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: #include "mozEnglishWordUtils.h"
michael@0: #include "nsReadableUtils.h"
michael@0: #include "nsIServiceManager.h"
michael@0: #include "nsUnicharUtils.h"
michael@0: #include "nsUnicharUtilCIID.h"
michael@0: #include "nsUnicodeProperties.h"
michael@0: #include "nsCRT.h"
michael@0: #include "mozilla/Likely.h"
michael@0: 
michael@0: NS_IMPL_CYCLE_COLLECTING_ADDREF(mozEnglishWordUtils)
michael@0: NS_IMPL_CYCLE_COLLECTING_RELEASE(mozEnglishWordUtils)
michael@0: 
michael@0: NS_INTERFACE_MAP_BEGIN(mozEnglishWordUtils)
michael@0:   NS_INTERFACE_MAP_ENTRY(mozISpellI18NUtil)
michael@0:   NS_INTERFACE_MAP_ENTRY_AMBIGUOUS(nsISupports, mozISpellI18NUtil)
michael@0:   NS_INTERFACE_MAP_ENTRIES_CYCLE_COLLECTION(mozEnglishWordUtils)
michael@0: NS_INTERFACE_MAP_END
michael@0: 
michael@0: NS_IMPL_CYCLE_COLLECTION(mozEnglishWordUtils,
michael@0:                          mURLDetector)
michael@0: 
michael@0: mozEnglishWordUtils::mozEnglishWordUtils()
michael@0: {
michael@0:   mLanguage.AssignLiteral("en");
michael@0: 
michael@0:   nsresult rv;
michael@0:   mURLDetector = do_CreateInstance(MOZ_TXTTOHTMLCONV_CONTRACTID, &rv);
michael@0: }
michael@0: 
michael@0: mozEnglishWordUtils::~mozEnglishWordUtils()
michael@0: {
michael@0: }
michael@0: 
michael@0: /* attribute wstring language; */
michael@0: NS_IMETHODIMP mozEnglishWordUtils::GetLanguage(char16_t * *aLanguage)
michael@0: {
michael@0:   nsresult rv = NS_OK;
michael@0:   NS_ENSURE_ARG_POINTER(aLanguage);
michael@0: 
michael@0:   *aLanguage = ToNewUnicode(mLanguage);
michael@0:   if(!aLanguage) rv = NS_ERROR_OUT_OF_MEMORY;
michael@0:   return rv;
michael@0:  }
michael@0: 
michael@0: /* void GetRootForm (in wstring aWord, in uint32_t type, [array, size_is (count)] out wstring words, out uint32_t count); */
michael@0: // return the possible root forms of aWord.
michael@0: NS_IMETHODIMP mozEnglishWordUtils::GetRootForm(const char16_t *aWord, uint32_t type, char16_t ***words, uint32_t *count)
michael@0: {
michael@0:   nsAutoString word(aWord);
michael@0:   char16_t **tmpPtr;
michael@0:   int32_t length = word.Length();
michael@0: 
michael@0:   *count = 0;
michael@0: 
michael@0:   mozEnglishWordUtils::myspCapitalization ct = captype(word);
michael@0:   switch (ct)
michael@0:     {
michael@0:     case HuhCap:
michael@0:     case NoCap: 
michael@0:       tmpPtr = (char16_t **)nsMemory::Alloc(sizeof(char16_t *));
michael@0:       if (!tmpPtr)
michael@0:         return NS_ERROR_OUT_OF_MEMORY;
michael@0:       tmpPtr[0] = ToNewUnicode(word);
michael@0:       if (!tmpPtr[0]) {
michael@0:         NS_FREE_XPCOM_ALLOCATED_POINTER_ARRAY(0, tmpPtr);
michael@0:         return NS_ERROR_OUT_OF_MEMORY;
michael@0:       }
michael@0:       *words = tmpPtr;
michael@0:       *count = 1;
michael@0:       break;
michael@0:     
michael@0: 
michael@0:     case AllCap:
michael@0:       tmpPtr = (char16_t **)nsMemory::Alloc(sizeof(char16_t *) * 3);
michael@0:       if (!tmpPtr)
michael@0:         return NS_ERROR_OUT_OF_MEMORY;
michael@0:       tmpPtr[0] = ToNewUnicode(word);
michael@0:       if (!tmpPtr[0]) {
michael@0:         NS_FREE_XPCOM_ALLOCATED_POINTER_ARRAY(0, tmpPtr);
michael@0:         return NS_ERROR_OUT_OF_MEMORY;
michael@0:       }
michael@0:       ToLowerCase(tmpPtr[0], tmpPtr[0], length);
michael@0: 
michael@0:       tmpPtr[1] = ToNewUnicode(word);
michael@0:       if (!tmpPtr[1]) {
michael@0:         NS_FREE_XPCOM_ALLOCATED_POINTER_ARRAY(1, tmpPtr);
michael@0:         return NS_ERROR_OUT_OF_MEMORY;
michael@0:       }
michael@0:       ToLowerCase(tmpPtr[1], tmpPtr[1], length);
michael@0:       ToUpperCase(tmpPtr[1], tmpPtr[1], 1);
michael@0: 
michael@0:       tmpPtr[2] = ToNewUnicode(word);
michael@0:       if (!tmpPtr[2]) {
michael@0:         NS_FREE_XPCOM_ALLOCATED_POINTER_ARRAY(2, tmpPtr);
michael@0:         return NS_ERROR_OUT_OF_MEMORY;
michael@0:       }
michael@0: 
michael@0:       *words = tmpPtr;
michael@0:       *count = 3;
michael@0:       break;
michael@0:  
michael@0:     case InitCap:  
michael@0:       tmpPtr = (char16_t **)nsMemory::Alloc(sizeof(char16_t *) * 2);
michael@0:       if (!tmpPtr)
michael@0:         return NS_ERROR_OUT_OF_MEMORY;
michael@0: 
michael@0:       tmpPtr[0] = ToNewUnicode(word);
michael@0:       if (!tmpPtr[0]) {
michael@0:         NS_FREE_XPCOM_ALLOCATED_POINTER_ARRAY(0, tmpPtr);
michael@0:         return NS_ERROR_OUT_OF_MEMORY;
michael@0:       }
michael@0:       ToLowerCase(tmpPtr[0], tmpPtr[0], length);
michael@0: 
michael@0:       tmpPtr[1] = ToNewUnicode(word);
michael@0:       if (!tmpPtr[1]) {
michael@0:         NS_FREE_XPCOM_ALLOCATED_POINTER_ARRAY(1, tmpPtr);
michael@0:         return NS_ERROR_OUT_OF_MEMORY;
michael@0:       }
michael@0: 
michael@0:       *words = tmpPtr;
michael@0:       *count = 2;
michael@0:       break;
michael@0:     default:
michael@0:       return NS_ERROR_FAILURE; // should never get here;
michael@0:     }
michael@0:   return NS_OK;
michael@0: }
michael@0: 
michael@0: // This needs vast improvement
michael@0: bool mozEnglishWordUtils::ucIsAlpha(char16_t aChar)
michael@0: {
michael@0:   // XXX we have to fix callers to handle the full Unicode range
michael@0:   return nsIUGenCategory::kLetter == mozilla::unicode::GetGenCategory(aChar);
michael@0: }
michael@0: 
michael@0: /* void FindNextWord (in wstring word, in uint32_t length, in uint32_t offset, out uint32_t begin, out uint32_t end); */
michael@0: NS_IMETHODIMP mozEnglishWordUtils::FindNextWord(const char16_t *word, uint32_t length, uint32_t offset, int32_t *begin, int32_t *end)
michael@0: {
michael@0:   const char16_t *p = word + offset;
michael@0:   const char16_t *endbuf = word + length;
michael@0:   const char16_t *startWord=p;
michael@0:   if(p<endbuf){
michael@0:     // XXX These loops should be modified to handle non-BMP characters.
michael@0:     // if previous character is a word character, need to advance out of the word
michael@0:     if (offset > 0 && ucIsAlpha(*(p-1))) {
michael@0:       while (p < endbuf && ucIsAlpha(*p))
michael@0:         p++;
michael@0:     }
michael@0:     while((p < endbuf) && (!ucIsAlpha(*p)))
michael@0:       {
michael@0:         p++;
michael@0:       }
michael@0:     startWord=p;
michael@0:     while((p < endbuf) && ((ucIsAlpha(*p))||(*p=='\'')))
michael@0:       { 
michael@0:         p++;
michael@0:       }
michael@0:     
michael@0:     // we could be trying to break down a url, we don't want to break a url into parts,
michael@0:     // instead we want to find out if it really is a url and if so, skip it, advancing startWord 
michael@0:     // to a point after the url.
michael@0: 
michael@0:     // before we spend more time looking to see if the word is a url, look for a url identifer
michael@0:     // and make sure that identifer isn't the last character in the word fragment.
michael@0:     if ( (*p == ':' || *p == '@' || *p == '.') &&  p < endbuf - 1) {
michael@0: 
michael@0:         // ok, we have a possible url...do more research to find out if we really have one
michael@0:         // and determine the length of the url so we can skip over it.
michael@0:        
michael@0:         if (mURLDetector)
michael@0:         {
michael@0:           int32_t startPos = -1;
michael@0:           int32_t endPos = -1;        
michael@0: 
michael@0:           mURLDetector->FindURLInPlaintext(startWord, endbuf - startWord, p - startWord, &startPos, &endPos);
michael@0: 
michael@0:           // ok, if we got a url, adjust the array bounds, skip the current url text and find the next word again
michael@0:           if (startPos != -1 && endPos != -1) { 
michael@0:             startWord = p + endPos + 1; // skip over the url
michael@0:             p = startWord; // reset p
michael@0: 
michael@0:             // now recursively call FindNextWord to search for the next word now that we have skipped the url
michael@0:             return FindNextWord(word, length, startWord - word, begin, end);
michael@0:           }
michael@0:         }
michael@0:     }
michael@0: 
michael@0:     while((p > startWord)&&(*(p-1) == '\'')){  // trim trailing apostrophes
michael@0:       p--;
michael@0:     }
michael@0:   }
michael@0:   else{
michael@0:     startWord = endbuf;
michael@0:   }
michael@0:   if(startWord == endbuf){
michael@0:     *begin = -1;
michael@0:     *end = -1;
michael@0:   }
michael@0:   else{
michael@0:     *begin = startWord-word;
michael@0:     *end = p-word;
michael@0:   }
michael@0:   return NS_OK;
michael@0: }
michael@0: 
michael@0: mozEnglishWordUtils::myspCapitalization 
michael@0: mozEnglishWordUtils::captype(const nsString &word)
michael@0: {
michael@0:   char16_t* lword=ToNewUnicode(word);  
michael@0:   ToUpperCase(lword,lword,word.Length());
michael@0:   if(word.Equals(lword)){
michael@0:     nsMemory::Free(lword);
michael@0:     return AllCap;
michael@0:   }
michael@0: 
michael@0:   ToLowerCase(lword,lword,word.Length());
michael@0:   if(word.Equals(lword)){
michael@0:     nsMemory::Free(lword);
michael@0:     return NoCap;
michael@0:   }
michael@0:   int32_t length=word.Length();
michael@0:   if(Substring(word,1,length-1).Equals(lword+1)){
michael@0:     nsMemory::Free(lword);
michael@0:     return InitCap;
michael@0:   }
michael@0:   nsMemory::Free(lword);
michael@0:   return HuhCap;
michael@0: }
michael@0: 
michael@0: // Convert the list of words in iwords to the same capitalization aWord and 
michael@0: // return them in owords.
michael@0: NS_IMETHODIMP mozEnglishWordUtils::FromRootForm(const char16_t *aWord, const char16_t **iwords, uint32_t icount, char16_t ***owords, uint32_t *ocount)
michael@0: {
michael@0:   nsAutoString word(aWord);
michael@0:   nsresult rv = NS_OK;
michael@0: 
michael@0:   int32_t length;
michael@0:   char16_t **tmpPtr  = (char16_t **)nsMemory::Alloc(sizeof(char16_t *)*icount);
michael@0:   if (!tmpPtr)
michael@0:     return NS_ERROR_OUT_OF_MEMORY;
michael@0: 
michael@0:   mozEnglishWordUtils::myspCapitalization ct = captype(word);
michael@0:   for(uint32_t i = 0; i < icount; ++i) {
michael@0:     length = NS_strlen(iwords[i]);
michael@0:     tmpPtr[i] = (char16_t *) nsMemory::Alloc(sizeof(char16_t) * (length + 1));
michael@0:     if (MOZ_UNLIKELY(!tmpPtr[i])) {
michael@0:       NS_FREE_XPCOM_ALLOCATED_POINTER_ARRAY(i, tmpPtr);
michael@0:       return NS_ERROR_OUT_OF_MEMORY;
michael@0:     }
michael@0:     memcpy(tmpPtr[i], iwords[i], (length + 1) * sizeof(char16_t));
michael@0: 
michael@0:     nsAutoString capTest(tmpPtr[i]);
michael@0:     mozEnglishWordUtils::myspCapitalization newCt=captype(capTest);
michael@0:     if(newCt == NoCap){
michael@0:       switch(ct) 
michael@0:         {
michael@0:         case HuhCap:
michael@0:         case NoCap:
michael@0:           break;
michael@0:         case AllCap:
michael@0:           ToUpperCase(tmpPtr[i],tmpPtr[i],length);
michael@0:           rv = NS_OK;
michael@0:           break;
michael@0:         case InitCap:  
michael@0:           ToUpperCase(tmpPtr[i],tmpPtr[i],1);
michael@0:           rv = NS_OK;
michael@0:           break;
michael@0:         default:
michael@0:           rv = NS_ERROR_FAILURE; // should never get here;
michael@0:           break;
michael@0: 
michael@0:         }
michael@0:     }
michael@0:   }
michael@0:   if (NS_SUCCEEDED(rv)){
michael@0:     *owords = tmpPtr;
michael@0:     *ocount = icount;
michael@0:   }
michael@0:   return rv;
michael@0: }
michael@0: