michael@0: /* -*- Mode: C++; tab-width: 2; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
michael@0: /* This Source Code Form is subject to the terms of the Mozilla Public
michael@0:  * License, v. 2.0. If a copy of the MPL was not distributed with this
michael@0:  * file, You can obtain one at http://mozilla.org/MPL/2.0/. */
michael@0: 
michael@0: 
michael@0: #include "nsSampleWordBreaker.h"
michael@0: 
michael@0: nsSampleWordBreaker::nsSampleWordBreaker()
michael@0: {
michael@0: }
michael@0: nsSampleWordBreaker::~nsSampleWordBreaker()
michael@0: {
michael@0: }
michael@0: 
michael@0: NS_IMPL_ISUPPORTS(nsSampleWordBreaker, nsIWordBreaker)
michael@0: 
michael@0: bool nsSampleWordBreaker::BreakInBetween(
michael@0:   const char16_t* aText1 , uint32_t aTextLen1,
michael@0:   const char16_t* aText2 , uint32_t aTextLen2)
michael@0: {
michael@0:   NS_PRECONDITION( nullptr != aText1, "null ptr");
michael@0:   NS_PRECONDITION( nullptr != aText2, "null ptr");
michael@0: 
michael@0:   if(!aText1 || !aText2 || (0 == aTextLen1) || (0 == aTextLen2))
michael@0:     return false;
michael@0: 
michael@0:   return (this->GetClass(aText1[aTextLen1-1]) != this->GetClass(aText2[0]));
michael@0: }
michael@0: 
michael@0: 
michael@0: #define IS_ASCII(c)            (0 == ( 0xFF80 & (c)))
michael@0: #define ASCII_IS_ALPHA(c)         ((( 'a' <= (c)) && ((c) <= 'z')) || (( 'A' <= (c)) && ((c) <= 'Z')))
michael@0: #define ASCII_IS_DIGIT(c)         (( '0' <= (c)) && ((c) <= '9'))
michael@0: #define ASCII_IS_SPACE(c)         (( ' ' == (c)) || ( '\t' == (c)) || ( '\r' == (c)) || ( '\n' == (c)))
michael@0: #define IS_ALPHABETICAL_SCRIPT(c) ((c) < 0x2E80) 
michael@0: 
michael@0: // we change the beginning of IS_HAN from 0x4e00 to 0x3400 to relfect Unicode 3.0 
michael@0: #define IS_HAN(c)              (( 0x3400 <= (c)) && ((c) <= 0x9fff))||(( 0xf900 <= (c)) && ((c) <= 0xfaff))
michael@0: #define IS_KATAKANA(c)         (( 0x30A0 <= (c)) && ((c) <= 0x30FF))
michael@0: #define IS_HIRAGANA(c)         (( 0x3040 <= (c)) && ((c) <= 0x309F))
michael@0: #define IS_HALFWIDTHKATAKANA(c)         (( 0xFF60 <= (c)) && ((c) <= 0xFF9F))
michael@0: #define IS_THAI(c)         (0x0E00 == (0xFF80 & (c) )) // Look at the higest 9 bits
michael@0: 
michael@0: uint8_t nsSampleWordBreaker::GetClass(char16_t c)
michael@0: {
michael@0:   // begin of the hack
michael@0: 
michael@0:   if (IS_ALPHABETICAL_SCRIPT(c))  {
michael@0: 	  if(IS_ASCII(c))  {
michael@0: 		  if(ASCII_IS_SPACE(c)) {
michael@0: 			  return kWbClassSpace;
michael@0: 		  } else if(ASCII_IS_ALPHA(c) || ASCII_IS_DIGIT(c)) {
michael@0: 			  return kWbClassAlphaLetter;
michael@0: 		  } else {
michael@0: 			  return kWbClassPunct;
michael@0: 		  }
michael@0: 	  } else if(IS_THAI(c))	{
michael@0: 		  return kWbClassThaiLetter;
michael@0: 	  } else if (c == 0x00A0/*NBSP*/) {
michael@0:       return kWbClassSpace;
michael@0:     } else {
michael@0: 		  return kWbClassAlphaLetter;
michael@0: 	  }
michael@0:   }  else {
michael@0: 	  if(IS_HAN(c)) {
michael@0: 		  return kWbClassHanLetter;
michael@0: 	  } else if(IS_KATAKANA(c))   {
michael@0: 		  return kWbClassKatakanaLetter;
michael@0: 	  } else if(IS_HIRAGANA(c))   {
michael@0: 		  return kWbClassHiraganaLetter;
michael@0: 	  } else if(IS_HALFWIDTHKATAKANA(c))  {
michael@0: 		  return kWbClassHWKatakanaLetter;
michael@0: 	  } else  {
michael@0: 		  return kWbClassAlphaLetter;
michael@0: 	  }
michael@0:   }
michael@0:   return 0;
michael@0: }
michael@0: 
michael@0: nsWordRange nsSampleWordBreaker::FindWord(
michael@0:   const char16_t* aText , uint32_t aTextLen,
michael@0:   uint32_t aOffset)
michael@0: {
michael@0:   nsWordRange range;
michael@0:   NS_PRECONDITION( nullptr != aText, "null ptr");
michael@0:   NS_PRECONDITION( 0 != aTextLen, "len = 0");
michael@0:   NS_PRECONDITION( aOffset <= aTextLen, "aOffset > aTextLen");
michael@0: 
michael@0:   range.mBegin = aTextLen + 1;
michael@0:   range.mEnd = aTextLen + 1;
michael@0: 
michael@0:   if(!aText || aOffset > aTextLen)
michael@0:     return range;
michael@0: 
michael@0:   uint8_t c = this->GetClass(aText[aOffset]);
michael@0:   uint32_t i;
michael@0:   // Scan forward
michael@0:   range.mEnd--;
michael@0:   for(i = aOffset +1;i <= aTextLen; i++)
michael@0:   {
michael@0:      if( c != this->GetClass(aText[i]))
michael@0:      {
michael@0:        range.mEnd = i;
michael@0:        break;
michael@0:      }
michael@0:   }
michael@0: 
michael@0:   // Scan backward
michael@0:   range.mBegin = 0;
michael@0:   for(i = aOffset ;i > 0; i--)
michael@0:   {
michael@0:      if( c != this->GetClass(aText[i-1]))
michael@0:      {
michael@0:        range.mBegin = i;
michael@0:        break;
michael@0:      }
michael@0:   }
michael@0:   if(kWbClassThaiLetter == c)
michael@0:   {
michael@0: 	// need to call Thai word breaker from here
michael@0: 	// we should pass the whole Thai segment to the thai word breaker to find a shorter answer
michael@0:   }
michael@0:   return range;
michael@0: }
michael@0: 
michael@0: int32_t nsSampleWordBreaker::NextWord( 
michael@0:   const char16_t* aText, uint32_t aLen, uint32_t aPos) 
michael@0: {
michael@0:   int8_t c1, c2;
michael@0:   uint32_t cur = aPos;
michael@0:   if (cur == aLen)
michael@0:     return NS_WORDBREAKER_NEED_MORE_TEXT;
michael@0:   c1 = this->GetClass(aText[cur]);
michael@0:  
michael@0:   for(cur++; cur <aLen; cur++)
michael@0:   {
michael@0:      c2 = this->GetClass(aText[cur]);
michael@0:      if(c2 != c1) 
michael@0:        break;
michael@0:   }
michael@0:   if(kWbClassThaiLetter == c1)
michael@0:   {
michael@0: 	// need to call Thai word breaker from here
michael@0: 	// we should pass the whole Thai segment to the thai word breaker to find a shorter answer
michael@0:   }
michael@0:   if (cur == aLen)
michael@0:     return NS_WORDBREAKER_NEED_MORE_TEXT;
michael@0:   return cur;
michael@0: }