X-Git-Url: https://git.saurik.com/wxWidgets.git/blobdiff_plain/63986ba653aa30be8bde24c21feabb8fc6067d03..84c5b38d579e140df28cfbb649587e8862148c89:/src/common/intl.cpp diff --git a/src/common/intl.cpp b/src/common/intl.cpp index 032bfb30c1..711788c1e4 100644 --- a/src/common/intl.cpp +++ b/src/common/intl.cpp @@ -1,5 +1,5 @@ ///////////////////////////////////////////////////////////////////////////// -// Name: intl.cpp +// Name: src/common/intl.cpp // Purpose: Internationalization and localisation for wxWindows // Author: Vadim Zeitlin // Modified by: @@ -31,26 +31,35 @@ #if wxUSE_INTL // standard headers -#include -#include +#include +#include +#include +#ifdef HAVE_LANGINFO_H + #include +#endif // wxWindows -#include "wx/defs.h" -#include "wx/string.h" -#include "wx/tokenzr.h" -#include "wx/intl.h" +#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 + #include "wx/file.h" -#include "wx/log.h" -#include "wx/debug.h" -#include "wx/utils.h" -#include "wx/dynarray.h" +#include "wx/tokenzr.h" +#include "wx/module.h" +#include "wx/fontmap.h" +#include "wx/encconv.h" + #ifdef __WIN32__ -#include "wx/msw/private.h" + #include "wx/msw/private.h" +#elif defined(__UNIX_LIKE__) + #include "wx/fontmap.h" // for CharsetToEncoding() #endif - -#include - // ---------------------------------------------------------------------------- // simple types // ---------------------------------------------------------------------------- @@ -85,9 +94,7 @@ typedef unsigned char size_t8; { // 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, + wxASSERT_MSG( wxAssertIsEqual(sizeof(int), 4), "size_t32 is incorrectly defined!" ); } } intsizechecker; @@ -105,6 +112,11 @@ 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 '_' + // ---------------------------------------------------------------------------- // global functions // ---------------------------------------------------------------------------- @@ -139,6 +151,23 @@ public: static wxLocale *wxSetLocale(wxLocale *pLocale); +// helper functions of GetSystemLanguage() +#ifdef __UNIX__ + +// get just the language part +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) +{ + return langFull.Mid(LEN_LANG); +} + +#endif // __UNIX__ + // ---------------------------------------------------------------------------- // wxMsgCatalog corresponds to one disk-file message catalog. // @@ -206,7 +235,7 @@ private: // utility functions // calculate the hash value of given string - static inline size_t32 GetHash(const char *sz); + static size_t32 GetHash(const char *sz); // big<->little endian inline size_t32 Swap(size_t32 ui) const; @@ -436,19 +465,17 @@ const char *wxMsgCatalog::GetString(const char *szOrig) const size_t32 nIncr = 1 + (nHashVal % (m_nHashSize - 2)); -#if defined(__VISAGECPP__) -// VA just can't stand while(1) or while(TRUE) - bool bOs2var = TRUE; - while(bOs2var) { -#else - while (1) { -#endif + for ( ;; ) { size_t32 nStr = Swap(m_pHashTable[nIndex]); if ( nStr == 0 ) return NULL; - if ( strcmp(szOrig, StringAtOfs(m_pOrigTable, nStr - 1)) == 0 ) - return StringAtOfs(m_pTransTable, nStr - 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; + } if ( nIndex >= m_nHashSize - nIncr) nIndex -= m_nHashSize - nIncr; @@ -467,8 +494,11 @@ const char *wxMsgCatalog::GetString(const char *szOrig) const top = current; else if ( res > 0 ) bottom = current + 1; - else // found! - return StringAtOfs(m_pTransTable, current); + else { // found! + // work around the same BC++ 5.5 bug as above + const char *tmp = StringAtOfs(m_pTransTable, current); + return tmp; + } } } @@ -476,24 +506,15 @@ const char *wxMsgCatalog::GetString(const char *szOrig) const return NULL; } - -#if wxUSE_GUI -#include "wx/fontmap.h" -#include "wx/encconv.h" -#endif - void wxMsgCatalog::ConvertEncoding() { -#if wxUSE_GUI - wxFontEncoding enc; - // first, find encoding header: const char *hdr = StringAtOfs(m_pOrigTable, 0); - if (hdr == NULL) return; // not supported by this catalog, does not have non-fuzzy header - if (hdr[0] != 0) return; // ditto + if ( hdr == NULL || hdr[0] != 0 ) { + // not supported by this catalog, does not have correct header + return; + } - /* 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=")); @@ -503,23 +524,28 @@ void wxMsgCatalog::ConvertEncoding() while (header[n] != wxT('\n')) charset << header[n++]; - enc = wxTheFontMapper->CharsetToEncoding(charset, FALSE); +#if wxUSE_FONTMAP + wxFontEncoding enc = wxTheFontMapper->CharsetToEncoding(charset, FALSE); if ( enc == wxFONTENCODING_SYSTEM ) return; // unknown encoding - 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 + wxFontEncoding targetEnc = wxLocale::GetSystemEncoding(); + if (targetEnc == wxFONTENCODING_SYSTEM) + { + 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]; + } wxEncodingConverter converter; + converter.Init(enc, targetEnc); - converter.Init(enc, a[0]); for (size_t i = 0; i < m_numStrings; i++) converter.Convert((char*)StringAtOfs(m_pTransTable, i)); -#endif +#endif // wxUSE_FONTMAP } @@ -531,13 +557,28 @@ void wxMsgCatalog::ConvertEncoding() WX_DECLARE_EXPORTED_OBJARRAY(wxLanguageInfo, wxLanguageInfoArray); WX_DEFINE_OBJARRAY(wxLanguageInfoArray); +wxLanguageInfoArray *wxLocale::ms_languagesDB = NULL; + +/*static*/ void wxLocale::CreateLanguagesDB() +{ + if (ms_languagesDB == NULL) + { + ms_languagesDB = new wxLanguageInfoArray; + InitLanguagesDB(); + } +} + +/*static*/ void wxLocale::DestroyLanguagesDB() +{ + delete ms_languagesDB; + ms_languagesDB = NULL; +} wxLocale::wxLocale() { m_pszOldLocale = NULL; m_pMsgCat = NULL; - m_languagesDB = NULL; m_language = wxLANGUAGE_UNKNOWN; } @@ -583,37 +624,37 @@ bool wxLocale::Init(const wxChar *szName, return bOk; } - - bool wxLocale::Init(int language, int flags) { wxLanguageInfo *info = NULL; int lang = language; - if (m_languagesDB == NULL) + CreateLanguagesDB(); + + if (lang == wxLANGUAGE_DEFAULT) { - m_languagesDB = new wxLanguageInfoArray; - InitLanguagesDB(); + // auto detect the language + lang = GetSystemLanguage(); + } + + // We failed to detect system language, so we will use English: + if (lang == wxLANGUAGE_UNKNOWN) + { + return FALSE; } - if (lang == wxLANGUAGE_DEFAULT) lang = GetSystemLanguage(); - if (lang != wxLANGUAGE_UNKNOWN) + if (lang != wxLANGUAGE_DEFAULT) { - for (size_t i = 0; i < m_languagesDB->GetCount(); i++) + for (size_t i = 0; i < ms_languagesDB->GetCount(); i++) { - if (m_languagesDB->Item(i).Language == lang) + if (ms_languagesDB->Item(i).Language == lang) { - info = &m_languagesDB->Item(i); + info = &ms_languagesDB->Item(i); break; } } } - // We failed to detect system language, so we will use English: - if (lang == wxLANGUAGE_UNKNOWN) - { - return FALSE; - } // Unknown language: if (info == NULL) { @@ -625,11 +666,13 @@ bool wxLocale::Init(int language, int flags) wxString canonical = info->CanonicalName; wxString locale; wxChar *retloc; - + // Set the locale: #ifdef __UNIX__ - if (language == wxLANGUAGE_DEFAULT) locale = wxEmptyString; - else locale = info->CanonicalName; + if (language == wxLANGUAGE_DEFAULT) + locale = wxEmptyString; + else + locale = info->CanonicalName; retloc = wxSetlocale(LC_ALL, locale); @@ -659,7 +702,6 @@ bool wxLocale::Init(int language, int flags) wxLogError(wxT("Cannot set locale to '%s'."), locale.c_str()); return FALSE; } - #elif defined(__WIN32__) if (language != wxLANGUAGE_DEFAULT) { @@ -669,8 +711,8 @@ bool wxLocale::Init(int language, int flags) retloc = wxT("C"); } else - { - wxUint32 lcid = MAKELCID(MAKELANGID(info->WinLang, info->WinSublang), + { + wxUint32 lcid = MAKELCID(MAKELANGID(info->WinLang, info->WinSublang), SORT_DEFAULT); if (SetThreadLocale(lcid)) retloc = wxSetlocale(LC_ALL, wxEmptyString); @@ -707,11 +749,11 @@ bool wxLocale::Init(int language, int flags) wxLogError(wxT("Cannot set locale to language %s."), name.c_str()); return FALSE; } - + #else return FALSE; #endif - + return Init(name, canonical, wxString(retloc), (flags & wxLOCALE_LOAD_DEFAULT) != 0, (flags & wxLOCALE_CONV_ENCODING) != 0); @@ -728,126 +770,248 @@ void wxLocale::AddCatalogLookupPathPrefix(const wxString& prefix) //else: already have it } - -int wxLocale::GetSystemLanguage() const +/*static*/ int wxLocale::GetSystemLanguage() { - int wxlang = wxLANGUAGE_UNKNOWN; - size_t i; + CreateLanguagesDB(); + + // init i to avoid compiler warning + size_t i = 0, + count = ms_languagesDB->GetCount(); - wxASSERT_MSG(m_languagesDB != NULL, "Languages DB not initialized, call wxLocale::Init!"); - #if defined(__UNIX__) - wxString lang; - if (!wxGetEnv(wxT("LC_ALL"), &lang) && - !wxGetEnv(wxT("LC_MESSAGES"), &lang) && - !wxGetEnv(wxT("LANG"), &lang)) - return wxLANGUAGE_UNKNOWN; - - bool is_abbrev = lang.Len() == 2 || - (lang.Len() == 5 && lang[2] == wxT('_')); - - // 0. Make sure the abbrev is according to latest ISO 639 - // (this is neccessary because glibc uses iw and in instead - // of he and id respectively). - if (is_abbrev) + // first get the string identifying the language from the environment + wxString langFull; + if (!wxGetEnv(wxT("LC_ALL"), &langFull) && + !wxGetEnv(wxT("LC_MESSAGES"), &langFull) && + !wxGetEnv(wxT("LANG"), &langFull)) + { + // no language specified, threat it as English + return wxLANGUAGE_ENGLISH; + } + + if ( langFull == _T("C") ) { - wxString mid = lang.Mid(0,2); - if (mid == wxT("iw")) lang = wxT("he") + lang.Mid(3); - else if (mid == wxT("in")) lang = wxT("id") + lang.Mid(3); - else if (mid == wxT("ji")) lang = wxT("yi") + lang.Mid(3); + // default C locale + return wxLANGUAGE_ENGLISH; } - // 1. Try to find the lang as is: - if (is_abbrev) + // the language string has the following form + // + // lang[_LANG[.encoding]] + // + // where lang is the primary language, LANG is a sublang + // + // for example, the following strings are valid: + // fr + // fr_FR + // de_DE.iso88591 + + // 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('.')); + + // 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 + // check for this + + // do we have just the language (or sublang too)? + bool justLang = langFull.Len() == LEN_LANG; + if ( justLang || + (langFull.Len() == LEN_FULL && langFull[LEN_LANG] == wxT('_')) ) { - for (i = 0; i < m_languagesDB->GetCount(); i++) + // 0. Make sure the lang is according to latest ISO 639 + // (this is neccessary because glibc uses iw and in instead + // of he and id respectively). + + // the language itself (second part is the dialect/sublang) + wxString langOrig = ExtractLang(langFull); + + wxString lang; + if ( langOrig == wxT("iw")) + lang = _T("he"); + else if ( langOrig == wxT("in") ) + lang = wxT("id"); + else if ( langOrig == wxT("ji") ) + lang = wxT("yi"); + else + lang = langOrig; + + // did we change it? + if ( lang != langOrig ) + { + langFull = lang + ExtractNotLang(langFull); + } + + // 1. Try to find the language either as is: + for ( i = 0; i < count; i++ ) { - if (m_languagesDB->Item(i).CanonicalName == lang) + if ( ms_languagesDB->Item(i).CanonicalName == langFull ) { - wxlang = m_languagesDB->Item(i).Language; break; } } - } - // 2. If lang is of the form xx_YY, try to find xx: - if (wxlang == wxLANGUAGE_UNKNOWN && is_abbrev && lang.Len() == 5) - { - wxString lang2 = lang.Mid(0,2); - for (i = 0; i < m_languagesDB->GetCount(); i++) + // 2. If langFull is of the form xx_YY, try to find xx: + if ( i == count && !justLang ) { - if (m_languagesDB->Item(i).CanonicalName == lang2) + for ( i = 0; i < count; i++ ) { - wxlang = m_languagesDB->Item(i).Language; - break; + if ( ms_languagesDB->Item(i).CanonicalName == lang ) + { + break; + } } } - } - // 3. If lang is of the form xx, try to find any xx_YY record: - if (wxlang == wxLANGUAGE_UNKNOWN && is_abbrev && lang.Len() == 2) - { - for (i = 0; i < m_languagesDB->GetCount(); i++) + // 3. If langFull is of the form xx, try to find any xx_YY record: + if ( i == count && justLang ) { - if (m_languagesDB->Item(i).CanonicalName.Mid(0,2) == lang) + for ( i = 0; i < count; i++ ) { - wxlang = m_languagesDB->Item(i).Language; - break; + if ( ExtractLang(ms_languagesDB->Item(i).CanonicalName) + == langFull ) + { + break; + } } } } - - // 4. If everything failed, try to find the name in verbose description - // (SuSE is known to use LANG="german"): - if (wxlang == wxLANGUAGE_UNKNOWN && !is_abbrev) + else // not standard format { - for (i = 0; i < m_languagesDB->GetCount(); i++) + // try to find the name in verbose description + for ( i = 0; i < count; i++ ) { - if (m_languagesDB->Item(i).Description.CmpNoCase(lang) == 0) + if (ms_languagesDB->Item(i).Description.CmpNoCase(langFull) == 0) { - wxlang = m_languagesDB->Item(i).Language; break; } } } - #elif defined(__WIN32__) LCID lcid = GetUserDefaultLCID(); - if (lcid == 0) return wxLANGUAGE_UNKNOWN; - wxUint32 lang = PRIMARYLANGID(LANGIDFROMLCID(lcid)); - wxUint32 sublang = SUBLANGID(LANGIDFROMLCID(lcid)); - - for (i = 0; i < m_languagesDB->GetCount(); i++) + if ( lcid != 0 ) { - if (m_languagesDB->Item(i).WinLang == lang && - m_languagesDB->Item(i).WinSublang == sublang) + wxUint32 lang = PRIMARYLANGID(LANGIDFROMLCID(lcid)); + wxUint32 sublang = SUBLANGID(LANGIDFROMLCID(lcid)); + + for ( i = 0; i < count; i++ ) { - wxlang = m_languagesDB->Item(i).Language; - break; + if (ms_languagesDB->Item(i).WinLang == lang && + ms_languagesDB->Item(i).WinSublang == sublang) + { + break; + } } } -#endif + //else: leave wxlang == wxLANGUAGE_UNKNOWN +#endif // Unix/Win32 - return wxlang; + if ( i < count ) + { + // we did find a matching entry, use it + return ms_languagesDB->Item(i).Language; + } + + // no info about this language in the database + return wxLANGUAGE_UNKNOWN; } +// ---------------------------------------------------------------------------- +// 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 he encoding info -void wxLocale::AddLanguage(const wxLanguageInfo& info) +/* static */ +wxString wxLocale::GetSystemEncodingName() { - wxASSERT_MSG(m_languagesDB != NULL, "Languages DB not initialized, call wxLocale::Init!"); - m_languagesDB->Add(info); + wxString encname; + +#ifdef __WIN32__ + // 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, ""); + char *alang = nl_langinfo(CODESET); + setlocale(LC_CTYPE, oldLocale); + free(oldLocale); + if (alang) + { + encname = wxConvLibc.cMB2WX(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) + wxChar *lang = wxGetenv(wxT("LC_ALL")); + wxChar *dot = lang ? wxStrchr(lang, wxT('.')) : (wxChar *)NULL; + if (!dot) + { + lang = wxGetenv(wxT("LC_CTYPE")); + if ( lang ) + dot = wxStrchr(lang, wxT('.')); + } + if (!dot) + { + lang = wxGetenv(wxT("LANG")); + if ( lang ) + dot = wxStrchr(lang, wxT('.')); + } + + if ( dot ) + { + encname = dot+1; + } + } +#endif // Win32/Unix + + return encname; } +/* static */ +wxFontEncoding wxLocale::GetSystemEncoding() +{ +#ifdef __WIN32__ + UINT codepage = ::GetACP(); + + // wxWindows only knows about CP1250-1257 + if ( codepage >= 1250 && codepage <= 1257 ) + { + return (wxFontEncoding)(wxFONTENCODING_CP1250 + codepage - 1250); + } +#elif defined(__UNIX_LIKE__) && wxUSE_FONTMAP + wxString encname = GetSystemEncodingName(); + if ( !encname.empty() ) + { + return wxTheFontMapper-> + CharsetToEncoding(encname, FALSE /* not interactive */); + } +#endif // Win32/Unix + return wxFONTENCODING_SYSTEM; +} + +/*static*/ void wxLocale::AddLanguage(const wxLanguageInfo& info) +{ + CreateLanguagesDB(); + ms_languagesDB->Add(info); +} wxString wxLocale::GetSysName() const { return wxSetlocale(LC_ALL, NULL); } - - // clean up wxLocale::~wxLocale() { @@ -859,8 +1023,6 @@ wxLocale::~wxLocale() delete pTmpCat; } - delete m_languagesDB; - // restore old locale wxSetLocale(m_pOldLocale); wxSetlocale(LC_ALL, m_pszOldLocale); @@ -966,11 +1128,6 @@ bool wxLocale::AddCatalog(const wxChar *szDomain) } } - - - - - // ---------------------------------------------------------------------------- // global functions and variables // ---------------------------------------------------------------------------- @@ -995,16 +1152,35 @@ wxLocale *wxSetLocale(wxLocale *pLocale) +// ---------------------------------------------------------------------------- +// wxLocale module (for lazy destruction of languagesDB) +// ---------------------------------------------------------------------------- + +class wxLocaleModule: public wxModule +{ + DECLARE_DYNAMIC_CLASS(wxLocaleModule) + public: + wxLocaleModule() {} + bool OnInit() { return TRUE; } + void OnExit() { wxLocale::DestroyLanguagesDB(); } +}; + +IMPLEMENT_DYNAMIC_CLASS(wxLocaleModule, wxModule) + + // ---------------------------------------------------------------------------- // default languages table & initialization // ---------------------------------------------------------------------------- + +// --- --- --- generated code begins here --- --- --- + // 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) @@ -1738,6 +1914,7 @@ void wxLocale::InitLanguagesDB() }; #undef LNG +// --- --- --- generated code ends here --- --- --- #endif // wxUSE_INTL