X-Git-Url: https://git.saurik.com/wxWidgets.git/blobdiff_plain/869c75498a12e470b610f22ff24f69401b359cd4..27c78e4552aaefac9a4db0d4453eff09cdfef2ad:/src/common/intl.cpp diff --git a/src/common/intl.cpp b/src/common/intl.cpp index b8d28fdb0d..548776f481 100644 --- a/src/common/intl.cpp +++ b/src/common/intl.cpp @@ -1,8 +1,9 @@ ///////////////////////////////////////////////////////////////////////////// // Name: src/common/intl.cpp -// Purpose: Internationalization and localisation for wxWindows +// Purpose: Internationalization and localisation for wxWidgets // Author: Vadim Zeitlin -// Modified by: +// Modified by: Michael N. Filippov +// (2003/09/30 - PluralForms support) // Created: 29/01/98 // RCS-ID: $Id$ // Copyright: (c) 1998 Vadim Zeitlin @@ -17,8 +18,16 @@ // headers // ---------------------------------------------------------------------------- -#ifdef __GNUG__ - #pragma implementation "intl.h" +#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". @@ -31,14 +40,18 @@ #if wxUSE_INTL // standard headers + +#ifndef __WXWINCE__ #include +#endif + #include #include #ifdef HAVE_LANGINFO_H #include #endif -// wxWindows +// wxWidgets #ifndef WX_PRECOMP #include "wx/string.h" #include "wx/intl.h" @@ -48,18 +61,22 @@ #include "wx/dynarray.h" #endif // WX_PRECOMP +#ifdef __WIN32__ + #include "wx/msw/private.h" +#elif defined(__UNIX_LIKE__) + #include "wx/fontmap.h" // for CharsetToEncoding() +#endif + #include "wx/file.h" +#include "wx/filename.h" #include "wx/tokenzr.h" #include "wx/module.h" #include "wx/fontmap.h" #include "wx/encconv.h" #include "wx/hashmap.h" - -#ifdef __WIN32__ - #include "wx/msw/private.h" -#elif defined(__UNIX_LIKE__) - #include "wx/fontmap.h" // for CharsetToEncoding() -#endif +#include "wx/ptr_scpd.h" +#include "wx/app.h" +#include "wx/apptrait.h" #if defined(__WXMAC__) #include "wx/mac/private.h" // includes mac headers @@ -81,14 +98,13 @@ typedef wxUint32 size_t32; const size_t32 MSGCATALOG_MAGIC = 0x950412de; const size_t32 MSGCATALOG_MAGIC_SW = 0xde120495; -// extension of ".mo" files -#define MSGCATALOG_EXTENSION _T(".mo") - // the constants describing the format of lang_LANG locale string static const size_t LEN_LANG = 2; static const size_t LEN_SUBLANG = 2; static const size_t LEN_FULL = LEN_LANG + 1 + LEN_SUBLANG; // 1 for '_' +#define TRACE_I18N _T("i18n") + // ---------------------------------------------------------------------------- // global functions // ---------------------------------------------------------------------------- @@ -132,13 +148,719 @@ static inline wxString ExtractLang(const wxString& langFull) return langFull.Left(LEN_LANG); } -// get everything else (including the leading '_') -static inline wxString ExtractNotLang(const wxString& langFull) +// get everything else (including the leading '_') +static inline wxString ExtractNotLang(const wxString& langFull) +{ + return langFull.Mid(LEN_LANG); +} + +#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); + +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) { - return langFull.Mid(LEN_LANG); + wxPluralFormsCalculatorPtr calculator(new wxPluralFormsCalculator); + if (s != NULL) + { + wxPluralFormsScanner scanner(s); + wxPluralFormsParser p(scanner); + if (!p.parse(*calculator)) + { + return NULL; + } + } + return calculator.release(); } -#endif // __UNIX__ @@ -158,10 +880,12 @@ public: ~wxMsgCatalogFile(); // load the catalog from disk (szDirPrefix corresponds to language) - bool Load(const wxChar *szDirPrefix, const wxChar *szName); + bool Load(const wxChar *szDirPrefix, const wxChar *szName, + wxPluralFormsCalculatorPtr& rPluralFormsCalculator); // fills the hash with string-translation pairs - void FillHash(wxMessagesHash& hash, bool convertEncoding) const; + void FillHash(wxMessagesHash& hash, const wxString& msgIdCharset, + bool convertEncoding) const; private: // this implementation is binary compatible with GNU gettext() version 0.10 @@ -196,6 +920,8 @@ private: 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 { @@ -211,13 +937,13 @@ private: // 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); } - wxString GetCharset() const; - bool m_bSwapped; // wrong endianness? DECLARE_NO_COPY_CLASS(wxMsgCatalogFile) @@ -235,13 +961,14 @@ class wxMsgCatalog { public: // load the catalog from disk (szDirPrefix corresponds to language) - bool Load(const wxChar *szDirPrefix, const wxChar *szName, bool bConvertEncoding = FALSE); + bool Load(const wxChar *szDirPrefix, const wxChar *szName, + const wxChar *msgIdCharset = NULL, bool bConvertEncoding = false); // get name of the catalog wxString GetName() const { return m_name; } // get the translated string: returns NULL if not found - const wxChar *GetString(const wxChar *sz) const; + 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; @@ -249,6 +976,7 @@ public: private: wxMessagesHash m_messages; // all messages in the catalog wxString m_name; // name of the domain + wxPluralFormsCalculatorPtr m_pluralFormsCalculator; }; // ---------------------------------------------------------------------------- @@ -306,11 +1034,15 @@ static wxString GetFullSearchPath(const wxChar *lang) << wxPATH_SEP; } + // TODO: use wxStandardPaths instead of all this mess!! + // 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 #ifdef __UNIX__ // add some standard ones and the one in the tree where wxWin was installed: @@ -323,31 +1055,43 @@ static wxString GetFullSearchPath(const wxChar *lang) // 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); +#if defined(__WXMAC__) + searchPath << GetAllMsgCatalogSubdirs(wxGetCwd(), lang); // generic search paths could be somewhere in the system folder preferences -#else // !Mac +#elif defined(__WXMSW__) + // look in the directory of the executable + wxString path; + wxSplitPath(wxGetFullModuleName(), &path, NULL, NULL); + searchPath << GetAllMsgCatalogSubdirs(path, lang); +#else // !Mac, !MSW searchPath << GetAllMsgCatalogSubdirs(wxT("."), lang); - #endif // platform return searchPath; } // open disk file and read in it's contents -bool wxMsgCatalogFile::Load(const wxChar *szDirPrefix, const wxChar *szName0) +bool wxMsgCatalogFile::Load(const wxChar *szDirPrefix, const wxChar *szName, + 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 - szName = szName.Left(szName.Find(wxT('.'))); + wxString searchPath; + +#if wxUSE_FONTMAP + // first look for the catalog for this language and the current locale: + // notice that we don't use the system name for the locale as this would + // force us to install catalogs in different locations depending on the + // system but always use the canonical name + wxFontEncoding encSys = wxLocale::GetSystemEncoding(); + if ( encSys != wxFONTENCODING_SYSTEM ) + { + wxString fullname(szDirPrefix); + fullname << _T('.') << wxFontMapperBase::GetEncodingName(encSys); + searchPath << GetFullSearchPath(fullname) << wxPATH_SEP; + } +#endif // wxUSE_FONTMAP - wxString searchPath = GetFullSearchPath(szDirPrefix); + + searchPath += GetFullSearchPath(szDirPrefix); const wxChar *sublocale = wxStrchr(szDirPrefix, wxT('_')); if ( sublocale ) { @@ -359,9 +1103,6 @@ bool wxMsgCatalogFile::Load(const wxChar *szDirPrefix, const wxChar *szName0) << wxPATH_SEP; } - wxString strFile = szName; - strFile += MSGCATALOG_EXTENSION; - // don't give translation errors here because the wxstd catalog might // not yet be loaded (and it's normal) // @@ -369,36 +1110,44 @@ bool wxMsgCatalogFile::Load(const wxChar *szDirPrefix, const wxChar *szName0) NoTransErr noTransErr; wxLogVerbose(_("looking for catalog '%s' in path '%s'."), - szName.c_str(), searchPath.c_str()); + szName, searchPath.c_str()); + wxLogTrace(TRACE_I18N, _T("Looking for \"%s.mo\" in \"%s\""), + szName, searchPath.c_str()); + wxFileName fn(szName); + fn.SetExt(_T("mo")); wxString strFullName; - if ( !wxFindFileInPath(&strFullName, searchPath, strFile) ) { - wxLogVerbose(_("catalog file for domain '%s' not found."), szName.c_str()); - return FALSE; + if ( !wxFindFileInPath(&strFullName, searchPath, fn.GetFullPath()) ) { + wxLogVerbose(_("catalog file for domain '%s' not found."), szName); + wxLogTrace(TRACE_I18N, _T("Catalog \"%s.mo\" not found"), szName); + return false; } // open file - wxLogVerbose(_("using catalog '%s' from '%s'."), - szName.c_str(), strFullName.c_str()); + wxLogVerbose(_("using catalog '%s' from '%s'."), szName, strFullName.c_str()); + wxLogTrace(TRACE_I18N, _T("Using catalog \"%s\"."), strFullName.c_str()); wxFile fileMsg(strFullName); if ( !fileMsg.IsOpened() ) - return FALSE; + return false; - // get the file size - off_t nSize = fileMsg.Length(); - if ( nSize == wxInvalidOffset ) - return FALSE; + // get the file size (assume it is less than 4Gb...) + wxFileOffset lenFile = fileMsg.Length(); + if ( lenFile == wxInvalidOffset ) + return false; + + size_t nSize = wx_truncate_cast(size_t, lenFile); + wxASSERT_MSG( nSize == lenFile + size_t(0), _T("message catalog bigger than 4GB?") ); // read the whole file in memory m_pData = new size_t8[nSize]; - if ( fileMsg.Read(m_pData, nSize) != nSize ) { + if ( fileMsg.Read(m_pData, nSize) != lenFile ) { wxDELETEA(m_pData); - return FALSE; + return false; } // examine header - bool bValid = (size_t)nSize > sizeof(wxMsgCatalogHeader); + bool bValid = nSize + (size_t)0 > sizeof(wxMsgCatalogHeader); wxMsgCatalogHeader *pHeader = (wxMsgCatalogHeader *)m_pData; if ( bValid ) { @@ -414,7 +1163,7 @@ bool wxMsgCatalogFile::Load(const wxChar *szDirPrefix, const wxChar *szName0) wxLogWarning(_("'%s' is not a valid message catalog."), strFullName.c_str()); wxDELETEA(m_pData); - return FALSE; + return false; } // initialize @@ -423,49 +1172,126 @@ bool wxMsgCatalogFile::Load(const wxChar *szDirPrefix, const wxChar *szName0) Swap(pHeader->ofsOrigTable)); m_pTransTable = (wxMsgTableEntry *)(m_pData + Swap(pHeader->ofsTransTable)); - m_nSize = nSize; + m_nSize = (size_t32)nSize; + + // now parse catalog's header and try to extract catalog charset and + // plural forms formula from it: + + 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 + + // 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; } -void wxMsgCatalogFile::FillHash(wxMessagesHash& hash, bool convertEncoding) const +void wxMsgCatalogFile::FillHash(wxMessagesHash& hash, + const wxString& msgIdCharset, + bool convertEncoding) const { - wxString charset = GetCharset(); +#if wxUSE_FONTMAP + // determine if we need any conversion at all + if ( convertEncoding ) + { + wxFontEncoding encCat = wxFontMapperBase::GetEncodingFromName(m_charset); + if ( encCat == wxLocale::GetSystemEncoding() ) + { + // no need to convert + convertEncoding = false; + } + } +#endif // wxUSE_FONTMAP #if wxUSE_WCHAR_T - wxCSConv *csConv = NULL; - if ( !!charset ) - csConv = new wxCSConv(charset); - - wxMBConv& inputConv = csConv ? *((wxMBConv*)csConv) : *wxConvCurrent; - - for (size_t i = 0; i < m_numStrings; i++) + // conversion to use to convert catalog strings to the GUI encoding + wxMBConv *inputConv, + *csConv = NULL; // another ptr just to be able to delete it later + if ( convertEncoding ) { - wxString key(StringAtOfs(m_pOrigTable, i), inputConv); - - #if wxUSE_UNICODE - hash[key] = wxString(StringAtOfs(m_pTransTable, i), inputConv); - #else - if ( convertEncoding ) - hash[key] = - wxString(inputConv.cMB2WC(StringAtOfs(m_pTransTable, i)), - wxConvLocal); + if ( m_charset.empty() ) + { + inputConv = wxConvCurrent; + } else - hash[key] = StringAtOfs(m_pTransTable, i); - #endif + { + inputConv = + csConv = new wxCSConv(m_charset); + } + } + else // no need to convert the encoding + { + // we still need the conversion for Unicode build +#if wxUSE_UNICODE + inputConv = wxConvCurrent; +#else // !wxUSE_UNICODE + inputConv = NULL; +#endif // wxUSE_UNICODE/!wxUSE_UNICODE } - delete csConv; -#else // !wxUSE_WCHAR_T - #if wxUSE_FONTMAP + // conversion to apply to msgid strings before looking them up: we only + // need it if the msgids are neither in 7 bit ASCII nor in the same + // encoding as the catalog + wxCSConv *sourceConv = msgIdCharset.empty() || (msgIdCharset == m_charset) + ? NULL + : new wxCSConv(msgIdCharset); + +#elif wxUSE_FONTMAP + wxASSERT_MSG( msgIdCharset == NULL, + _T("non-ASCII msgid languages only supported if wxUSE_WCHAR_T=1") ); + + wxEncodingConverter converter; if ( convertEncoding ) { wxFontEncoding targetEnc = wxFONTENCODING_SYSTEM; - wxFontEncoding enc = wxFontMapper::Get()->CharsetToEncoding(charset, FALSE); + wxFontEncoding enc = wxFontMapperBase::Get()->CharsetToEncoding(m_charset, false); if ( enc == wxFONTENCODING_SYSTEM ) { - convertEncoding = FALSE; // unknown encoding + convertEncoding = false; // unknown encoding } else { @@ -475,95 +1301,115 @@ void wxMsgCatalogFile::FillHash(wxMessagesHash& hash, bool convertEncoding) cons wxFontEncodingArray a = wxEncodingConverter::GetPlatformEquivalents(enc); if (a[0] == enc) // no conversion needed, locale uses native encoding - convertEncoding = FALSE; + convertEncoding = false; if (a.GetCount() == 0) // we don't know common equiv. under this platform - convertEncoding = FALSE; + convertEncoding = false; targetEnc = a[0]; } } if ( convertEncoding ) { - wxEncodingConverter converter; converter.Init(enc, targetEnc); - - for (size_t i = 0; i < m_numStrings; i++) - { - wxString key(StringAtOfs(m_pOrigTable, i)); - hash[key] = - converter.Convert(wxString(StringAtOfs(m_pTransTable, i))); - } } } +#endif // wxUSE_WCHAR_T/!wxUSE_WCHAR_T + (void)convertEncoding; // get rid of warnings about unused parameter - if ( !convertEncoding ) - #endif // wxUSE_FONTMAP/!wxUSE_FONTMAP + for (size_t32 i = 0; i < m_numStrings; i++) { - for (size_t i = 0; i < m_numStrings; i++) + const char *data = StringAtOfs(m_pOrigTable, i); +#if wxUSE_UNICODE + wxString msgid(data, *inputConv); +#else // ASCII + wxString msgid; + #if wxUSE_WCHAR_T + if ( inputConv && 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 key(StringAtOfs(m_pOrigTable, i)); - hash[key] = StringAtOfs(m_pTransTable, i); - } - } + wxString msgstr; +#if wxUSE_WCHAR_T + #if wxUSE_UNICODE + msgstr = wxString(data + offset, *inputConv); + #else + if ( inputConv ) + 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 -} - -wxString wxMsgCatalogFile::GetCharset() const -{ - // 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 correct header - return wxEmptyString; - } - - wxString header = wxString::FromAscii( StringAtOfs(m_pTransTable, 0)); - wxString charset; - int pos = header.Find(wxT("Content-Type: text/plain; charset=")); - if ( pos == wxNOT_FOUND ) - { - // incorrectly filled Content-Type header - return wxEmptyString; - } - size_t n = pos + 34; /*strlen("Content-Type: text/plain; charset=")*/ - while ( header[n] != wxT('\n') ) - charset << header[n++]; - - if ( charset == wxT("CHARSET") ) - { - // "CHARSET" is not valid charset, but lazy translator - return wxEmptyString; + if ( !msgstr.empty() ) + { + hash[index == 0 ? msgid : msgid + wxChar(index)] = msgstr; + } + offset += strlen(data + offset) + 1; + ++index; + } } - return charset; +#if wxUSE_WCHAR_T + delete sourceConv; + delete csConv; +#endif } + // ---------------------------------------------------------------------------- // wxMsgCatalog class // ---------------------------------------------------------------------------- bool wxMsgCatalog::Load(const wxChar *szDirPrefix, const wxChar *szName, - bool bConvertEncoding) + const wxChar *msgIdCharset, bool bConvertEncoding) { wxMsgCatalogFile file; m_name = szName; - if ( file.Load(szDirPrefix, szName) ) + if ( file.Load(szDirPrefix, szName, m_pluralFormsCalculator) ) { - file.FillHash(m_messages, bConvertEncoding); - return TRUE; + file.FillHash(m_messages, msgIdCharset, bConvertEncoding); + return true; } - return FALSE; + return false; } -const wxChar *wxMsgCatalog::GetString(const wxChar *sz) const +const wxChar *wxMsgCatalog::GetString(const wxChar *sz, size_t n) const { - wxMessagesHash::const_iterator i = m_messages.find(sz); + int index = 0; + if (n != size_t(-1)) + { + index = m_pluralFormsCalculator->evaluate(n); + } + wxMessagesHash::const_iterator i; + if (index != 0) + { + i = m_messages.find(wxString(sz) + wxChar(index)); // plural + } + else + { + i = m_messages.find(sz); + } + if ( i != m_messages.end() ) { return i->second.c_str(); @@ -578,7 +1424,7 @@ const wxChar *wxMsgCatalog::GetString(const wxChar *sz) const #include "wx/arrimpl.cpp" WX_DECLARE_EXPORTED_OBJARRAY(wxLanguageInfo, wxLanguageInfoArray); -WX_DEFINE_OBJARRAY(wxLanguageInfoArray); +WX_DEFINE_OBJARRAY(wxLanguageInfoArray) wxLanguageInfoArray *wxLocale::ms_languagesDB = NULL; @@ -598,11 +1444,15 @@ wxLanguageInfoArray *wxLocale::ms_languagesDB = NULL; } -wxLocale::wxLocale() +void wxLocale::DoCommonInit() { m_pszOldLocale = NULL; + + m_pOldLocale = wxSetLocale(this); + m_pMsgCat = NULL; m_language = wxLANGUAGE_UNKNOWN; + m_initialized = false; } // NB: this function has (desired) side effect of changing current locale @@ -612,6 +1462,10 @@ bool wxLocale::Init(const wxChar *szName, 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; @@ -623,19 +1477,40 @@ bool wxLocale::Init(const wxChar *szName, // the argument to setlocale() szLocale = szShort; - wxCHECK_MSG( szLocale, FALSE, _T("no locale to set in wxLocale::Init()") ); + 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); } + 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 - m_pszOldLocale = wxStrdup(wxSetlocale(LC_ALL, szLocale)); if ( m_pszOldLocale == NULL ) wxLogError(_("locale '%s' can not be set."), szLocale); // the short name will be used to look for catalog files as well, // so we need something here - if ( m_strShort.IsEmpty() ) { + if ( m_strShort.empty() ) { // FIXME I don't know how these 2 letter abbreviations are formed, // this wild guess is surely wrong - if ( szLocale[0] ) + if ( szLocale && szLocale[0] ) { m_strShort += (wxChar)wxTolower(szLocale[0]); if ( szLocale[1] ) @@ -643,28 +1518,52 @@ bool wxLocale::Init(const wxChar *szName, } } - // 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")); + // there may be a catalog with toolkit specific overrides, it is not + // an error if this does not exist + if ( bOk && wxTheApp ) + { + wxAppTraits *traits = wxTheApp->GetTraits(); + if (traits) + AddCatalog(traits->GetToolkitInfo().name.BeforeFirst(wxT('/')).MakeLower()); + } + } + return bOk; } -#if defined(__UNIX__) && wxUSE_UNICODE +#if defined(__UNIX__) && wxUSE_UNICODE && !defined(__WXMAC__) static wxWCharBuffer wxSetlocaleTryUTF(int c, const wxChar *lc) { wxMB2WXbuf l = wxSetlocale(c, lc); - if ( lc && lc[0] != 0 && !l ) + if ( !l && lc && lc[0] != 0 ) { - wxString buf(lc); - buf += wxT(".utf8"); - l = wxSetlocale(c, buf.c_str()); + 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; } @@ -684,7 +1583,7 @@ bool wxLocale::Init(int language, int flags) // We failed to detect system language, so we will use English: if (lang == wxLANGUAGE_UNKNOWN) { - return FALSE; + return false; } const wxLanguageInfo *info = GetLanguageInfo(lang); @@ -693,7 +1592,7 @@ bool wxLocale::Init(int language, int flags) if (info == NULL) { wxLogError(wxT("Unknown language %i."), lang); - return FALSE; + return false; } wxString name = info->Description; @@ -701,48 +1600,83 @@ bool wxLocale::Init(int language, int flags) wxString locale; // Set the locale: -#if defined(__UNIX__) && !defined(__WXMAC__) - if (language == wxLANGUAGE_DEFAULT) - locale = wxEmptyString; - else +#if defined(__OS2__) + wxMB2WXbuf retloc = wxSetlocale(LC_ALL , wxEmptyString); +#elif defined(__UNIX__) && !defined(__WXMAC__) + if (language != wxLANGUAGE_DEFAULT) locale = info->CanonicalName; wxMB2WXbuf retloc = wxSetlocaleTryUTF(LC_ALL, locale); + const wxString langOnly = locale.Left(2); if ( !retloc ) { // Some C libraries don't like xx_YY form and require xx only - retloc = wxSetlocaleTryUTF(LC_ALL, locale.Mid(0,2)); + retloc = wxSetlocaleTryUTF(LC_ALL, langOnly); } + +#if wxUSE_FONTMAP + // some systems (e.g. FreeBSD and HP-UX) don't have xx_YY aliases but + // require the full xx_YY.encoding form, so try using UTF-8 because this is + // the only thing we can do generically + // + // TODO: add encodings applicable to each language to the lang DB and try + // them all in turn here 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); - else if (mid == wxT("nb")) - locale = wxT("no_NO"); - else if (mid == wxT("nn")) - locale = wxT("no_NY"); - - retloc = wxSetlocaleTryUTF(LC_ALL, locale); + const wxChar **names = + wxFontMapperBase::GetAllEncodingNames(wxFONTENCODING_UTF8); + while ( *names ) + { + retloc = wxSetlocale(LC_ALL, locale + _T('.') + *names++); + if ( retloc ) + break; + } } +#endif // wxUSE_FONTMAP + 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 = wxSetlocaleTryUTF(LC_ALL, locale.Mid(0,2)); + // Some C libraries (namely glibc) still use old ISO 639, + // so will translate the abbrev for them + wxString localeAlt; + if ( langOnly == wxT("he") ) + localeAlt = wxT("iw") + locale.Mid(3); + else if ( langOnly == wxT("id") ) + localeAlt = wxT("in") + locale.Mid(3); + else if ( langOnly == wxT("yi") ) + localeAlt = wxT("ji") + locale.Mid(3); + else if ( langOnly == wxT("nb") ) + localeAlt = wxT("no_NO"); + else if ( langOnly == wxT("nn") ) + localeAlt = wxT("no_NY"); + + if ( !localeAlt.empty() ) + { + retloc = wxSetlocaleTryUTF(LC_ALL, localeAlt); + if ( !retloc ) + retloc = wxSetlocaleTryUTF(LC_ALL, locale.Left(2)); + } } + if ( !retloc ) { wxLogError(wxT("Cannot set locale to '%s'."), locale.c_str()); - return FALSE; + return false; } + +#ifdef __AIX__ + // at least in AIX 5.2 libc is buggy and the string returned from setlocale(LC_ALL) + // can't be passed back to it because it returns 6 strings (one for each locale + // category), i.e. for C locale we get back "C C C C C C" + // + // this contradicts IBM own docs but this is not of much help, so just work around + // it in the crudest possible manner + wxChar *p = wxStrchr((wxChar *)retloc, _T(' ')); + if ( p ) + *p = _T('\0'); +#endif // __AIX__ + #elif defined(__WIN32__) #if wxUSE_UNICODE && (defined(__VISUALC__) || defined(__MINGW32__)) @@ -754,7 +1688,10 @@ bool wxLocale::Init(int language, int flags) // #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) { @@ -765,10 +1702,17 @@ bool wxLocale::Init(int language, int flags) } else { - int codepage = -1; + int codepage + #ifdef SETLOCALE_FAILS_ON_UNICODE_LANGS + = -1 + #endif + ; wxUint32 lcid = MAKELCID(MAKELANGID(info->WinLang, info->WinSublang), SORT_DEFAULT); + // 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: @@ -784,15 +1728,18 @@ bool wxLocale::Init(int language, int flags) if (codepage != 0) locale << wxT(".") << buffer; } - if (locale.IsEmpty()) + if (locale.empty()) { wxLogLastError(wxT("SetThreadLocale")); wxLogError(wxT("Cannot set locale to language %s."), name.c_str()); - return FALSE; + return false; } else { + // FIXME +#ifndef __WXWINCE__ retloc = wxSetlocale(LC_ALL, locale); +#endif #ifdef SETLOCALE_FAILS_ON_UNICODE_LANGS if (codepage == 0 && (const wxChar*)retloc == NULL) { @@ -804,7 +1751,12 @@ bool wxLocale::Init(int language, int flags) } else { + // FIXME +#ifndef __WXWINCE__ retloc = wxSetlocale(LC_ALL, wxEmptyString); +#else + retloc = NULL; +#endif #ifdef SETLOCALE_FAILS_ON_UNICODE_LANGS if ((const wxChar*)retloc == NULL) { @@ -822,12 +1774,28 @@ bool wxLocale::Init(int language, int flags) if ( !retloc ) { wxLogError(wxT("Cannot set locale to language %s."), name.c_str()); - return FALSE; + return false; + } +#elif defined(__WXMAC__) + if (lang == wxLANGUAGE_DEFAULT) + locale = wxEmptyString; + else + locale = info->CanonicalName; + + wxMB2WXbuf retloc = wxSetlocale(LC_ALL, locale); + + if ( !retloc ) + { + // Some C libraries don't like xx_YY form and require xx only + retloc = wxSetlocale(LC_ALL, locale.Mid(0,2)); + } + if ( !retloc ) + { + wxLogError(wxT("Cannot set locale to '%s'."), locale.c_str()); + return false; } -#elif defined(__WXMAC__) || defined(__WXPM__) - wxMB2WXbuf retloc = wxSetlocale(LC_ALL , wxEmptyString); #else - return FALSE; + return false; #define WX_NO_LOCALE_SUPPORT #endif @@ -838,7 +1806,7 @@ bool wxLocale::Init(int language, int flags) (flags & wxLOCALE_CONV_ENCODING) != 0); free(szLocale); - if ( ret ) + if (IsOk()) // setlocale() succeeded m_language = lang; return ret; @@ -919,7 +1887,7 @@ void wxLocale::AddCatalogLookupPathPrefix(const wxString& prefix) (langFull.Len() == LEN_FULL && langFull[LEN_LANG] == wxT('_')) ) { // 0. Make sure the lang is according to latest ISO 639 - // (this is neccessary because glibc uses iw and in instead + // (this is necessary because glibc uses iw and in instead // of he and id respectively). // the language itself (second part is the dialect/sublang) @@ -1313,7 +2281,7 @@ void wxLocale::AddCatalogLookupPathPrefix(const wxString& prefix) // 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 he encoding info +// reflects the way different systems provide the encoding info /* static */ wxString wxLocale::GetSystemEncodingName() @@ -1324,6 +2292,9 @@ wxString wxLocale::GetSystemEncodingName() // FIXME: what is the error return value for GetACP()? UINT codepage = ::GetACP(); encname.Printf(_T("windows-%u"), codepage); +#elif defined(__WXMAC__) + // default is just empty string, this resolves to the default system + // encoding later #elif defined(__UNIX_LIKE__) #if defined(HAVE_LANGINFO_H) && defined(CODESET) @@ -1331,31 +2302,15 @@ wxString wxLocale::GetSystemEncodingName() // to Unix98) char *oldLocale = strdup(setlocale(LC_CTYPE, NULL)); setlocale(LC_CTYPE, ""); - char *alang = nl_langinfo(CODESET); + 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... - if ( strcmp(alang, "646") == 0 || - strcmp(alang, "ANSI_X3.4-1968") == 0 ) - { - encname = _T("US-ASCII"); - } - else - { - encname = wxString::FromAscii( alang ); - } + encname = wxString::FromAscii( alang ); } - else + else // nl_langinfo() failed #endif // HAVE_LANGINFO_H { // if we can't get at the character set directly, try to see if it's in @@ -1392,12 +2347,17 @@ wxFontEncoding wxLocale::GetSystemEncoding() #if defined(__WIN32__) && !defined(__WXMICROWIN__) UINT codepage = ::GetACP(); - // wxWindows only knows about CP1250-1257, 932, 936, 949, 950 + // wxWidgets only knows about CP1250-1257, 874, 932, 936, 949, 950 if ( codepage >= 1250 && codepage <= 1257 ) { return (wxFontEncoding)(wxFONTENCODING_CP1250 + codepage - 1250); } + if ( codepage == 874 ) + { + return wxFONTENCODING_CP874; + } + if ( codepage == 932 ) { return wxFONTENCODING_CP932; @@ -1417,17 +2377,24 @@ wxFontEncoding wxLocale::GetSystemEncoding() { 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(); + const wxString encname = GetSystemEncodingName(); if ( !encname.empty() ) { - wxFontEncoding enc = wxFontMapper::Get()-> - CharsetToEncoding(encname, FALSE /* not interactive */); + wxFontEncoding enc = wxFontMapperBase::GetEncodingFromName(encname); // 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 !wxUSE_UNICODE && defined(__WXGTK__) if ( enc == wxFONTENCODING_UTF8 ) { // the most similar supported encoding... @@ -1435,13 +2402,11 @@ wxFontEncoding wxLocale::GetSystemEncoding() } #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 ) + // GetEncodingFromName() returns wxFONTENCODING_DEFAULT for C locale + // (a.k.a. US-ASCII) which is arguably a bug but keep it like this for + // backwards compatibility and just take care to not return + // wxFONTENCODING_DEFAULT from here as this surely doesn't make sense + if ( enc != wxFONTENCODING_MAX && enc != wxFONTENCODING_DEFAULT ) { return enc; } @@ -1464,6 +2429,11 @@ 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++ ) { @@ -1524,7 +2494,12 @@ const wxLanguageInfo *wxLocale::FindLanguageInfo(const wxString& locale) wxString wxLocale::GetSysName() const { + // FIXME +#ifndef __WXWINCE__ return wxSetlocale(LC_ALL, NULL); +#else + return wxEmptyString; +#endif } // clean up @@ -1538,18 +2513,30 @@ wxLocale::~wxLocale() delete pTmpCat; } - // restore old locale + // restore old locale pointer 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 wxChar *wxLocale::GetString(const wxChar *szOrigString, const wxChar *szDomain) const +{ + return GetString(szOrigString, szOrigString, size_t(-1), szDomain); +} + +const wxChar *wxLocale::GetString(const wxChar *szOrigString, + const wxChar *szOrigString2, + size_t n, + const wxChar *szDomain) const { if ( wxIsEmpty(szOrigString) ) - return _T(""); + return wxEmptyString; const wxChar *pszTrans = NULL; wxMsgCatalog *pMsgCat; @@ -1560,14 +2547,14 @@ const wxChar *wxLocale::GetString(const wxChar *szOrigString, // does the catalog exist? if ( pMsgCat != NULL ) - pszTrans = pMsgCat->GetString(szOrigString); + pszTrans = pMsgCat->GetString(szOrigString, n); } else { // search in all domains for ( pMsgCat = m_pMsgCat; pMsgCat != NULL; pMsgCat = pMsgCat->m_pNext ) { - pszTrans = pMsgCat->GetString(szOrigString); + pszTrans = pMsgCat->GetString(szOrigString, n); if ( pszTrans != NULL ) // take the first found break; } @@ -1580,27 +2567,76 @@ const wxChar *wxLocale::GetString(const wxChar *szOrigString, { NoTransErr noTransErr; - if ( szDomain != NULL ) - { - wxLogTrace(_T("i18n"), - _T("string '%s' not found in domain '%s' for locale '%s'."), - szOrigString, szDomain, m_strLocale.c_str()); - } - else - { - wxLogTrace(_T("i18n"), - _T("string '%s' not found in locale '%s'."), - szOrigString, m_strLocale.c_str()); - } + wxLogTrace(TRACE_I18N, + _T("string \"%s\"[%ld] not found in %slocale '%s'."), + szOrigString, (long)n, + szDomain ? wxString::Format(_T("domain '%s' "), szDomain).c_str() + : _T(""), + m_strLocale.c_str()); } #endif // __WXDEBUG__ - return szOrigString; + if (n == size_t(-1)) + return szOrigString; + else + return n == 1 ? szOrigString : szOrigString2; } 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(wxEmptyString, (size_t)-1); + } + else + { + // search in all domains + for ( pMsgCat = m_pMsgCat; pMsgCat != NULL; pMsgCat = pMsgCat->m_pNext ) + { + pszTrans = pMsgCat->GetString(wxEmptyString, (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 { @@ -1623,22 +2659,47 @@ bool wxLocale::IsLoaded(const wxChar *szDomain) const // 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; } } @@ -1646,12 +2707,10 @@ bool wxLocale::AddCatalog(const wxChar *szDomain) // accessors for locale-dependent data // ---------------------------------------------------------------------------- -#if 0 - #ifdef __WXMSW__ /* static */ -wxString wxLocale::GetInfo(wxLocaleInfo index) +wxString wxLocale::GetInfo(wxLocaleInfo index, wxLocaleCategory WXUNUSED(cat)) { wxString str; wxChar buffer[256]; @@ -1659,29 +2718,31 @@ wxString wxLocale::GetInfo(wxLocaleInfo index) buffer[0] = wxT('\0'); switch (index) { - case wxSYS_DECIMAL_SEPARATOR: + case wxLOCALE_DECIMAL_POINT: count = ::GetLocaleInfo(LOCALE_USER_DEFAULT, LOCALE_SDECIMAL, buffer, 256); if (!count) - str << "."; + str << wxT("."); else str << buffer; break; +#if 0 case wxSYS_LIST_SEPARATOR: count = ::GetLocaleInfo(LOCALE_USER_DEFAULT, LOCALE_SLIST, buffer, 256); if (!count) - str << ","; + 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 << "0"; + str << wxT("0"); else str << buffer; break; +#endif default: - wxFAIL_MSG("Unknown System String !"); + wxFAIL_MSG(wxT("Unknown System String !")); } return str; } @@ -1689,15 +2750,42 @@ wxString wxLocale::GetInfo(wxLocaleInfo index) #else // !__WXMSW__ /* static */ -wxString wxLocale::GetInfo(wxLocaleInfo index, wxLocaleCategory) +wxString wxLocale::GetInfo(wxLocaleInfo index, wxLocaleCategory cat) { - return wxEmptyString; + 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__ -#endif // 0 - // ---------------------------------------------------------------------------- // global functions and variables // ---------------------------------------------------------------------------- @@ -1731,7 +2819,7 @@ class wxLocaleModule: public wxModule DECLARE_DYNAMIC_CLASS(wxLocaleModule) public: wxLocaleModule() {} - bool OnInit() { return TRUE; } + bool OnInit() { return true; } void OnExit() { wxLocale::DestroyLanguagesDB(); } }; @@ -2293,7 +3381,7 @@ void wxLocale::InitLanguagesDB() 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_TW", LANG_CHINESE , SUBLANG_CHINESE_TRADITIONAL , "Chinese (Traditional)") LNG(wxLANGUAGE_CHINESE_HONGKONG, "zh_HK", LANG_CHINESE , SUBLANG_CHINESE_HONGKONG , "Chinese (Hongkong)") @@ -2450,7 +3538,7 @@ void wxLocale::InitLanguagesDB() 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") @@ -2481,10 +3569,9 @@ void wxLocale::InitLanguagesDB() LNG(wxLANGUAGE_ZHUANG, "za" , 0 , 0 , "Zhuang") LNG(wxLANGUAGE_ZULU, "zu" , 0 , 0 , "Zulu") -}; +} #undef LNG // --- --- --- generated code ends here --- --- --- #endif // wxUSE_INTL -