jbe@0: /*
jbe@7:  *  Copyright (c) 2006-2007 Jan Behrens, FlexiGuided GmbH, Berlin
jbe@0:  *
jbe@7:  *  Permission is hereby granted, free of charge, to any person obtaining a
jbe@7:  *  copy of this software and associated documentation files (the "Software"),
jbe@7:  *  to deal in the Software without restriction, including without limitation
jbe@7:  *  the rights to use, copy, modify, merge, publish, distribute, sublicense,
jbe@7:  *  and/or sell copies of the Software, and to permit persons to whom the
jbe@7:  *  Software is furnished to do so, subject to the following conditions:
jbe@0:  *
jbe@7:  *  The above copyright notice and this permission notice shall be included in
jbe@7:  *  all copies or substantial portions of the Software.
jbe@0:  *
jbe@7:  *  THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
jbe@7:  *  IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
jbe@7:  *  FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
jbe@7:  *  AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
jbe@7:  *  LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
jbe@7:  *  FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
jbe@7:  *  DEALINGS IN THE SOFTWARE.
jbe@0:  */
jbe@7: 
jbe@0: 
jbe@0: /*
jbe@0:  *  File name:    ruby/utf8proc_native.c
jbe@7:  *  Version:      1.1.1
jbe@7:  *  Last changed: 2007-07-22
jbe@0:  *
jbe@0:  *  Description:
jbe@0:  *  Native part of the ruby wrapper for libutf8proc.
jbe@0:  */
jbe@0: 
jbe@0: 
jbe@0: #include "../utf8proc.c"
jbe@0: #include "ruby.h"
jbe@0: 
jbe@0: typedef struct utf8proc_ruby_mapenv_struct {
jbe@0:   int32_t *buffer;
jbe@0: } utf8proc_ruby_mapenv_t;
jbe@0: 
jbe@0: void utf8proc_ruby_mapenv_free(utf8proc_ruby_mapenv_t *env) {
jbe@0:   free(env->buffer);
jbe@0:   free(env);
jbe@0: }
jbe@0: 
jbe@0: VALUE utf8proc_ruby_module;
jbe@0: VALUE utf8proc_ruby_options;
jbe@0: VALUE utf8proc_ruby_eUnicodeError;
jbe@0: VALUE utf8proc_ruby_eInvalidUtf8Error;
jbe@0: VALUE utf8proc_ruby_eCodeNotAssignedError;
jbe@0: 
jbe@0: VALUE utf8proc_ruby_map_error(ssize_t result) {
jbe@0:   VALUE excpt_class;
jbe@0:   switch (result) {
jbe@0:     case UTF8PROC_ERROR_NOMEM:
jbe@0:     excpt_class = rb_eNoMemError; break;
jbe@0:     case UTF8PROC_ERROR_OVERFLOW:
jbe@3:     case UTF8PROC_ERROR_INVALIDOPTS:
jbe@0:     excpt_class = rb_eArgError; break;
jbe@0:     case UTF8PROC_ERROR_INVALIDUTF8:
jbe@0:     excpt_class = utf8proc_ruby_eInvalidUtf8Error; break;
jbe@0:     case UTF8PROC_ERROR_NOTASSIGNED:
jbe@0:     excpt_class = utf8proc_ruby_eCodeNotAssignedError; break;
jbe@0:     default:
jbe@0:     excpt_class = rb_eRuntimeError;
jbe@0:   }
jbe@0:   rb_raise(excpt_class, "%s", utf8proc_errmsg(result));
jbe@0:   return Qnil;
jbe@0: }
jbe@0: 
jbe@0: VALUE utf8proc_ruby_map(VALUE self, VALUE str_param, VALUE options_param) {
jbe@0:   VALUE str;
jbe@0:   int options;
jbe@0:   VALUE env_obj;
jbe@0:   utf8proc_ruby_mapenv_t *env;
jbe@0:   ssize_t result;
jbe@0:   VALUE retval;
jbe@0:   str = StringValue(str_param);
jbe@0:   options = NUM2INT(options_param) & ~UTF8PROC_NULLTERM;
jbe@0:   env_obj = Data_Make_Struct(rb_cObject, utf8proc_ruby_mapenv_t, NULL,
jbe@0:     utf8proc_ruby_mapenv_free, env);
jbe@0:   result = utf8proc_decompose(RSTRING(str)->ptr, RSTRING(str)->len,
jbe@0:     NULL, 0, options);
jbe@2:   if (result < 0) {
jbe@2:     utf8proc_ruby_map_error(result);
jbe@2:     return Qnil;  // needed to prevent problems with optimization
jbe@2:   }
jbe@0:   env->buffer = ALLOC_N(int32_t, result+1);
jbe@0:   result = utf8proc_decompose(RSTRING(str)->ptr, RSTRING(str)->len,
jbe@0:     env->buffer, result, options);
jbe@0:   if (result < 0) {
jbe@0:     free(env->buffer);
jbe@0:     env->buffer = 0;
jbe@1:     utf8proc_ruby_map_error(result);
jbe@2:     return Qnil;  // needed to prevent problems with optimization
jbe@0:   }
jbe@0:   result = utf8proc_reencode(env->buffer, result, options);
jbe@0:   if (result < 0) {
jbe@0:     free(env->buffer);
jbe@0:     env->buffer = 0;
jbe@1:     utf8proc_ruby_map_error(result);
jbe@2:     return Qnil;  // needed to prevent problems with optimization
jbe@0:   }
jbe@0:   retval = rb_str_new((char *)env->buffer, result);
jbe@0:   free(env->buffer);
jbe@0:   env->buffer = 0;
jbe@0:   return retval;
jbe@0: }
jbe@0: 
jbe@0: static VALUE utf8proc_ruby_char(VALUE self, VALUE code_param) {
jbe@0:   char buffer[4];
jbe@0:   ssize_t result;
jbe@0:   int uc;
jbe@0:   uc = NUM2INT(code_param);
jbe@7:   if (!utf8proc_codepoint_valid(uc))
jbe@0:     rb_raise(rb_eArgError, "Invalid Unicode code point");
jbe@0:   result = utf8proc_encode_char(uc, buffer);
jbe@0:   return rb_str_new(buffer, result);
jbe@0: }
jbe@0: 
jbe@0: #define register_utf8proc_option(sym, field) \
jbe@0:   rb_hash_aset(utf8proc_ruby_options, ID2SYM(rb_intern(sym)), INT2FIX(field))
jbe@0: 
jbe@0: void Init_utf8proc_native() {
jbe@0:   utf8proc_ruby_module = rb_define_module("Utf8Proc");
jbe@0:   rb_define_module_function(utf8proc_ruby_module, "utf8map",
jbe@0:     utf8proc_ruby_map, 2);
jbe@0:   rb_define_module_function(utf8proc_ruby_module, "utf8char",
jbe@0:     utf8proc_ruby_char, 1);
jbe@0:   utf8proc_ruby_eUnicodeError = rb_define_class_under(utf8proc_ruby_module,
jbe@0:     "UnicodeError", rb_eStandardError);
jbe@0:   utf8proc_ruby_eInvalidUtf8Error = rb_define_class_under(
jbe@0:     utf8proc_ruby_module, "InvalidUtf8Error", utf8proc_ruby_eUnicodeError);
jbe@0:   utf8proc_ruby_eCodeNotAssignedError = rb_define_class_under(
jbe@0:     utf8proc_ruby_module, "CodeNotAssignedError",
jbe@0:     utf8proc_ruby_eUnicodeError);
jbe@0:   utf8proc_ruby_options = rb_hash_new();
jbe@2:   register_utf8proc_option("stable",    UTF8PROC_STABLE);
jbe@2:   register_utf8proc_option("compat",    UTF8PROC_COMPAT);
jbe@2:   register_utf8proc_option("compose",   UTF8PROC_COMPOSE);
jbe@2:   register_utf8proc_option("decompose", UTF8PROC_DECOMPOSE);
jbe@2:   register_utf8proc_option("ignore",    UTF8PROC_IGNORE);
jbe@2:   register_utf8proc_option("rejectna",  UTF8PROC_REJECTNA);
jbe@2:   register_utf8proc_option("nlf2ls",    UTF8PROC_NLF2LS);
jbe@2:   register_utf8proc_option("nlf2ps",    UTF8PROC_NLF2PS);
jbe@2:   register_utf8proc_option("nlf2lf",    UTF8PROC_NLF2LF);
jbe@2:   register_utf8proc_option("stripcc",   UTF8PROC_STRIPCC);
jbe@2:   register_utf8proc_option("casefold",  UTF8PROC_CASEFOLD);
jbe@2:   register_utf8proc_option("charbound", UTF8PROC_CHARBOUND);
jbe@3:   register_utf8proc_option("lump",      UTF8PROC_LUMP);
jbe@3:   register_utf8proc_option("stripmark", UTF8PROC_STRIPMARK);
jbe@0:   OBJ_FREEZE(utf8proc_ruby_options);
jbe@0:   rb_define_const(utf8proc_ruby_module, "Options", utf8proc_ruby_options);
jbe@0: }
jbe@0: