Merge pull request #640 from htacg/locale_fix

Address #639.

console/tidy.c

@@ -2003,7 +2003,6 @@ int main( int argc, char** argv )
     ctmbstr cfgfil = NULL, errfil = NULL, htmlfil = NULL;
     TidyDoc tdoc = NULL;
     int status = 0;
-    tmbstr locale = NULL;
 
     uint contentErrors = 0;
     uint contentWarnings = 0;
@@ -2023,14 +2022,6 @@ int main( int argc, char** argv )
     /* Set an atexit handler. */
     atexit( tidy_cleanup );
 
-    /*************************************/
-    /* Set the locale for tidy's output. */
-    /*************************************/
-    locale = tidySystemLocale(locale);
-    tidySetLanguage(locale);
-    if ( locale )
-        free( locale );
-
 #if defined(_WIN32)
     /* Force Windows console to use UTF, otherwise many characters will
      * be garbage. Note that East Asian languages *are* supported, but

include/tidy.h

@@ -336,9 +336,10 @@ TIDY_EXPORT Bool TIDY_CALL tidySetPanicCall( TidyPanic fpanic );
  */
     
 /** The primary creation of a document instance. Instances of a TidyDoc are used
- ** throughout the API as a token to represent a particular document. When done
- ** using a TidyDoc instance, be sure to `tidyRelease(myTidyDoc);` in order
- ** to free related memory.
+ ** throughout the API as a token to represent a particular document. You must
+ ** create at least one TidyDoc instance to initialize the library and begin
+ ** interaction with the API. When done using a TidyDoc instance, be sure to
+ ** `tidyRelease(myTidyDoc);` in order to free related memory.
  ** @result Returns a TidyDoc instance.
  */
 TIDY_EXPORT TidyDoc TIDY_CALL     tidyCreate(void);
@@ -1969,12 +1970,6 @@ TIDY_EXPORT uint TIDY_CALL getNextErrorCode( TidyIterator* iter );
  ** @{
  */
 
-/** Determines the current locale without affecting the C locale.
- ** @param  result The buffer to use to return the string, or NULL on failure.
- ** @return The same buffer for convenience.
- */
-TIDY_EXPORT tmbstr TIDY_CALL tidySystemLocale(tmbstr result);
-
 /** Tells Tidy to use a different language for output.
  ** @param  languageCode A Windows or POSIX language code, and must match
  **         a TIDY_LANGUAGE for an installed language.

src/language.c

@@ -24,6 +24,7 @@
  *  This structure type provides universal access to all of Tidy's strings.
  */
 typedef struct {
+    Bool manually_set;
     languageDefinition *currentLanguage;
     languageDefinition *fallbackLanguage;
     languageDefinition *languages[];
@@ -35,6 +36,7 @@ typedef struct {
  *  `.currentLanguage` to language_en, which is Tidy's default language.
  */
 static tidyLanguagesType tidyLanguages = {
+    no,           /* library language was NOT manually set */
     &language_en, /* current language */
     &language_en, /* first fallback language */
     {
@@ -295,38 +297,6 @@ ctmbstr TY_(tidyLocalizedString)( uint messageType )
 }
 
 
-/**
- **  Determines the current locale without affecting the C locale.
- **  Tidy has always used the default C locale, and at this point
- **  in its development we're not going to tamper with that.
- **  @note this routine uses default allocator, see tidySetMallocCall.
- **  @param  result The buffer to use to return the string.
- **          Returns NULL on failure.
- **  @return The same buffer for convenience.
- */
-tmbstr TY_(tidySystemLocale)(tmbstr result)
-{
-    ctmbstr temp;
-    TidyAllocator* allocator = &TY_(g_default_allocator);
-    
-    /* This should set the OS locale. */
-    setlocale( LC_ALL, "" );
-    
-    /* This should read the current locale. */
-    temp = setlocale( LC_ALL, NULL);
-    
-    /* Make a new copy of the string, because temp
-     always points to the current locale. */
-    if (( result = TidyAlloc( allocator, strlen( temp ) + 1 ) ))
-        strcpy(result, temp);
-    
-    /* This should restore the C locale. */
-    setlocale( LC_ALL, "C" );
-    
-    return result;
-}
-
-
 /**
  *  Retrieves the POSIX name for a string. Result is a static char so please
  *  don't try to free it. If the name looks like a cc_ll identifier, we will
@@ -483,6 +453,26 @@ ctmbstr TY_(tidyGetLanguage)()
 }
 
 
+/**
+ *  Indicates whether or not the current language was set by a
+ *  LibTidy user (yes) or internally by the library (no).
+ */
+Bool TY_(tidyGetLanguageSetByUser)()
+{
+    return tidyLanguages.manually_set;
+}
+
+
+/**
+ *  Specifies to LibTidy that the user (rather than the library)
+ *  selected the current language.
+ */
+void TY_(tidySetLanguageSetByUser)( void )
+{
+    tidyLanguages.manually_set = yes;
+}
+
+
 /**
  *  Provides a string given `messageType` in the default
  *  localization (which is `en`), for single plural form.

src/language.h

@@ -115,6 +115,26 @@ Bool TY_(tidySetLanguage)( ctmbstr languageCode );
  */
 ctmbstr TY_(tidyGetLanguage)(void);
 
+
+/**
+ *  Indicates whether or not the current language was set by a
+ *  LibTidy user or internally by the library. This flag prevents
+ *  subsequently created instances of TidyDocument from changing the
+ *  user's language.
+ *  @returns Returns yes to indicate that the current language was
+ *    specified by an API user.
+ */
+Bool TY_(tidyGetLanguageSetByUser)(void);
+
+
+/**
+ *  Specifies to LibTidy that the user (rather than the library)
+ *  selected the current language. This flag prevents subsequently
+ *  created instances of TidyDocument from changing the user's language.
+ */
+void TY_(tidySetLanguageSetByUser)( void );
+
+
 /**
  *  Provides a string given `messageType` in the current
  *  localization for `quantity`.

src/tidylib.c

@@ -34,6 +34,9 @@
 #include "mappedio.h"
 #include "language.h"
 #include "sprtf.h"
+#if SUPPORT_LOCALIZATIONS
+#  include "locale.h"
+#endif
 
 /* Create/Destroy a Tidy "document" object */
 static TidyDocImpl* tidyDocCreate( TidyAllocator *allocator );
@@ -108,6 +111,17 @@ TidyDocImpl* tidyDocCreate( TidyAllocator *allocator )
     TY_(InitConfig)( doc );
     TY_(InitPrintBuf)( doc );
 
+    /* Set the locale for tidy's output. This both configures
+    ** LibTidy to use the environment's locale as well as the
+    ** standard library.
+    */
+#if SUPPORT_LOCALIZATIONS
+    if ( TY_(tidyGetLanguageSetByUser)() == no )
+    {
+        TY_(tidySetLanguage)( setlocale( LC_ALL, "") );
+    }
+#endif
+
     /* By default, wire tidy messages to standard error.
     ** Document input will be set by parsing routines.
     ** Document output will be set by pretty print routines.
@@ -2472,14 +2486,14 @@ uint TIDY_CALL getNextErrorCode( TidyIterator* iter )
  *******************************************************************/
 
 
-tmbstr TIDY_CALL tidySystemLocale(tmbstr result)
-{
-    return TY_(tidySystemLocale)( result );
-}
-
 Bool TIDY_CALL tidySetLanguage( ctmbstr languageCode )
 {
-    return TY_(tidySetLanguage)( languageCode );
+    Bool result = TY_(tidySetLanguage)( languageCode );
+
+    if ( result )
+        TY_(tidySetLanguageSetByUser)();
+
+    return result;
 }
 
 ctmbstr TIDY_CALL tidyGetLanguage()