(token) {}
+ const wxPluralFormsToken& token() const { return m_token; }
+ const wxPluralFormsNode* node(size_t i) const
+ { return m_nodes[i].get(); }
+ void setNode(size_t i, wxPluralFormsNode* n);
+ wxPluralFormsNode* releaseNode(size_t i);
+ wxPluralFormsToken::Number evaluate(wxPluralFormsToken::Number n) const;
+
+private:
+ wxPluralFormsToken m_token;
+ wxPluralFormsNodePtr m_nodes[3];
+};
+
+wxPluralFormsNodePtr::~wxPluralFormsNodePtr()
+{
+ delete m_p;
+}
+wxPluralFormsNode* wxPluralFormsNodePtr::release()
+{
+ wxPluralFormsNode *p = m_p;
+ m_p = NULL;
+ return p;
+}
+void wxPluralFormsNodePtr::reset(wxPluralFormsNode *p)
+{
+ if (p != m_p)
+ {
+ delete m_p;
+ m_p = p;
+ }
+}
+
+
+void wxPluralFormsNode::setNode(size_t i, wxPluralFormsNode* n)
+{
+ m_nodes[i].reset(n);
+}
+
+wxPluralFormsNode* wxPluralFormsNode::releaseNode(size_t i)
+{
+ return m_nodes[i].release();
+}
+
+wxPluralFormsToken::Number
+wxPluralFormsNode::evaluate(wxPluralFormsToken::Number n) const
+{
+ switch (token().type())
+ {
+ // leaf
+ case wxPluralFormsToken::T_NUMBER:
+ return token().number();
+ case wxPluralFormsToken::T_N:
+ return n;
+ // 2 args
+ case wxPluralFormsToken::T_EQUAL:
+ return node(0)->evaluate(n) == node(1)->evaluate(n);
+ case wxPluralFormsToken::T_NOT_EQUAL:
+ return node(0)->evaluate(n) != node(1)->evaluate(n);
+ case wxPluralFormsToken::T_GREATER:
+ return node(0)->evaluate(n) > node(1)->evaluate(n);
+ case wxPluralFormsToken::T_GREATER_OR_EQUAL:
+ return node(0)->evaluate(n) >= node(1)->evaluate(n);
+ case wxPluralFormsToken::T_LESS:
+ return node(0)->evaluate(n) < node(1)->evaluate(n);
+ case wxPluralFormsToken::T_LESS_OR_EQUAL:
+ return node(0)->evaluate(n) <= node(1)->evaluate(n);
+ case wxPluralFormsToken::T_REMINDER:
+ {
+ wxPluralFormsToken::Number number = node(1)->evaluate(n);
+ if (number != 0)
+ {
+ return node(0)->evaluate(n) % number;
+ }
+ else
+ {
+ return 0;
+ }
+ }
+ case wxPluralFormsToken::T_LOGICAL_AND:
+ return node(0)->evaluate(n) && node(1)->evaluate(n);
+ case wxPluralFormsToken::T_LOGICAL_OR:
+ return node(0)->evaluate(n) || node(1)->evaluate(n);
+ // 3 args
+ case wxPluralFormsToken::T_QUESTION:
+ return node(0)->evaluate(n)
+ ? node(1)->evaluate(n)
+ : node(2)->evaluate(n);
+ default:
+ return 0;
+ }
+}
+
+
+class wxPluralFormsCalculator
+{
+public:
+ wxPluralFormsCalculator() : m_nplurals(0), m_plural(0) {}
+
+ // input: number, returns msgstr index
+ int evaluate(int n) const;
+
+ // input: text after "Plural-Forms:" (e.g. "nplurals=2; plural=(n != 1);"),
+ // if s == 0, creates default handler
+ // returns 0 if error
+ static wxPluralFormsCalculator* make(const char* s = 0);
+
+ ~wxPluralFormsCalculator() {}
+
+ void init(wxPluralFormsToken::Number nplurals, wxPluralFormsNode* plural);
+ wxString getString() const;
+
+private:
+ wxPluralFormsToken::Number m_nplurals;
+ wxPluralFormsNodePtr m_plural;
+};
+
+wxDEFINE_SCOPED_PTR_TYPE(wxPluralFormsCalculator);
+
+void wxPluralFormsCalculator::init(wxPluralFormsToken::Number nplurals,
+ wxPluralFormsNode* plural)
+{
+ m_nplurals = nplurals;
+ m_plural.reset(plural);
+}
+
+int wxPluralFormsCalculator::evaluate(int n) const
+{
+ if (m_plural.get() == 0)
+ {
+ return 0;
+ }
+ wxPluralFormsToken::Number number = m_plural->evaluate(n);
+ if (number < 0 || number > m_nplurals)
+ {
+ return 0;
+ }
+ return number;
+}
+
+
+class wxPluralFormsParser
+{
+public:
+ wxPluralFormsParser(wxPluralFormsScanner& scanner) : m_scanner(scanner) {}
+ bool parse(wxPluralFormsCalculator& rCalculator);
+
+private:
+ wxPluralFormsNode* parsePlural();
+ // stops at T_SEMICOLON, returns 0 if error
+ wxPluralFormsScanner& m_scanner;
+ const wxPluralFormsToken& token() const;
+ bool nextToken();
+
+ wxPluralFormsNode* expression();
+ wxPluralFormsNode* logicalOrExpression();
+ wxPluralFormsNode* logicalAndExpression();
+ wxPluralFormsNode* equalityExpression();
+ wxPluralFormsNode* multiplicativeExpression();
+ wxPluralFormsNode* relationalExpression();
+ wxPluralFormsNode* pmExpression();
+};
+
+bool wxPluralFormsParser::parse(wxPluralFormsCalculator& rCalculator)
+{
+ if (token().type() != wxPluralFormsToken::T_NPLURALS)
+ return false;
+ if (!nextToken())
+ return false;
+ if (token().type() != wxPluralFormsToken::T_ASSIGN)
+ return false;
+ if (!nextToken())
+ return false;
+ if (token().type() != wxPluralFormsToken::T_NUMBER)
+ return false;
+ wxPluralFormsToken::Number nplurals = token().number();
+ if (!nextToken())
+ return false;
+ if (token().type() != wxPluralFormsToken::T_SEMICOLON)
+ return false;
+ if (!nextToken())
+ return false;
+ if (token().type() != wxPluralFormsToken::T_PLURAL)
+ return false;
+ if (!nextToken())
+ return false;
+ if (token().type() != wxPluralFormsToken::T_ASSIGN)
+ return false;
+ if (!nextToken())
+ return false;
+ wxPluralFormsNode* plural = parsePlural();
+ if (plural == 0)
+ return false;
+ if (token().type() != wxPluralFormsToken::T_SEMICOLON)
+ return false;
+ if (!nextToken())
+ return false;
+ if (token().type() != wxPluralFormsToken::T_EOF)
+ return false;
+ rCalculator.init(nplurals, plural);
+ return true;
+}
+
+wxPluralFormsNode* wxPluralFormsParser::parsePlural()
+{
+ wxPluralFormsNode* p = expression();
+ if (p == NULL)
+ {
+ return NULL;
+ }
+ wxPluralFormsNodePtr n(p);
+ if (token().type() != wxPluralFormsToken::T_SEMICOLON)
+ {
+ return NULL;
+ }
+ return n.release();
+}
+
+const wxPluralFormsToken& wxPluralFormsParser::token() const
+{
+ return m_scanner.token();
+}
+
+bool wxPluralFormsParser::nextToken()
+{
+ if (!m_scanner.nextToken())
+ return false;
+ return true;
+}
+
+wxPluralFormsNode* wxPluralFormsParser::expression()
+{
+ wxPluralFormsNode* p = logicalOrExpression();
+ if (p == NULL)
+ return NULL;
+ wxPluralFormsNodePtr n(p);
+ if (token().type() == wxPluralFormsToken::T_QUESTION)
+ {
+ wxPluralFormsNodePtr qn(new wxPluralFormsNode(token()));
+ if (!nextToken())
+ {
+ return 0;
+ }
+ p = expression();
+ if (p == 0)
+ {
+ return 0;
+ }
+ qn->setNode(1, p);
+ if (token().type() != wxPluralFormsToken::T_COLON)
+ {
+ return 0;
+ }
+ if (!nextToken())
+ {
+ return 0;
+ }
+ p = expression();
+ if (p == 0)
+ {
+ return 0;
+ }
+ qn->setNode(2, p);
+ qn->setNode(0, n.release());
+ return qn.release();
+ }
+ return n.release();
+}
+
+wxPluralFormsNode*wxPluralFormsParser::logicalOrExpression()
+{
+ wxPluralFormsNode* p = logicalAndExpression();
+ if (p == NULL)
+ return NULL;
+ wxPluralFormsNodePtr ln(p);
+ if (token().type() == wxPluralFormsToken::T_LOGICAL_OR)
+ {
+ wxPluralFormsNodePtr un(new wxPluralFormsNode(token()));
+ if (!nextToken())
+ {
+ return 0;
+ }
+ p = logicalOrExpression();
+ if (p == 0)
+ {
+ return 0;
+ }
+ wxPluralFormsNodePtr rn(p); // right
+ if (rn->token().type() == wxPluralFormsToken::T_LOGICAL_OR)
+ {
+ // see logicalAndExpression comment
+ un->setNode(0, ln.release());
+ un->setNode(1, rn->releaseNode(0));
+ rn->setNode(0, un.release());
+ return rn.release();
+ }
+
+
+ un->setNode(0, ln.release());
+ un->setNode(1, rn.release());
+ return un.release();
+ }
+ return ln.release();
+}
+
+wxPluralFormsNode* wxPluralFormsParser::logicalAndExpression()
+{
+ wxPluralFormsNode* p = equalityExpression();
+ if (p == NULL)
+ return NULL;
+ wxPluralFormsNodePtr ln(p); // left
+ if (token().type() == wxPluralFormsToken::T_LOGICAL_AND)
+ {
+ wxPluralFormsNodePtr un(new wxPluralFormsNode(token())); // up
+ if (!nextToken())
+ {
+ return NULL;
+ }
+ p = logicalAndExpression();
+ if (p == 0)
+ {
+ return NULL;
+ }
+ wxPluralFormsNodePtr rn(p); // right
+ if (rn->token().type() == wxPluralFormsToken::T_LOGICAL_AND)
+ {
+// transform 1 && (2 && 3) -> (1 && 2) && 3
+// u r
+// l r -> u 3
+// 2 3 l 2
+ un->setNode(0, ln.release());
+ un->setNode(1, rn->releaseNode(0));
+ rn->setNode(0, un.release());
+ return rn.release();
+ }
+
+ un->setNode(0, ln.release());
+ un->setNode(1, rn.release());
+ return un.release();
+ }
+ return ln.release();
+}
+
+wxPluralFormsNode* wxPluralFormsParser::equalityExpression()
+{
+ wxPluralFormsNode* p = relationalExpression();
+ if (p == NULL)
+ return NULL;
+ wxPluralFormsNodePtr n(p);
+ if (token().type() == wxPluralFormsToken::T_EQUAL
+ || token().type() == wxPluralFormsToken::T_NOT_EQUAL)
+ {
+ wxPluralFormsNodePtr qn(new wxPluralFormsNode(token()));
+ if (!nextToken())
+ {
+ return NULL;
+ }
+ p = relationalExpression();
+ if (p == NULL)
+ {
+ return NULL;
+ }
+ qn->setNode(1, p);
+ qn->setNode(0, n.release());
+ return qn.release();
+ }
+ return n.release();
+}
+
+wxPluralFormsNode* wxPluralFormsParser::relationalExpression()
+{
+ wxPluralFormsNode* p = multiplicativeExpression();
+ if (p == NULL)
+ return NULL;
+ wxPluralFormsNodePtr n(p);
+ if (token().type() == wxPluralFormsToken::T_GREATER
+ || token().type() == wxPluralFormsToken::T_LESS
+ || token().type() == wxPluralFormsToken::T_GREATER_OR_EQUAL
+ || token().type() == wxPluralFormsToken::T_LESS_OR_EQUAL)
+ {
+ wxPluralFormsNodePtr qn(new wxPluralFormsNode(token()));
+ if (!nextToken())
+ {
+ return NULL;
+ }
+ p = multiplicativeExpression();
+ if (p == NULL)
+ {
+ return NULL;
+ }
+ qn->setNode(1, p);
+ qn->setNode(0, n.release());
+ return qn.release();
+ }
+ return n.release();
+}
+
+wxPluralFormsNode* wxPluralFormsParser::multiplicativeExpression()
+{
+ wxPluralFormsNode* p = pmExpression();
+ if (p == NULL)
+ return NULL;
+ wxPluralFormsNodePtr n(p);
+ if (token().type() == wxPluralFormsToken::T_REMINDER)
+ {
+ wxPluralFormsNodePtr qn(new wxPluralFormsNode(token()));
+ if (!nextToken())
+ {
+ return NULL;
+ }
+ p = pmExpression();
+ if (p == NULL)
+ {
+ return NULL;
+ }
+ qn->setNode(1, p);
+ qn->setNode(0, n.release());
+ return qn.release();
+ }
+ return n.release();
+}
+
+wxPluralFormsNode* wxPluralFormsParser::pmExpression()
+{
+ wxPluralFormsNodePtr n;
+ if (token().type() == wxPluralFormsToken::T_N
+ || token().type() == wxPluralFormsToken::T_NUMBER)
+ {
+ n.reset(new wxPluralFormsNode(token()));
+ if (!nextToken())
+ {
+ return NULL;
+ }
+ }
+ else if (token().type() == wxPluralFormsToken::T_LEFT_BRACKET) {
+ if (!nextToken())
+ {
+ return NULL;
+ }
+ wxPluralFormsNode* p = expression();
+ if (p == NULL)
+ {
+ return NULL;
+ }
+ n.reset(p);
+ if (token().type() != wxPluralFormsToken::T_RIGHT_BRACKET)
+ {
+ return NULL;
+ }
+ if (!nextToken())
+ {
+ return NULL;
+ }
+ }
+ else
+ {
+ return NULL;
+ }
+ return n.release();
+}
+
+wxPluralFormsCalculator* wxPluralFormsCalculator::make(const char* s)
+{
+ wxPluralFormsCalculatorPtr calculator(new wxPluralFormsCalculator);
+ if (s != NULL)
+ {
+ wxPluralFormsScanner scanner(s);
+ wxPluralFormsParser p(scanner);
+ if (!p.parse(*calculator))
+ {
+ return NULL;
+ }
+ }
+ return calculator.release();
+}
+
+
+
+
// ----------------------------------------------------------------------------
-// wxMsgCatalog corresponds to one disk-file message catalog.
+// wxMsgCatalogFile corresponds to one disk-file message catalog.
+//
+// This is a "low-level" class and is used only by wxMsgCatalog
+// ----------------------------------------------------------------------------
+
+WX_DECLARE_EXPORTED_STRING_HASH_MAP(wxString, wxMessagesHash);
+
+class wxMsgCatalogFile
+{
+public:
+ // ctor & dtor
+ wxMsgCatalogFile();
+ ~wxMsgCatalogFile();
+
+ // load the catalog from disk (szDirPrefix corresponds to language)
+ bool Load(const wxChar *szDirPrefix, const wxChar *szName,
+ wxPluralFormsCalculatorPtr& rPluralFormsCalculator);
+
+ // fills the hash with string-translation pairs
+ void FillHash(wxMessagesHash& hash, const wxString& msgIdCharset,
+ bool convertEncoding) const;
+
+private:
+ // this implementation is binary compatible with GNU gettext() version 0.10
+
+ // an entry in the string table
+ struct wxMsgTableEntry
+ {
+ size_t32 nLen; // length of the string
+ size_t32 ofsString; // pointer to the string
+ };
+
+ // header of a .mo file
+ struct wxMsgCatalogHeader
+ {
+ size_t32 magic, // offset +00: magic id
+ revision, // +04: revision
+ numStrings; // +08: number of strings in the file
+ size_t32 ofsOrigTable, // +0C: start of original string table
+ ofsTransTable; // +10: start of translated string table
+ size_t32 nHashSize, // +14: hash table size
+ ofsHashTable; // +18: offset of hash table start
+ };
+
+ // all data is stored here, NULL if no data loaded
+ size_t8 *m_pData;
+
+ // amount of memory pointed to by m_pData.
+ size_t32 m_nSize;
+
+ // data description
+ size_t32 m_numStrings; // number of strings in this domain
+ wxMsgTableEntry *m_pOrigTable, // pointer to original strings
+ *m_pTransTable; // translated
+
+ wxString m_charset;
+
+ // swap the 2 halves of 32 bit integer if needed
+ size_t32 Swap(size_t32 ui) const
+ {
+ return m_bSwapped ? (ui << 24) | ((ui & 0xff00) << 8) |
+ ((ui >> 8) & 0xff00) | (ui >> 24)
+ : ui;
+ }
+
+ const char *StringAtOfs(wxMsgTableEntry *pTable, size_t32 n) const
+ {
+ const wxMsgTableEntry * const ent = pTable + n;
+
+ // this check could fail for a corrupt message catalog
+ size_t32 ofsString = Swap(ent->ofsString);
+ if ( ofsString + Swap(ent->nLen) > m_nSize)
+ {
+ return NULL;
+ }
+
+ return (const char *)(m_pData + ofsString);
+ }
+
+ bool m_bSwapped; // wrong endianness?
+
+ DECLARE_NO_COPY_CLASS(wxMsgCatalogFile)
+};
+
+
+// ----------------------------------------------------------------------------
+// wxMsgCatalog corresponds to one loaded message catalog.
//
// This is a "low-level" class and is used only by wxLocale (that's why
// it's designed to be stored in a linked list)
class wxMsgCatalog
{
public:
- // ctor & dtor
- wxMsgCatalog();
- ~wxMsgCatalog();
-
- // load the catalog from disk (szDirPrefix corresponds to language)
- bool Load(const wxChar *szDirPrefix, const wxChar *szName, bool bConvertEncoding = FALSE);
- bool IsLoaded() const { return m_pData != NULL; }
+ // load the catalog from disk (szDirPrefix corresponds to language)
+ bool Load(const wxChar *szDirPrefix, const wxChar *szName,
+ const wxChar *msgIdCharset = NULL, bool bConvertEncoding = false);
- // get name of the catalog
- const wxChar *GetName() const { return m_pszName; }
+ // get name of the catalog
+ wxString GetName() const { return m_name; }
- // get the translated string: returns NULL if not found
- const char *GetString(const char *sz) const;
+ // get the translated string: returns NULL if not found
+ const wxChar *GetString(const wxChar *sz, size_t n = size_t(-1)) const;
- // public variable pointing to the next element in a linked list (or NULL)
- wxMsgCatalog *m_pNext;
+ // public variable pointing to the next element in a linked list (or NULL)
+ wxMsgCatalog *m_pNext;
private:
- // this implementation is binary compatible with GNU gettext() version 0.10
-
- // an entry in the string table
- struct wxMsgTableEntry
- {
- size_t32 nLen; // length of the string
- size_t32 ofsString; // pointer to the string
- };
-
- // header of a .mo file
- struct wxMsgCatalogHeader
- {
- size_t32 magic, // offset +00: magic id
- revision, // +04: revision
- numStrings; // +08: number of strings in the file
- size_t32 ofsOrigTable, // +0C: start of original string table
- ofsTransTable; // +10: start of translated string table
- size_t32 nHashSize, // +14: hash table size
- ofsHashTable; // +18: offset of hash table start
- };
-
- // all data is stored here, NULL if no data loaded
- size_t8 *m_pData;
-
- // data description
- size_t32 m_numStrings, // number of strings in this domain
- m_nHashSize; // number of entries in hash table
- size_t32 *m_pHashTable; // pointer to hash table
- wxMsgTableEntry *m_pOrigTable, // pointer to original strings
- *m_pTransTable; // translated
-
- const char *StringAtOfs(wxMsgTableEntry *pTable, size_t32 index) const
- { return (const char *)(m_pData + Swap(pTable[index].ofsString)); }
-
- // convert encoding to platform native one, if neccessary
- void ConvertEncoding();
-
- // utility functions
- // calculate the hash value of given string
- static size_t32 GetHash(const char *sz);
- // big<->little endian
- inline size_t32 Swap(size_t32 ui) const;
-
- // internal state
- bool HasHashTable() const // true if hash table is present
- { return m_nHashSize > 2 && m_pHashTable != NULL; }
-
- bool m_bSwapped; // wrong endianness?
-
- wxChar *m_pszName; // name of the domain
+ wxMessagesHash m_messages; // all messages in the catalog
+ wxString m_name; // name of the domain
+ wxPluralFormsCalculatorPtr m_pluralFormsCalculator;
};
// ----------------------------------------------------------------------------
// ============================================================================
// ----------------------------------------------------------------------------
-// wxMsgCatalog class
+// wxMsgCatalogFile class
// ----------------------------------------------------------------------------
-// calculate hash value using the so called hashpjw function by P.J. Weinberger
-// [see Aho/Sethi/Ullman, COMPILERS: Principles, Techniques and Tools]
-size_t32 wxMsgCatalog::GetHash(const char *sz)
+wxMsgCatalogFile::wxMsgCatalogFile()
{
- #define HASHWORDBITS 32 // the length of size_t32
-
- size_t32 hval = 0;
- size_t32 g;
- while ( *sz != '\0' ) {
- hval <<= 4;
- hval += (size_t32)*sz++;
- g = hval & ((size_t32)0xf << (HASHWORDBITS - 4));
- if ( g != 0 ) {
- hval ^= g >> (HASHWORDBITS - 8);
- hval ^= g;
- }
- }
-
- return hval;
+ m_pData = NULL;
+ m_nSize = 0;
}
-// swap the 2 halves of 32 bit integer if needed
-size_t32 wxMsgCatalog::Swap(size_t32 ui) const
+wxMsgCatalogFile::~wxMsgCatalogFile()
{
- return m_bSwapped ? (ui << 24) | ((ui & 0xff00) << 8) |
- ((ui >> 8) & 0xff00) | (ui >> 24)
- : ui;
-}
-
-wxMsgCatalog::wxMsgCatalog()
-{
- m_pData = NULL;
- m_pszName = NULL;
-}
-
-wxMsgCatalog::~wxMsgCatalog()
-{
- wxDELETEA(m_pData);
- wxDELETEA(m_pszName);
+ wxDELETEA(m_pData);
}
// return all directories to search for given prefix
// LC_PATH is a standard env var containing the search path for the .mo
// files
+#ifndef __WXWINCE__
const wxChar *pszLcPath = wxGetenv(wxT("LC_PATH"));
if ( pszLcPath != NULL )
searchPath << GetAllMsgCatalogSubdirs(pszLcPath, lang);
+#endif
- // then take the current directory
- // FIXME it should be the directory of the executable
- searchPath << GetAllMsgCatalogSubdirs(wxT("."), lang);
-
- // and finally add some standard ones
+#ifdef __UNIX__
+ // add some standard ones and the one in the tree where wxWin was installed:
searchPath
+ << GetAllMsgCatalogSubdirs(wxString(wxGetInstallPrefix()) + wxT("/share/locale"), lang)
<< GetAllMsgCatalogSubdirs(wxT("/usr/share/locale"), lang)
<< GetAllMsgCatalogSubdirs(wxT("/usr/lib/locale"), lang)
<< GetAllMsgCatalogSubdirs(wxT("/usr/local/share/locale"), lang);
+#endif // __UNIX__
+
+ // then take the current directory
+ // FIXME it should be the directory of the executable
+#ifdef __WXMAC__
+ wxChar cwd[512] ;
+ wxGetWorkingDirectory( cwd , sizeof( cwd ) ) ;
+ searchPath << GetAllMsgCatalogSubdirs(cwd, lang);
+ // generic search paths could be somewhere in the system folder preferences
+#else // !Mac
+ searchPath << GetAllMsgCatalogSubdirs(wxT("."), lang);
+
+#endif // platform
return searchPath;
}
// open disk file and read in it's contents
-bool wxMsgCatalog::Load(const wxChar *szDirPrefix, const wxChar *szName0, bool bConvertEncoding)
+bool wxMsgCatalogFile::Load(const wxChar *szDirPrefix, const wxChar *szName0,
+ wxPluralFormsCalculatorPtr& rPluralFormsCalculator)
{
/* We need to handle locales like de_AT.iso-8859-1
For this we first chop off the .CHARSET specifier and ignore it.
FIXME: UNICODE SUPPORT: must use CHARSET specifier!
*/
wxString szName = szName0;
- if(szName.Find(wxT('.')) != -1) // contains a dot
+ if(szName.Find(wxT('.')) != wxNOT_FOUND) // contains a dot
szName = szName.Left(szName.Find(wxT('.')));
wxString searchPath = GetFullSearchPath(szDirPrefix);
wxString strFullName;
if ( !wxFindFileInPath(&strFullName, searchPath, strFile) ) {
wxLogVerbose(_("catalog file for domain '%s' not found."), szName.c_str());
- return FALSE;
+ return false;
}
// open file
wxFile fileMsg(strFullName);
if ( !fileMsg.IsOpened() )
- return FALSE;
+ return false;
// get the file size
off_t nSize = fileMsg.Length();
if ( nSize == wxInvalidOffset )
- return FALSE;
+ return false;
// read the whole file in memory
m_pData = new size_t8[nSize];
if ( fileMsg.Read(m_pData, nSize) != nSize ) {
wxDELETEA(m_pData);
- return FALSE;
+ return false;
}
// examine header
wxLogWarning(_("'%s' is not a valid message catalog."), strFullName.c_str());
wxDELETEA(m_pData);
- return FALSE;
+ return false;
}
// initialize
Swap(pHeader->ofsOrigTable));
m_pTransTable = (wxMsgTableEntry *)(m_pData +
Swap(pHeader->ofsTransTable));
+ m_nSize = nSize;
- m_nHashSize = Swap(pHeader->nHashSize);
- m_pHashTable = (size_t32 *)(m_pData + Swap(pHeader->ofsHashTable));
+ // now parse catalog's header and try to extract catalog charset and
+ // plural forms formula from it:
- m_pszName = new wxChar[wxStrlen(szName) + 1];
- wxStrcpy(m_pszName, szName);
+ const char* headerData = StringAtOfs(m_pOrigTable, 0);
+ if (headerData && headerData[0] == 0)
+ {
+ // Extract the charset:
+ wxString header = wxString::FromAscii(StringAtOfs(m_pTransTable, 0));
+ int begin = header.Find(wxT("Content-Type: text/plain; charset="));
+ if (begin != wxNOT_FOUND)
+ {
+ begin += 34; //strlen("Content-Type: text/plain; charset=")
+ size_t end = header.find('\n', begin);
+ if (end != size_t(-1))
+ {
+ m_charset.assign(header, begin, end - begin);
+ if (m_charset == wxT("CHARSET"))
+ {
+ // "CHARSET" is not valid charset, but lazy translator
+ m_charset.Clear();
+ }
+ }
+ }
+ // else: incorrectly filled Content-Type header
- if (bConvertEncoding)
- ConvertEncoding();
+ // Extract plural forms:
+ begin = header.Find(wxT("Plural-Forms:"));
+ if (begin != wxNOT_FOUND)
+ {
+ begin += 13;
+ size_t end = header.find('\n', begin);
+ if (end != size_t(-1))
+ {
+ wxString pfs(header, begin, end - begin);
+ wxPluralFormsCalculator* pCalculator = wxPluralFormsCalculator
+ ::make(pfs.ToAscii());
+ if (pCalculator != 0)
+ {
+ rPluralFormsCalculator.reset(pCalculator);
+ }
+ else
+ {
+ wxLogVerbose(_("Cannot parse Plural-Forms:'%s'"),
+ pfs.c_str());
+ }
+ }
+ }
+ if (rPluralFormsCalculator.get() == NULL)
+ {
+ rPluralFormsCalculator.reset(wxPluralFormsCalculator::make());
+ }
+ }
// everything is fine
- return TRUE;
+ return true;
}
-// search for a string
-const char *wxMsgCatalog::GetString(const char *szOrig) const
+void wxMsgCatalogFile::FillHash(wxMessagesHash& hash,
+ const wxString& msgIdCharset,
+ bool convertEncoding) const
{
- if ( szOrig == NULL )
- return NULL;
+#if wxUSE_WCHAR_T
+ wxCSConv *csConv = NULL;
+ if ( !m_charset.IsEmpty() )
+ csConv = new wxCSConv(m_charset);
- if ( HasHashTable() ) { // use hash table for lookup if possible
- size_t32 nHashVal = GetHash(szOrig);
- size_t32 nIndex = nHashVal % m_nHashSize;
+ wxMBConv& inputConv = csConv ? *((wxMBConv*)csConv) : *wxConvCurrent;
- size_t32 nIncr = 1 + (nHashVal % (m_nHashSize - 2));
+ wxCSConv *sourceConv = NULL;
+ if ( !msgIdCharset.empty() && (m_charset != msgIdCharset) )
+ sourceConv = new wxCSConv(msgIdCharset);
- for ( ;; ) {
- size_t32 nStr = Swap(m_pHashTable[nIndex]);
- if ( nStr == 0 )
- return NULL;
+#elif wxUSE_FONTMAP
+ wxASSERT_MSG( msgIdCharset == NULL,
+ _T("non-ASCII msgid languages only supported if wxUSE_WCHAR_T=1") );
- if ( strcmp(szOrig, StringAtOfs(m_pOrigTable, nStr - 1)) == 0 ) {
- // work around for BC++ 5.5 bug: without a temp var, the optimizer
- // breaks the code and the return value is incorrect
- const char *tmp = StringAtOfs(m_pTransTable, nStr - 1);
- return tmp;
- }
+ wxEncodingConverter converter;
+ if ( convertEncoding )
+ {
+ wxFontEncoding targetEnc = wxFONTENCODING_SYSTEM;
+ wxFontEncoding enc = wxFontMapper::Get()->CharsetToEncoding(m_charset, false);
+ if ( enc == wxFONTENCODING_SYSTEM )
+ {
+ convertEncoding = false; // unknown encoding
+ }
+ else
+ {
+ targetEnc = wxLocale::GetSystemEncoding();
+ if (targetEnc == wxFONTENCODING_SYSTEM)
+ {
+ wxFontEncodingArray a = wxEncodingConverter::GetPlatformEquivalents(enc);
+ if (a[0] == enc)
+ // no conversion needed, locale uses native encoding
+ convertEncoding = false;
+ if (a.GetCount() == 0)
+ // we don't know common equiv. under this platform
+ convertEncoding = false;
+ targetEnc = a[0];
+ }
+ }
- if ( nIndex >= m_nHashSize - nIncr)
- nIndex -= m_nHashSize - nIncr;
- else
- nIndex += nIncr;
+ if ( convertEncoding )
+ {
+ converter.Init(enc, targetEnc);
+ }
}
- }
- else { // no hash table: use default binary search
- size_t32 bottom = 0,
- top = m_numStrings,
- current;
- while ( bottom < top ) {
- current = (bottom + top) / 2;
- int res = strcmp(szOrig, StringAtOfs(m_pOrigTable, current));
- if ( res < 0 )
- top = current;
- else if ( res > 0 )
- bottom = current + 1;
- else { // found!
- // work around the same BC++ 5.5 bug as above
- const char *tmp = StringAtOfs(m_pTransTable, current);
- return tmp;
- }
+#endif // wxUSE_WCHAR_T/!wxUSE_WCHAR_T
+ (void)convertEncoding; // get rid of warnings about unused parameter
+
+ for (size_t i = 0; i < m_numStrings; i++)
+ {
+ const char *data = StringAtOfs(m_pOrigTable, i);
+#if wxUSE_UNICODE
+ wxString msgid(data, inputConv);
+#else
+ wxString msgid;
+#if wxUSE_WCHAR_T
+ if ( convertEncoding && sourceConv )
+ msgid = wxString(inputConv.cMB2WC(data), *sourceConv);
+ else
+#endif
+ msgid = data;
+#endif // wxUSE_UNICODE
+
+ data = StringAtOfs(m_pTransTable, i);
+ size_t length = Swap(m_pTransTable[i].nLen);
+ size_t offset = 0;
+ size_t index = 0;
+ while (offset < length)
+ {
+ wxString msgstr;
+#if wxUSE_WCHAR_T
+ #if wxUSE_UNICODE
+ msgstr = wxString(data + offset, inputConv);
+ #else
+ if ( convertEncoding )
+ msgstr = wxString(inputConv.cMB2WC(data + offset), wxConvLocal);
+ else
+ msgstr = wxString(data + offset);
+ #endif
+#else // !wxUSE_WCHAR_T
+ #if wxUSE_FONTMAP
+ if ( convertEncoding )
+ msgstr = wxString(converter.Convert(data + offset));
+ else
+ #endif
+ msgstr = wxString(data + offset);
+#endif // wxUSE_WCHAR_T/!wxUSE_WCHAR_T
+
+ if ( !msgstr.empty() )
+ {
+ hash[index == 0 ? msgid : msgid + wxChar(index)] = msgstr;
+ }
+ offset += strlen(data + offset) + 1;
+ ++index;
+ }
}
- }
- // not found
- return NULL;
+#if wxUSE_WCHAR_T
+ delete sourceConv;
+ delete csConv;
+#endif
}
-#if wxUSE_GUI
-#include "wx/fontmap.h"
-#include "wx/encconv.h"
-#endif
+// ----------------------------------------------------------------------------
+// wxMsgCatalog class
+// ----------------------------------------------------------------------------
-void wxMsgCatalog::ConvertEncoding()
+bool wxMsgCatalog::Load(const wxChar *szDirPrefix, const wxChar *szName,
+ const wxChar *msgIdCharset, bool bConvertEncoding)
{
-#if wxUSE_GUI
- wxFontEncoding enc;
+ wxMsgCatalogFile file;
- // first, find encoding header:
- const char *hdr = StringAtOfs(m_pOrigTable, 0);
- if ( hdr == NULL || hdr[0] != 0 ) {
- // not supported by this catalog, does not have non-fuzzy header
- return;
- }
+ m_name = szName;
- /*
- we support catalogs with header (msgid "") that is _not_ marked as "#,
- fuzzy" (otherwise the string would not be included into compiled
- catalog)
- */
- wxString header(StringAtOfs(m_pTransTable, 0));
- wxString charset;
- int pos = header.Find(wxT("Content-Type: text/plain; charset="));
- if (pos == wxNOT_FOUND)
- return; // incorrectly filled Content-Type header
- size_t n = pos + 34; /*strlen("Content-Type: text/plain; charset=")*/
- while (header[n] != wxT('\n'))
- charset << header[n++];
+ if ( file.Load(szDirPrefix, szName, m_pluralFormsCalculator) )
+ {
+ file.FillHash(m_messages, msgIdCharset, bConvertEncoding);
+ return true;
+ }
- enc = wxTheFontMapper->CharsetToEncoding(charset, FALSE);
- if ( enc == wxFONTENCODING_SYSTEM )
- return; // unknown encoding
+ return false;
+}
- wxFontEncoding targetEnc = wxFONTENCODING_SYSTEM;
-#ifdef __UNIX__
- wxString langFull;
- if (wxGetEnv(wxT("LC_ALL"), &langFull) ||
- wxGetEnv(wxT("LC_CTYPE"), &langFull) ||
- wxGetEnv(wxT("LANG"), &langFull))
+const wxChar *wxMsgCatalog::GetString(const wxChar *sz, size_t n) const
+{
+ int index = 0;
+ if (n != size_t(-1))
{
- wxString lcharset = langFull.AfterFirst(wxT('.')).BeforeFirst(wxT('@'));
- if (!lcharset.IsEmpty())
- targetEnc = wxTheFontMapper->CharsetToEncoding(lcharset, FALSE);
+ index = m_pluralFormsCalculator->evaluate(n);
}
-#endif
-
- if (targetEnc == wxFONTENCODING_SYSTEM)
+ wxMessagesHash::const_iterator i;
+ if (index != 0)
{
- wxFontEncodingArray a = wxEncodingConverter::GetPlatformEquivalents(enc);
- if (a[0] == enc)
- return; // no conversion needed, locale uses native encoding
- if (a.GetCount() == 0)
- return; // we don't know common equiv. under this platform
- targetEnc = a[0];
+ i = m_messages.find(wxString(sz) + wxChar(index)); // plural
+ }
+ else
+ {
+ i = m_messages.find(sz);
}
- wxEncodingConverter converter;
- converter.Init(enc, targetEnc);
-
- for (size_t i = 0; i < m_numStrings; i++)
- converter.Convert((char*)StringAtOfs(m_pTransTable, i));
-#endif // wxUSE_GUI
+ if ( i != m_messages.end() )
+ {
+ return i->second.c_str();
+ }
+ else
+ return NULL;
}
-
// ----------------------------------------------------------------------------
// wxLocale
// ----------------------------------------------------------------------------
}
-wxLocale::wxLocale()
+void wxLocale::DoCommonInit()
{
m_pszOldLocale = NULL;
m_pMsgCat = NULL;
m_language = wxLANGUAGE_UNKNOWN;
+ m_initialized = false;
}
// NB: this function has (desired) side effect of changing current locale
bool bLoadDefault,
bool bConvertEncoding)
{
+ wxASSERT_MSG( !m_initialized,
+ _T("you can't call wxLocale::Init more than once") );
+
+ m_initialized = true;
m_strLocale = szName;
m_strShort = szShort;
m_bConvertEncoding = bConvertEncoding;
{
// the argument to setlocale()
szLocale = szShort;
+
+ wxCHECK_MSG( szLocale, false, _T("no locale to set in wxLocale::Init()") );
+ }
+
+#ifdef __WXWINCE__
+ // FIXME: I'm guessing here
+ wxChar localeName[256];
+ int ret = GetLocaleInfo(LOCALE_USER_DEFAULT, LOCALE_SLANGUAGE, localeName,
+ 256);
+ if (ret != 0)
+ {
+ m_pszOldLocale = wxStrdup(localeName);
}
- m_pszOldLocale = wxSetlocale(LC_ALL, szLocale);
+ else
+ m_pszOldLocale = NULL;
+
+ // TODO: how to find languageId
+ // SetLocaleInfo(languageId, SORT_DEFAULT, localeName);
+#else
+ wxMB2WXbuf oldLocale = wxSetlocale(LC_ALL, szLocale);
+ if ( oldLocale )
+ m_pszOldLocale = wxStrdup(oldLocale);
+ else
+ m_pszOldLocale = NULL;
+#endif
+
if ( m_pszOldLocale == NULL )
wxLogError(_("locale '%s' can not be set."), szLocale);
if ( m_strShort.IsEmpty() ) {
// FIXME I don't know how these 2 letter abbreviations are formed,
// this wild guess is surely wrong
- m_strShort = tolower(szLocale[0]) + tolower(szLocale[1]);
+ if ( szLocale && szLocale[0] )
+ {
+ m_strShort += (wxChar)wxTolower(szLocale[0]);
+ if ( szLocale[1] )
+ m_strShort += (wxChar)wxTolower(szLocale[1]);
+ }
}
// save the old locale to be able to restore it later
m_pOldLocale = wxSetLocale(this);
- // load the default catalog with wxWindows standard messages
+ // load the default catalog with wxWidgets standard messages
m_pMsgCat = NULL;
- bool bOk = TRUE;
+ bool bOk = true;
if ( bLoadDefault )
bOk = AddCatalog(wxT("wxstd"));
return bOk;
}
+
+#if defined(__UNIX__) && wxUSE_UNICODE
+static wxWCharBuffer wxSetlocaleTryUTF(int c, const wxChar *lc)
+{
+ wxMB2WXbuf l = wxSetlocale(c, lc);
+ if ( !l && lc && lc[0] != 0 )
+ {
+ wxString buf(lc);
+ wxString buf2;
+ buf2 = buf + wxT(".UTF-8");
+ l = wxSetlocale(c, buf2.c_str());
+ if ( !l )
+ {
+ buf2 = buf + wxT(".utf-8");
+ l = wxSetlocale(c, buf2.c_str());
+ }
+ if ( !l )
+ {
+ buf2 = buf + wxT(".UTF8");
+ l = wxSetlocale(c, buf2.c_str());
+ }
+ if ( !l )
+ {
+ buf2 = buf + wxT(".utf8");
+ l = wxSetlocale(c, buf2.c_str());
+ }
+ }
+ return l;
+}
+#else
+#define wxSetlocaleTryUTF(c, lc) wxSetlocale(c, lc)
+#endif
+
bool wxLocale::Init(int language, int flags)
{
- wxLanguageInfo *info = NULL;
int lang = language;
-
- CreateLanguagesDB();
-
if (lang == wxLANGUAGE_DEFAULT)
{
// auto detect the language
// We failed to detect system language, so we will use English:
if (lang == wxLANGUAGE_UNKNOWN)
{
- return FALSE;
+ return false;
}
- if (lang != wxLANGUAGE_DEFAULT)
- {
- for (size_t i = 0; i < ms_languagesDB->GetCount(); i++)
- {
- if (ms_languagesDB->Item(i).Language == lang)
- {
- info = &ms_languagesDB->Item(i);
- break;
- }
- }
- }
+ const wxLanguageInfo *info = GetLanguageInfo(lang);
// Unknown language:
if (info == NULL)
{
wxLogError(wxT("Unknown language %i."), lang);
- return FALSE;
+ return false;
}
wxString name = info->Description;
wxString canonical = info->CanonicalName;
wxString locale;
- wxChar *retloc;
// Set the locale:
-#ifdef __UNIX__
+#if defined(__UNIX__) && !defined(__WXMAC__)
if (language == wxLANGUAGE_DEFAULT)
locale = wxEmptyString;
else
locale = info->CanonicalName;
- retloc = wxSetlocale(LC_ALL, locale);
+ wxMB2WXbuf retloc = wxSetlocaleTryUTF(LC_ALL, locale);
- if (retloc == NULL)
+ if ( !retloc )
{
// Some C libraries don't like xx_YY form and require xx only
- retloc = wxSetlocale(LC_ALL, locale.Mid(0,2));
+ retloc = wxSetlocaleTryUTF(LC_ALL, locale.Mid(0,2));
}
- if (retloc == NULL)
+ if ( !retloc )
{
// Some C libraries (namely glibc) still use old ISO 639,
// so will translate the abbrev for them
wxString mid = locale.Mid(0,2);
- if (mid == wxT("he")) locale = wxT("iw") + locale.Mid(3);
- else if (mid == wxT("id")) locale = wxT("in") + locale.Mid(3);
- else if (mid == wxT("yi")) locale = wxT("ji") + locale.Mid(3);
- retloc = wxSetlocale(LC_ALL, locale);
+ if (mid == wxT("he"))
+ locale = wxT("iw") + locale.Mid(3);
+ else if (mid == wxT("id"))
+ locale = wxT("in") + locale.Mid(3);
+ else if (mid == wxT("yi"))
+ locale = wxT("ji") + locale.Mid(3);
+ else if (mid == wxT("nb"))
+ locale = wxT("no_NO");
+ else if (mid == wxT("nn"))
+ locale = wxT("no_NY");
+
+ retloc = wxSetlocaleTryUTF(LC_ALL, locale);
}
- if (retloc == NULL)
+ if ( !retloc )
{
// (This time, we changed locale in previous if-branch, so try again.)
// Some C libraries don't like xx_YY form and require xx only
- retloc = wxSetlocale(LC_ALL, locale.Mid(0,2));
+ retloc = wxSetlocaleTryUTF(LC_ALL, locale.Mid(0,2));
}
- if (retloc == NULL)
+ if ( !retloc )
{
wxLogError(wxT("Cannot set locale to '%s'."), locale.c_str());
- return FALSE;
+ return false;
}
#elif defined(__WIN32__)
+
+ #if wxUSE_UNICODE && (defined(__VISUALC__) || defined(__MINGW32__))
+ // NB: setlocale() from msvcrt.dll (used by VC++ and Mingw)
+ // can't set locale to language that can only be written using
+ // Unicode. Therefore wxSetlocale call failed, but we don't want
+ // to report it as an error -- so that at least message catalogs
+ // can be used. Watch for code marked with
+ // #ifdef SETLOCALE_FAILS_ON_UNICODE_LANGS bellow.
+ #define SETLOCALE_FAILS_ON_UNICODE_LANGS
+ #endif
+
+#if !wxUSE_UNICODE
+ const
+#endif
+ wxMB2WXbuf retloc = wxT("C");
if (language != wxLANGUAGE_DEFAULT)
{
if (info->WinLang == 0)
{
wxLogWarning(wxT("Locale '%s' not supported by OS."), name.c_str());
- retloc = wxT("C");
+ // retloc already set to "C"
}
else
{
+ int codepage
+ #ifdef SETLOCALE_FAILS_ON_UNICODE_LANGS
+ = -1
+ #endif
+ ;
wxUint32 lcid = MAKELCID(MAKELANGID(info->WinLang, info->WinSublang),
SORT_DEFAULT);
- if (SetThreadLocale(lcid))
- retloc = wxSetlocale(LC_ALL, wxEmptyString);
+ // FIXME
+#ifndef __WXWINCE__
+ SetThreadLocale(lcid);
+#endif
+ // NB: we must translate LCID to CRT's setlocale string ourselves,
+ // because SetThreadLocale does not modify change the
+ // interpretation of setlocale(LC_ALL, "") call:
+ wxChar buffer[256];
+ buffer[0] = wxT('\0');
+ GetLocaleInfo(lcid, LOCALE_SENGLANGUAGE, buffer, 256);
+ locale << buffer;
+ if (GetLocaleInfo(lcid, LOCALE_SENGCOUNTRY, buffer, 256) > 0)
+ locale << wxT("_") << buffer;
+ if (GetLocaleInfo(lcid, LOCALE_IDEFAULTANSICODEPAGE, buffer, 256) > 0)
+ {
+ codepage = wxAtoi(buffer);
+ if (codepage != 0)
+ locale << wxT(".") << buffer;
+ }
+ if (locale.IsEmpty())
+ {
+ wxLogLastError(wxT("SetThreadLocale"));
+ wxLogError(wxT("Cannot set locale to language %s."), name.c_str());
+ return false;
+ }
else
{
- // Windows9X doesn't support SetThreadLocale, so we must
- // translate LCID to CRT's setlocale string ourselves
- locale.Empty();
- if (GetLastError() == ERROR_CALL_NOT_IMPLEMENTED)
- {
- wxChar buffer[256];
- buffer[0] = wxT('\0');
- GetLocaleInfo(lcid, LOCALE_SENGLANGUAGE, buffer, 256);
- locale << buffer;
- if (GetLocaleInfo(lcid, LOCALE_SENGCOUNTRY, buffer, 256) > 0)
- locale << wxT("_") << buffer;
- }
- if (locale.IsEmpty())
+ // FIXME
+#ifndef __WXWINCE__
+ retloc = wxSetlocale(LC_ALL, locale);
+#endif
+#ifdef SETLOCALE_FAILS_ON_UNICODE_LANGS
+ if (codepage == 0 && (const wxChar*)retloc == NULL)
{
- wxLogLastError(wxT("SetThreadLocale"));
- wxLogError(wxT("Cannot set locale to language %s."), name.c_str());
- return FALSE;
+ retloc = wxT("C");
}
- else
- retloc = wxSetlocale(LC_ALL, locale);
+#endif
}
}
}
else
+ {
+ // FIXME
+#ifndef __WXWINCE__
retloc = wxSetlocale(LC_ALL, wxEmptyString);
+#else
+ retloc = NULL;
+#endif
+#ifdef SETLOCALE_FAILS_ON_UNICODE_LANGS
+ if ((const wxChar*)retloc == NULL)
+ {
+ wxChar buffer[16];
+ if (GetLocaleInfo(LOCALE_USER_DEFAULT,
+ LOCALE_IDEFAULTANSICODEPAGE, buffer, 16) > 0 &&
+ wxStrcmp(buffer, wxT("0")) == 0)
+ {
+ retloc = wxT("C");
+ }
+ }
+#endif
+ }
- if (retloc == NULL)
+ if ( !retloc )
{
wxLogError(wxT("Cannot set locale to language %s."), name.c_str());
- return FALSE;
+ return false;
}
-
+#elif defined(__WXMAC__) || defined(__WXPM__)
+ wxMB2WXbuf retloc = wxSetlocale(LC_ALL , wxEmptyString);
#else
- return FALSE;
+ return false;
+ #define WX_NO_LOCALE_SUPPORT
#endif
- return Init(name, canonical, wxString(retloc),
- (flags & wxLOCALE_LOAD_DEFAULT) != 0,
- (flags & wxLOCALE_CONV_ENCODING) != 0);
+#ifndef WX_NO_LOCALE_SUPPORT
+ wxChar *szLocale = retloc ? wxStrdup(retloc) : NULL;
+ bool ret = Init(name, canonical, retloc,
+ (flags & wxLOCALE_LOAD_DEFAULT) != 0,
+ (flags & wxLOCALE_CONV_ENCODING) != 0);
+ free(szLocale);
+
+ if (IsOk()) // setlocale() succeeded
+ m_language = lang;
+
+ return ret;
+#endif
}
size_t i = 0,
count = ms_languagesDB->GetCount();
-#if defined(__UNIX__)
+#if defined(__UNIX__) && !defined(__WXMAC__)
// first get the string identifying the language from the environment
wxString langFull;
if (!wxGetEnv(wxT("LC_ALL"), &langFull) &&
return wxLANGUAGE_ENGLISH;
}
- if ( langFull == _T("C") )
+ if ( langFull == _T("C") || langFull == _T("POSIX") )
{
// default C locale
return wxLANGUAGE_ENGLISH;
// the language string has the following form
//
- // lang[_LANG[.encoding]]
+ // lang[_LANG][.encoding][@modifier]
//
- // where lang is the primary language, LANG is a sublang
+ // (see environ(5) in the Open Unix specification)
+ //
+ // where lang is the primary language, LANG is a sublang/territory,
+ // encoding is the charset to use and modifier "allows the user to select
+ // a specific instance of localization data within a single category"
//
// for example, the following strings are valid:
// fr
// fr_FR
// de_DE.iso88591
+ // de_DE@euro
+ // de_DE.iso88591@euro
// for now we don't use the encoding, although we probably should (doing
// translations of the msg catalogs on the fly as required) (TODO)
- langFull = langFull.BeforeFirst(_T('.'));
+ //
+ // we don't use the modifiers neither but we probably should translate
+ // "euro" into iso885915
+ size_t posEndLang = langFull.find_first_of(_T("@."));
+ if ( posEndLang != wxString::npos )
+ {
+ langFull.Truncate(posEndLang);
+ }
// in addition to the format above, we also can have full language names
// in LANG env var - for example, SuSE is known to use LANG="german" - so
wxString lang;
if ( langOrig == wxT("iw"))
lang = _T("he");
- else if ( langOrig == wxT("in") )
+ else if (langOrig == wxT("in"))
lang = wxT("id");
- else if ( langOrig == wxT("ji") )
+ else if (langOrig == wxT("ji"))
lang = wxT("yi");
+ else if (langOrig == wxT("no_NO"))
+ lang = wxT("nb_NO");
+ else if (langOrig == wxT("no_NY"))
+ lang = wxT("nn_NO");
+ else if (langOrig == wxT("no"))
+ lang = wxT("nb_NO");
else
lang = langOrig;
}
}
}
+#elif defined(__WXMAC__)
+ const wxChar * lc = NULL ;
+ long lang = GetScriptVariable( smSystemScript, smScriptLang) ;
+ switch( GetScriptManagerVariable( smRegionCode ) ) {
+ case verUS :
+ lc = wxT("en_US") ;
+ break ;
+ case verFrance :
+ lc = wxT("fr_FR") ;
+ break ;
+ case verBritain :
+ lc = wxT("en_GB") ;
+ break ;
+ case verGermany :
+ lc = wxT("de_DE") ;
+ break ;
+ case verItaly :
+ lc = wxT("it_IT") ;
+ break ;
+ case verNetherlands :
+ lc = wxT("nl_NL") ;
+ break ;
+ case verFlemish :
+ lc = wxT("nl_BE") ;
+ break ;
+ case verSweden :
+ lc = wxT("sv_SE" );
+ break ;
+ case verSpain :
+ lc = wxT("es_ES" );
+ break ;
+ case verDenmark :
+ lc = wxT("da_DK") ;
+ break ;
+ case verPortugal :
+ lc = wxT("pt_PT") ;
+ break ;
+ case verFrCanada:
+ lc = wxT("fr_CA") ;
+ break ;
+ case verNorway:
+ lc = wxT("nb_NO") ;
+ break ;
+ case verIsrael:
+ lc = wxT("iw_IL") ;
+ break ;
+ case verJapan:
+ lc = wxT("ja_JP") ;
+ break ;
+ case verAustralia:
+ lc = wxT("en_AU") ;
+ break ;
+ case verArabic:
+ lc = wxT("ar") ;
+ break ;
+ case verFinland:
+ lc = wxT("fi_FI") ;
+ break ;
+ case verFrSwiss:
+ lc = wxT("fr_CH") ;
+ break ;
+ case verGrSwiss:
+ lc = wxT("de_CH") ;
+ break ;
+ case verGreece:
+ lc = wxT("el_GR") ;
+ break ;
+ case verIceland:
+ lc = wxT("is_IS") ;
+ break ;
+ case verMalta:
+ lc = wxT("mt_MT") ;
+ break ;
+ case verCyprus:
+ // _CY is not part of wx, so we have to translate according to the system language
+ if ( lang == langGreek ) {
+ lc = wxT("el_GR") ;
+ }
+ else if ( lang == langTurkish ) {
+ lc = wxT("tr_TR") ;
+ }
+ break ;
+ case verTurkey:
+ lc = wxT("tr_TR") ;
+ break ;
+ case verYugoCroatian:
+ lc = wxT("hr_HR") ;
+ break ;
+ case verIndiaHindi:
+ lc = wxT("hi_IN") ;
+ break ;
+ case verPakistanUrdu:
+ lc = wxT("ur_PK") ;
+ break ;
+ case verTurkishModified:
+ lc = wxT("tr_TR") ;
+ break ;
+ case verItalianSwiss:
+ lc = wxT("it_CH") ;
+ break ;
+ case verInternational:
+ lc = wxT("en") ;
+ break ;
+ case verRomania:
+ lc = wxT("ro_RO") ;
+ break ;
+ case verGreecePoly:
+ lc = wxT("el_GR") ;
+ break ;
+ case verLithuania:
+ lc = wxT("lt_LT") ;
+ break ;
+ case verPoland:
+ lc = wxT("pl_PL") ;
+ break ;
+ case verMagyar :
+ case verHungary:
+ lc = wxT("hu_HU") ;
+ break ;
+ case verEstonia:
+ lc = wxT("et_EE") ;
+ break ;
+ case verLatvia:
+ lc = wxT("lv_LV") ;
+ break ;
+ case verSami:
+ // not known
+ break ;
+ case verFaroeIsl:
+ lc = wxT("fo_FO") ;
+ break ;
+ case verIran:
+ lc = wxT("fa_IR") ;
+ break ;
+ case verRussia:
+ lc = wxT("ru_RU") ;
+ break ;
+ case verIreland:
+ lc = wxT("ga_IE") ;
+ break ;
+ case verKorea:
+ lc = wxT("ko_KR") ;
+ break ;
+ case verChina:
+ lc = wxT("zh_CN") ;
+ break ;
+ case verTaiwan:
+ lc = wxT("zh_TW") ;
+ break ;
+ case verThailand:
+ lc = wxT("th_TH") ;
+ break ;
+ case verCzech:
+ lc = wxT("cs_CZ") ;
+ break ;
+ case verSlovak:
+ lc = wxT("sk_SK") ;
+ break ;
+ case verBengali:
+ lc = wxT("bn") ;
+ break ;
+ case verByeloRussian:
+ lc = wxT("be_BY") ;
+ break ;
+ case verUkraine:
+ lc = wxT("uk_UA") ;
+ break ;
+ case verGreeceAlt:
+ lc = wxT("el_GR") ;
+ break ;
+ case verSerbian:
+ lc = wxT("sr_YU") ;
+ break ;
+ case verSlovenian:
+ lc = wxT("sl_SI") ;
+ break ;
+ case verMacedonian:
+ lc = wxT("mk_MK") ;
+ break ;
+ case verCroatia:
+ lc = wxT("hr_HR") ;
+ break ;
+ case verBrazil:
+ lc = wxT("pt_BR ") ;
+ break ;
+ case verBulgaria:
+ lc = wxT("bg_BG") ;
+ break ;
+ case verCatalonia:
+ lc = wxT("ca_ES") ;
+ break ;
+ case verScottishGaelic:
+ lc = wxT("gd") ;
+ break ;
+ case verManxGaelic:
+ lc = wxT("gv") ;
+ break ;
+ case verBreton:
+ lc = wxT("br") ;
+ break ;
+ case verNunavut:
+ lc = wxT("iu_CA") ;
+ break ;
+ case verWelsh:
+ lc = wxT("cy") ;
+ break ;
+ case verIrishGaelicScript:
+ lc = wxT("ga_IE") ;
+ break ;
+ case verEngCanada:
+ lc = wxT("en_CA") ;
+ break ;
+ case verBhutan:
+ lc = wxT("dz_BT") ;
+ break ;
+ case verArmenian:
+ lc = wxT("hy_AM") ;
+ break ;
+ case verGeorgian:
+ lc = wxT("ka_GE") ;
+ break ;
+ case verSpLatinAmerica:
+ lc = wxT("es_AR") ;
+ break ;
+ case verTonga:
+ lc = wxT("to_TO" );
+ break ;
+ case verFrenchUniversal:
+ lc = wxT("fr_FR") ;
+ break ;
+ case verAustria:
+ lc = wxT("de_AT") ;
+ break ;
+ case verGujarati:
+ lc = wxT("gu_IN") ;
+ break ;
+ case verPunjabi:
+ lc = wxT("pa") ;
+ break ;
+ case verIndiaUrdu:
+ lc = wxT("ur_IN") ;
+ break ;
+ case verVietnam:
+ lc = wxT("vi_VN") ;
+ break ;
+ case verFrBelgium:
+ lc = wxT("fr_BE") ;
+ break ;
+ case verUzbek:
+ lc = wxT("uz_UZ") ;
+ break ;
+ case verSingapore:
+ lc = wxT("zh_SG") ;
+ break ;
+ case verNynorsk:
+ lc = wxT("nn_NO") ;
+ break ;
+ case verAfrikaans:
+ lc = wxT("af_ZA") ;
+ break ;
+ case verEsperanto:
+ lc = wxT("eo") ;
+ break ;
+ case verMarathi:
+ lc = wxT("mr_IN") ;
+ break ;
+ case verTibetan:
+ lc = wxT("bo") ;
+ break ;
+ case verNepal:
+ lc = wxT("ne_NP") ;
+ break ;
+ case verGreenland:
+ lc = wxT("kl_GL") ;
+ break ;
+ default :
+ break ;
+ }
+ for ( i = 0; i < count; i++ )
+ {
+ if ( ms_languagesDB->Item(i).CanonicalName == lc )
+ {
+ break;
+ }
+ }
+
#elif defined(__WIN32__)
LCID lcid = GetUserDefaultLCID();
if ( lcid != 0 )
return wxLANGUAGE_UNKNOWN;
}
-/*static*/ void wxLocale::AddLanguage(const wxLanguageInfo& info)
+// ----------------------------------------------------------------------------
+// encoding stuff
+// ----------------------------------------------------------------------------
+
+// this is a bit strange as under Windows we get the encoding name using its
+// numeric value and under Unix we do it the other way round, but this just
+// reflects the way different systems provide the encoding info
+
+/* static */
+wxString wxLocale::GetSystemEncodingName()
+{
+ wxString encname;
+
+#if defined(__WIN32__) && !defined(__WXMICROWIN__)
+ // FIXME: what is the error return value for GetACP()?
+ UINT codepage = ::GetACP();
+ encname.Printf(_T("windows-%u"), codepage);
+#elif defined(__UNIX_LIKE__)
+
+#if defined(HAVE_LANGINFO_H) && defined(CODESET)
+ // GNU libc provides current character set this way (this conforms
+ // to Unix98)
+ char *oldLocale = strdup(setlocale(LC_CTYPE, NULL));
+ setlocale(LC_CTYPE, "");
+ const char *alang = nl_langinfo(CODESET);
+ setlocale(LC_CTYPE, oldLocale);
+ free(oldLocale);
+
+ if ( alang )
+ {
+ // 7 bit ASCII encoding has several alternative names which we should
+ // recognize to avoid warnings about unrecognized encoding on each
+ // program startup
+
+ // nl_langinfo() under Solaris returns 646 by default which stands for
+ // ISO-646, i.e. 7 bit ASCII
+ //
+ // and recent glibc call it ANSI_X3.4-1968...
+ //
+ // HP-UX uses HP-Roman8 cset which is not the same as ASCII (see RFC
+ // 1345 for its definition) but must be recognized as otherwise HP
+ // users get a warning about it on each program startup, so handle it
+ // here -- but it would be obviously better to add real supprot to it,
+ // of course!
+ if ( strcmp(alang, "646") == 0
+ || strcmp(alang, "ANSI_X3.4-1968") == 0
+#ifdef __HPUX__
+ || strcmp(alang, "roman8") == 0
+#endif // __HPUX__
+ )
+ {
+ encname = _T("US-ASCII");
+ }
+ else
+ {
+ encname = wxString::FromAscii( alang );
+ }
+ }
+ else
+#endif // HAVE_LANGINFO_H
+ {
+ // if we can't get at the character set directly, try to see if it's in
+ // the environment variables (in most cases this won't work, but I was
+ // out of ideas)
+ char *lang = getenv( "LC_ALL");
+ char *dot = lang ? strchr(lang, '.') : (char *)NULL;
+ if (!dot)
+ {
+ lang = getenv( "LC_CTYPE" );
+ if ( lang )
+ dot = strchr(lang, '.' );
+ }
+ if (!dot)
+ {
+ lang = getenv( "LANG");
+ if ( lang )
+ dot = strchr(lang, '.');
+ }
+
+ if ( dot )
+ {
+ encname = wxString::FromAscii( dot+1 );
+ }
+ }
+#endif // Win32/Unix
+
+ return encname;
+}
+
+/* static */
+wxFontEncoding wxLocale::GetSystemEncoding()
+{
+#if defined(__WIN32__) && !defined(__WXMICROWIN__)
+ UINT codepage = ::GetACP();
+
+ // wxWidgets only knows about CP1250-1257, 932, 936, 949, 950
+ if ( codepage >= 1250 && codepage <= 1257 )
+ {
+ return (wxFontEncoding)(wxFONTENCODING_CP1250 + codepage - 1250);
+ }
+
+ if ( codepage == 932 )
+ {
+ return wxFONTENCODING_CP932;
+ }
+
+ if ( codepage == 936 )
+ {
+ return wxFONTENCODING_CP936;
+ }
+
+ if ( codepage == 949 )
+ {
+ return wxFONTENCODING_CP949;
+ }
+
+ if ( codepage == 950 )
+ {
+ return wxFONTENCODING_CP950;
+ }
+#elif defined(__WXMAC__)
+ TextEncoding encoding = 0 ;
+#if TARGET_CARBON
+ encoding = CFStringGetSystemEncoding() ;
+#else
+ UpgradeScriptInfoToTextEncoding ( smSystemScript , kTextLanguageDontCare , kTextRegionDontCare , NULL , &encoding ) ;
+#endif
+ return wxMacGetFontEncFromSystemEnc( encoding ) ;
+#elif defined(__UNIX_LIKE__) && wxUSE_FONTMAP
+ wxString encname = GetSystemEncodingName();
+ if ( !encname.empty() )
+ {
+ wxFontEncoding enc = wxFontMapper::Get()->
+ CharsetToEncoding(encname, false /* not interactive */);
+
+ // on some modern Linux systems (RedHat 8) the default system locale
+ // is UTF8 -- but it isn't supported by wxGTK in ANSI build at all so
+ // don't even try to use it in this case
+#if !wxUSE_UNICODE
+ if ( enc == wxFONTENCODING_UTF8 )
+ {
+ // the most similar supported encoding...
+ enc = wxFONTENCODING_ISO8859_1;
+ }
+#endif // !wxUSE_UNICODE
+
+ // this should probably be considered as a bug in CharsetToEncoding():
+ // it shouldn't return wxFONTENCODING_DEFAULT at all - but it does it
+ // for US-ASCII charset
+ //
+ // we, OTOH, definitely shouldn't return it as it doesn't make sense at
+ // all (which encoding is it?)
+ if ( enc != wxFONTENCODING_DEFAULT )
+ {
+ return enc;
+ }
+ //else: return wxFONTENCODING_SYSTEM below
+ }
+#endif // Win32/Unix
+
+ return wxFONTENCODING_SYSTEM;
+}
+
+/* static */
+void wxLocale::AddLanguage(const wxLanguageInfo& info)
{
CreateLanguagesDB();
ms_languagesDB->Add(info);
}
+/* static */
+const wxLanguageInfo *wxLocale::GetLanguageInfo(int lang)
+{
+ CreateLanguagesDB();
+
+ // calling GetLanguageInfo(wxLANGUAGE_DEFAULT) is a natural thing to do, so
+ // make it work
+ if ( lang == wxLANGUAGE_DEFAULT )
+ lang = GetSystemLanguage();
+
+ const size_t count = ms_languagesDB->GetCount();
+ for ( size_t i = 0; i < count; i++ )
+ {
+ if ( ms_languagesDB->Item(i).Language == lang )
+ {
+ return &ms_languagesDB->Item(i);
+ }
+ }
+
+ return NULL;
+}
+
+/* static */
+wxString wxLocale::GetLanguageName(int lang)
+{
+ const wxLanguageInfo *info = GetLanguageInfo(lang);
+ if ( !info )
+ return wxEmptyString;
+ else
+ return info->Description;
+}
+
+/* static */
+const wxLanguageInfo *wxLocale::FindLanguageInfo(const wxString& locale)
+{
+ CreateLanguagesDB();
+
+ const wxLanguageInfo *infoRet = NULL;
+
+ const size_t count = ms_languagesDB->GetCount();
+ for ( size_t i = 0; i < count; i++ )
+ {
+ const wxLanguageInfo *info = &ms_languagesDB->Item(i);
+
+ if ( wxStricmp(locale, info->CanonicalName) == 0 ||
+ wxStricmp(locale, info->Description) == 0 )
+ {
+ // exact match, stop searching
+ infoRet = info;
+ break;
+ }
+
+ if ( wxStricmp(locale, info->CanonicalName.BeforeFirst(_T('_'))) == 0 )
+ {
+ // a match -- but maybe we'll find an exact one later, so continue
+ // looking
+ //
+ // OTOH, maybe we had already found a language match and in this
+ // case don't overwrite it becauce the entry for the default
+ // country always appears first in ms_languagesDB
+ if ( !infoRet )
+ infoRet = info;
+ }
+ }
+
+ return infoRet;
+}
+
wxString wxLocale::GetSysName() const
{
+ // FIXME
+#ifndef __WXWINCE__
return wxSetlocale(LC_ALL, NULL);
+#else
+ return wxEmptyString;
+#endif
}
// clean up
// restore old locale
wxSetLocale(m_pOldLocale);
+ // FIXME
+#ifndef __WXWINCE__
wxSetlocale(LC_ALL, m_pszOldLocale);
+#endif
+ free((wxChar *)m_pszOldLocale); // const_cast
}
// get the translation of given string in current locale
-const wxMB2WXbuf wxLocale::GetString(const wxChar *szOrigString,
- const wxChar *szDomain) const
+const wxChar *wxLocale::GetString(const wxChar *szOrigString,
+ const wxChar *szDomain) const
{
- if ( wxIsEmpty(szOrigString) )
- return szDomain;
+ return GetString(szOrigString, szOrigString, size_t(-1), szDomain);
+}
- const char *pszTrans = NULL;
-#if wxUSE_UNICODE
- const wxWX2MBbuf szOrgString = wxConvCurrent->cWX2MB(szOrigString);
-#else // ANSI
- #define szOrgString szOrigString
-#endif // Unicode/ANSI
-
- wxMsgCatalog *pMsgCat;
- if ( szDomain != NULL ) {
- pMsgCat = FindCatalog(szDomain);
-
- // does the catalog exist?
- if ( pMsgCat != NULL )
- pszTrans = pMsgCat->GetString(szOrgString);
- }
- else {
- // search in all domains
- for ( pMsgCat = m_pMsgCat; pMsgCat != NULL; pMsgCat = pMsgCat->m_pNext ) {
- pszTrans = pMsgCat->GetString(szOrgString);
- if ( pszTrans != NULL ) // take the first found
- break;
+const wxChar *wxLocale::GetString(const wxChar *szOrigString,
+ const wxChar *szOrigString2,
+ size_t n,
+ const wxChar *szDomain) const
+{
+ if ( wxIsEmpty(szOrigString) )
+ return wxEmptyString;
+
+ const wxChar *pszTrans = NULL;
+ wxMsgCatalog *pMsgCat;
+
+ if ( szDomain != NULL )
+ {
+ pMsgCat = FindCatalog(szDomain);
+
+ // does the catalog exist?
+ if ( pMsgCat != NULL )
+ pszTrans = pMsgCat->GetString(szOrigString, n);
+ }
+ else
+ {
+ // search in all domains
+ for ( pMsgCat = m_pMsgCat; pMsgCat != NULL; pMsgCat = pMsgCat->m_pNext )
+ {
+ pszTrans = pMsgCat->GetString(szOrigString, n);
+ if ( pszTrans != NULL ) // take the first found
+ break;
+ }
}
- }
- if ( pszTrans == NULL ) {
+ if ( pszTrans == NULL )
+ {
#ifdef __WXDEBUG__
- if ( !NoTransErr::Suppress() ) {
- NoTransErr noTransErr;
+ if ( !NoTransErr::Suppress() )
+ {
+ NoTransErr noTransErr;
- if ( szDomain != NULL )
- {
- wxLogDebug(_T("string '%s' not found in domain '%s' for locale '%s'."),
- szOrigString, szDomain, m_strLocale.c_str());
- }
- else
- {
- wxLogDebug(_T("string '%s' not found in locale '%s'."),
- szOrigString, m_strLocale.c_str());
- }
- }
+ if ( szDomain != NULL )
+ {
+ wxLogTrace(_T("i18n"),
+ _T("string '%s'[%lu] not found in domain '%s' for locale '%s'."),
+ szOrigString, (unsigned long)n,
+ szDomain, m_strLocale.c_str());
+
+ }
+ else
+ {
+ wxLogTrace(_T("i18n"),
+ _T("string '%s'[%lu] not found in locale '%s'."),
+ szOrigString, (unsigned long)n, m_strLocale.c_str());
+ }
+ }
#endif // __WXDEBUG__
- return (wxMB2WXbuf)(szOrigString);
- }
- else
- {
- return wxConvertMB2WX(pszTrans); // or preferably wxCSConv(charset).cMB2WX(pszTrans) or something,
- // a macro similar to wxConvertMB2WX could be written for that
- }
+ if (n == size_t(-1))
+ return szOrigString;
+ else
+ return n == 1 ? szOrigString : szOrigString2;
+ }
- #undef szOrgString
+ return pszTrans;
}
+wxString wxLocale::GetHeaderValue( const wxChar* szHeader,
+ const wxChar* szDomain ) const
+{
+ if ( wxIsEmpty(szHeader) )
+ return wxEmptyString;
+
+ wxChar const * pszTrans = NULL;
+ wxMsgCatalog *pMsgCat;
+
+ if ( szDomain != NULL )
+ {
+ pMsgCat = FindCatalog(szDomain);
+
+ // does the catalog exist?
+ if ( pMsgCat == NULL )
+ return wxEmptyString;
+
+ pszTrans = pMsgCat->GetString(wxT(""), (size_t)-1);
+ }
+ else
+ {
+ // search in all domains
+ for ( pMsgCat = m_pMsgCat; pMsgCat != NULL; pMsgCat = pMsgCat->m_pNext )
+ {
+ pszTrans = pMsgCat->GetString(wxT(""), (size_t)-1);
+ if ( pszTrans != NULL ) // take the first found
+ break;
+ }
+ }
+
+ if ( wxIsEmpty(pszTrans) )
+ return wxEmptyString;
+
+ wxChar const * pszFound = wxStrstr(pszTrans, szHeader);
+ if ( pszFound == NULL )
+ return wxEmptyString;
+
+ pszFound += wxStrlen(szHeader) + 2 /* ': ' */;
+
+ // Every header is separated by \n
+
+ wxChar const * pszEndLine = wxStrchr(pszFound, wxT('\n'));
+ if ( pszEndLine == NULL ) pszEndLine = pszFound + wxStrlen(pszFound);
+
+
+ // wxString( wxChar*, length);
+ wxString retVal( pszFound, pszEndLine - pszFound );
+
+ return retVal;
+}
+
+
// find catalog by name in a linked list, return NULL if !found
wxMsgCatalog *wxLocale::FindCatalog(const wxChar *szDomain) const
{
-// linear search in the linked list
- wxMsgCatalog *pMsgCat;
- for ( pMsgCat = m_pMsgCat; pMsgCat != NULL; pMsgCat = pMsgCat->m_pNext ) {
- if ( wxStricmp(pMsgCat->GetName(), szDomain) == 0 )
- return pMsgCat;
- }
+ // linear search in the linked list
+ wxMsgCatalog *pMsgCat;
+ for ( pMsgCat = m_pMsgCat; pMsgCat != NULL; pMsgCat = pMsgCat->m_pNext )
+ {
+ if ( wxStricmp(pMsgCat->GetName(), szDomain) == 0 )
+ return pMsgCat;
+ }
- return NULL;
+ return NULL;
}
// check if the given catalog is loaded
// add a catalog to our linked list
bool wxLocale::AddCatalog(const wxChar *szDomain)
+{
+ return AddCatalog(szDomain, wxLANGUAGE_ENGLISH, NULL);
+}
+
+// add a catalog to our linked list
+bool wxLocale::AddCatalog(const wxChar *szDomain,
+ wxLanguage msgIdLanguage,
+ const wxChar *msgIdCharset)
+
{
wxMsgCatalog *pMsgCat = new wxMsgCatalog;
- if ( pMsgCat->Load(m_strShort, szDomain, m_bConvertEncoding) ) {
+ if ( pMsgCat->Load(m_strShort, szDomain, msgIdCharset, m_bConvertEncoding) ) {
// add it to the head of the list so that in GetString it will
// be searched before the catalogs added earlier
pMsgCat->m_pNext = m_pMsgCat;
m_pMsgCat = pMsgCat;
- return TRUE;
+ return true;
}
else {
// don't add it because it couldn't be loaded anyway
delete pMsgCat;
- return FALSE;
+ // It is OK to not load catalog if the msgid language and m_language match,
+ // in which case we can directly display the texts embedded in program's
+ // source code:
+ if (m_language == msgIdLanguage)
+ return true;
+
+ // If there's no exact match, we may still get partial match where the
+ // (basic) language is same, but the country differs. For example, it's
+ // permitted to use en_US strings from sources even if m_language is en_GB:
+ const wxLanguageInfo *msgIdLangInfo = GetLanguageInfo(msgIdLanguage);
+ if ( msgIdLangInfo &&
+ msgIdLangInfo->CanonicalName.Mid(0, 2) == m_strShort.Mid(0, 2) )
+ {
+ return true;
+ }
+
+ return false;
}
}
+// ----------------------------------------------------------------------------
+// accessors for locale-dependent data
+// ----------------------------------------------------------------------------
+
+#ifdef __WXMSW__
+
+/* static */
+wxString wxLocale::GetInfo(wxLocaleInfo index, wxLocaleCategory WXUNUSED(cat))
+{
+ wxString str;
+ wxChar buffer[256];
+ size_t count;
+ buffer[0] = wxT('\0');
+ switch (index)
+ {
+ case wxLOCALE_DECIMAL_POINT:
+ count = ::GetLocaleInfo(LOCALE_USER_DEFAULT, LOCALE_SDECIMAL, buffer, 256);
+ if (!count)
+ str << wxT(".");
+ else
+ str << buffer;
+ break;
+#if 0
+ case wxSYS_LIST_SEPARATOR:
+ count = ::GetLocaleInfo(LOCALE_USER_DEFAULT, LOCALE_SLIST, buffer, 256);
+ if (!count)
+ str << wxT(",");
+ else
+ str << buffer;
+ break;
+ case wxSYS_LEADING_ZERO: // 0 means no leading zero, 1 means leading zero
+ count = ::GetLocaleInfo(LOCALE_USER_DEFAULT, LOCALE_ILZERO, buffer, 256);
+ if (!count)
+ str << wxT("0");
+ else
+ str << buffer;
+ break;
+#endif
+ default:
+ wxFAIL_MSG(wxT("Unknown System String !"));
+ }
+ return str;
+}
+
+#else // !__WXMSW__
+
+/* static */
+wxString wxLocale::GetInfo(wxLocaleInfo index, wxLocaleCategory cat)
+{
+ struct lconv *locale_info = localeconv();
+ switch (cat)
+ {
+ case wxLOCALE_CAT_NUMBER:
+ switch (index)
+ {
+ case wxLOCALE_THOUSANDS_SEP:
+ return wxString(locale_info->thousands_sep,
+ *wxConvCurrent);
+ case wxLOCALE_DECIMAL_POINT:
+ return wxString(locale_info->decimal_point,
+ *wxConvCurrent);
+ default:
+ return wxEmptyString;
+ }
+ case wxLOCALE_CAT_MONEY:
+ switch (index)
+ {
+ case wxLOCALE_THOUSANDS_SEP:
+ return wxString(locale_info->mon_thousands_sep,
+ *wxConvCurrent);
+ case wxLOCALE_DECIMAL_POINT:
+ return wxString(locale_info->mon_decimal_point,
+ *wxConvCurrent);
+ default:
+ return wxEmptyString;
+ }
+ default:
+ return wxEmptyString;
+ }
+}
+
+#endif // __WXMSW__/!__WXMSW__
+
// ----------------------------------------------------------------------------
// global functions and variables
// ----------------------------------------------------------------------------
DECLARE_DYNAMIC_CLASS(wxLocaleModule)
public:
wxLocaleModule() {}
- bool OnInit() { return TRUE; }
+ bool OnInit() { return true; }
void OnExit() { wxLocale::DestroyLanguagesDB(); }
};
// This table is generated by misc/languages/genlang.py
// When making changes, please put them into misc/languages/langtabl.txt
-#ifndef __WIN32__
+#if !defined(__WIN32__) || defined(__WXMICROWIN__)
#define SETWINLANG(info,lang,sublang)
#define LNG(wxlang, canonical, winlang, winsublang, desc) \
info.Language = wxlang; \
info.CanonicalName = wxT(canonical); \
- info.Description = desc; \
+ info.Description = wxT(desc); \
SETWINLANG(info, winlang, winsublang) \
AddLanguage(info);
wxLanguageInfo info;
wxStringTokenizer tkn;
- LNG(wxLANGUAGE_ABKHAZIAN, "ab" , 0 , 0 , "Abkhazian")
+ LNG(wxLANGUAGE_ABKHAZIAN, "ab" , 0 , 0 , "Abkhazian")
LNG(wxLANGUAGE_AFAR, "aa" , 0 , 0 , "Afar")
LNG(wxLANGUAGE_AFRIKAANS, "af_ZA", LANG_AFRIKAANS , SUBLANG_DEFAULT , "Afrikaans")
LNG(wxLANGUAGE_ALBANIAN, "sq_AL", LANG_ALBANIAN , SUBLANG_DEFAULT , "Albanian")
LNG(wxLANGUAGE_BURMESE, "my" , 0 , 0 , "Burmese")
LNG(wxLANGUAGE_CAMBODIAN, "km" , 0 , 0 , "Cambodian")
LNG(wxLANGUAGE_CATALAN, "ca_ES", LANG_CATALAN , SUBLANG_DEFAULT , "Catalan")
- LNG(wxLANGUAGE_CHINESE, "zh_CN", LANG_CHINESE , SUBLANG_DEFAULT , "Chinese")
+ LNG(wxLANGUAGE_CHINESE, "zh_TW", LANG_CHINESE , SUBLANG_DEFAULT , "Chinese")
LNG(wxLANGUAGE_CHINESE_SIMPLIFIED, "zh_CN", LANG_CHINESE , SUBLANG_CHINESE_SIMPLIFIED , "Chinese (Simplified)")
- LNG(wxLANGUAGE_CHINESE_TRADITIONAL, "zh_CN", LANG_CHINESE , SUBLANG_CHINESE_TRADITIONAL , "Chinese (Traditional)")
+ LNG(wxLANGUAGE_CHINESE_TRADITIONAL, "zh_TW", LANG_CHINESE , SUBLANG_CHINESE_TRADITIONAL , "Chinese (Traditional)")
LNG(wxLANGUAGE_CHINESE_HONGKONG, "zh_HK", LANG_CHINESE , SUBLANG_CHINESE_HONGKONG , "Chinese (Hongkong)")
LNG(wxLANGUAGE_CHINESE_MACAU, "zh_MO", LANG_CHINESE , SUBLANG_CHINESE_MACAU , "Chinese (Macau)")
LNG(wxLANGUAGE_CHINESE_SINGAPORE, "zh_SG", LANG_CHINESE , SUBLANG_CHINESE_SINGAPORE , "Chinese (Singapore)")
- LNG(wxLANGUAGE_CHINESE_TAIWAN, "zh_TW", 0 , 0 , "Chinese (Taiwan)")
+ LNG(wxLANGUAGE_CHINESE_TAIWAN, "zh_TW", LANG_CHINESE , SUBLANG_CHINESE_TRADITIONAL , "Chinese (Taiwan)")
LNG(wxLANGUAGE_CORSICAN, "co" , 0 , 0 , "Corsican")
LNG(wxLANGUAGE_CROATIAN, "hr_HR", LANG_CROATIAN , SUBLANG_DEFAULT , "Croatian")
LNG(wxLANGUAGE_CZECH, "cs_CZ", LANG_CZECH , SUBLANG_DEFAULT , "Czech")
LNG(wxLANGUAGE_NAURU, "na" , 0 , 0 , "Nauru")
LNG(wxLANGUAGE_NEPALI, "ne" , LANG_NEPALI , SUBLANG_DEFAULT , "Nepali")
LNG(wxLANGUAGE_NEPALI_INDIA, "ne_IN", LANG_NEPALI , SUBLANG_NEPALI_INDIA , "Nepali (India)")
- LNG(wxLANGUAGE_NORWEGIAN_BOKMAL, "no_NO", LANG_NORWEGIAN , SUBLANG_NORWEGIAN_BOKMAL , "Norwegian (Bokmal)")
+ LNG(wxLANGUAGE_NORWEGIAN_BOKMAL, "nb_NO", LANG_NORWEGIAN , SUBLANG_NORWEGIAN_BOKMAL , "Norwegian (Bokmal)")
LNG(wxLANGUAGE_NORWEGIAN_NYNORSK, "nn_NO", LANG_NORWEGIAN , SUBLANG_NORWEGIAN_NYNORSK , "Norwegian (Nynorsk)")
LNG(wxLANGUAGE_OCCITAN, "oc" , 0 , 0 , "Occitan")
LNG(wxLANGUAGE_ORIYA, "or" , LANG_ORIYA , SUBLANG_DEFAULT , "Oriya")
LNG(wxLANGUAGE_SWAHILI, "sw_KE", LANG_SWAHILI , SUBLANG_DEFAULT , "Swahili")
LNG(wxLANGUAGE_SWEDISH, "sv_SE", LANG_SWEDISH , SUBLANG_SWEDISH , "Swedish")
LNG(wxLANGUAGE_SWEDISH_FINLAND, "sv_FI", LANG_SWEDISH , SUBLANG_SWEDISH_FINLAND , "Swedish (Finland)")
- LNG(wxLANGUAGE_TAGALOG, "tl" , 0 , 0 , "Tagalog")
+ LNG(wxLANGUAGE_TAGALOG, "tl_PH", 0 , 0 , "Tagalog")
LNG(wxLANGUAGE_TAJIK, "tg" , 0 , 0 , "Tajik")
LNG(wxLANGUAGE_TAMIL, "ta" , LANG_TAMIL , SUBLANG_DEFAULT , "Tamil")
LNG(wxLANGUAGE_TATAR, "tt" , LANG_TATAR , SUBLANG_DEFAULT , "Tatar")
LNG(wxLANGUAGE_YORUBA, "yo" , 0 , 0 , "Yoruba")
LNG(wxLANGUAGE_ZHUANG, "za" , 0 , 0 , "Zhuang")
LNG(wxLANGUAGE_ZULU, "zu" , 0 , 0 , "Zulu")
-
+
};
#undef LNG
// --- --- --- generated code ends here --- --- ---
-
-
#endif // wxUSE_INTL