Bug 355178 - Don't break words on hyphens before passing them to hunspell, as it can handle word breaking on hyphens on its own; r=smaug

--HG--
extra : rebase_source : 298f0832ac6f0b12497e9c883bbebea8e168284c

extensions/spellcheck/locales/en-US/hunspell/dictionary-sources/upstream-hunspell.diff

@@ -9463,101 +9463,103 @@
 < schrod's
 ---
 > schrod/SM
-42883,42885c48694
+41998a42010
+> scot-free
+42883,42885c48695
 < shit's
 < shit/S!
 < shite/S!
 ---
 > shit/MS!
-42887,42888c48696,48697
+42887,42888c48697,48698
 < shithead/S!
 < shitload/!
 ---
 > shithead/MS!
 > shitload/MS!
-42891c48700
+42891c48701
 < shitty/RT!
 ---
 > shitty/TR!
-42976a48786
+42976a48787
 > should've
-43008c48818
+43008c48819
 < showtime
 ---
 > showtime/MS
-43724,43726c49534
+43724,43726c49535
 < smoulder's
 < smouldered
 < smoulders
 ---
 > smoulder/GSMD
-44062c49870
+44062c49871
 < sonofabitch
 ---
 > sonofabitch/!
-44371a50180
+44371a50181
 > spick/S!
-44383c50192
+44383c50193
 < spik/S
 ---
 > spik/S!
-46106a51916
+46106a51917
 > syllabi
-46160c51970
+46160c51971
 < synch/GMD
 ---
 > synch/GMDS
-46167d51976
+46167d51977
 < synchs
-46203,46204c52012,52013
+46203,46204c52013,52014
 < sysadmin/S
 < sysop/S
 ---
 > sysadmin/MS
 > sysop/MS
-46752a52562
+46752a52563
 > terabit/MS
-46753a52564,52565
+46753a52565,52566
 > terahertz/M
 > terapixel/MS
-46817a52630
+46817a52631
 > testcase/MS
-46831a52645
+46831a52646
 > testsuite/MS
-46925a52740
+46925a52741
 > theremin/MS
-47755a53571
+47755a53572
 > transfect/DSMG
-47774a53591,53592
+47774a53592,53593
 > transgenderism
 > transgene/MS
-47951c53769
+47951c53770
 < triage/M
 ---
 > triage/MG
-48869a54688
+48869a54689
 > unlikeable
-49211c55030
+49211c55031
 < vagina/M
 ---
 > vagina/MS
-49368,49369c55187
+49368,49369c55188
 < velour's
 < velours's
 ---
 > velour/MS
-49478a55297
+49478a55298
 > vertices
-50148a55968
+50148a55969
 > weaponize/DSG
-50260,50261d56079
+50260,50261d56080
 < werwolf/M
 < werwolves
-50728c56546
+50728c56547
 < women
 ---
 > women/M
-50794c56612
+50794c56613
 < wop/S!
 ---
 > wop/MS!

extensions/spellcheck/locales/en-US/hunspell/en-US.dic

@@ -1,4 +1,4 @@
-57434
+57435
 0/nm
 0th/pt
 1/n1
@@ -48141,6 +48141,7 @@ scorn/MDRSZG
 scorner/M
 scornful/Y
 scorpion/MS
+scot-free
 scotch/MDSG
 scotchs
 scoundrel/MS

extensions/spellcheck/src/mozInlineSpellWordUtil.cpp

@@ -888,23 +888,55 @@ WordSplitState::ClassifyCharacter(PRInt32 aIndex, PRBool aRecurse) const
       return CHAR_CLASS_SEPARATOR;
     if (ClassifyCharacter(aIndex - 1, false) != CHAR_CLASS_WORD)
       return CHAR_CLASS_SEPARATOR;
+    // If the previous charatcer is a word-char, make sure that it's not a
+    // special dot character.
+    if (mDOMWordText[aIndex - 1] == '.')
+      return CHAR_CLASS_SEPARATOR;
 
     // now we know left char is a word-char, check the right-hand character
     if (aIndex == PRInt32(mDOMWordText.Length()) - 1)
       return CHAR_CLASS_SEPARATOR;
     if (ClassifyCharacter(aIndex + 1, false) != CHAR_CLASS_WORD)
       return CHAR_CLASS_SEPARATOR;
+    // If the next charatcer is a word-char, make sure that it's not a
+    // special dot character.
+    if (mDOMWordText[aIndex + 1] == '.')
+      return CHAR_CLASS_SEPARATOR;
 
     // char on either side is a word, this counts as a word
     return CHAR_CLASS_WORD;
   }
 
+  // The dot character, if appearing at the end of a word, should
+  // be considered part of that word.  Example: "etc.", or
+  // abbreviations
+  if (aIndex > 0 &&
+      mDOMWordText[aIndex] == '.' &&
+      mDOMWordText[aIndex - 1] != '.' &&
+      ClassifyCharacter(aIndex - 1, false) != CHAR_CLASS_WORD) {
+    return CHAR_CLASS_WORD;
+  }
+
   // all other punctuation
   if (charCategory == nsIUGenCategory::kSeparator ||
       charCategory == nsIUGenCategory::kOther ||
       charCategory == nsIUGenCategory::kPunctuation ||
-      charCategory == nsIUGenCategory::kSymbol)
+      charCategory == nsIUGenCategory::kSymbol) {
+    // Don't break on hyphens, as hunspell handles them on its own.
+    if (aIndex > 0 &&
+        mDOMWordText[aIndex] == '-' &&
+        mDOMWordText[aIndex - 1] != '-' &&
+        ClassifyCharacter(aIndex - 1, false) == CHAR_CLASS_WORD) {
+      // A hyphen is only meaningful as a separator inside a word
+      // if the previous and next characters are a word character.
+      if (aIndex == PRInt32(mDOMWordText.Length()) - 1)
         return CHAR_CLASS_SEPARATOR;
+      if (mDOMWordText[aIndex + 1] != '.' &&
+          ClassifyCharacter(aIndex + 1, false) == CHAR_CLASS_WORD)
+        return CHAR_CLASS_WORD;
+    }
+    return CHAR_CLASS_SEPARATOR;
+  }
 
   // any other character counts as a word
   return CHAR_CLASS_WORD;

layout/reftests/editor/reftest.list

@@ -47,7 +47,7 @@ needs-focus == input-text-notheme-onfocus-reframe.html input-text-notheme-onfocu
 needs-focus == caret_after_reframe.html caret_after_reframe-ref.html
 == nobogusnode-1.html nobogusnode-ref.html
 == nobogusnode-2.html nobogusnode-ref.html
-fails-if(!Android) == spellcheck-hyphen-valid.html spellcheck-hyphen-valid-ref.html
+fails-if(Android) == spellcheck-hyphen-valid.html spellcheck-hyphen-valid-ref.html
 fails-if(Android) != spellcheck-hyphen-invalid.html spellcheck-hyphen-invalid-ref.html
 == spellcheck-slash-valid.html spellcheck-slash-valid-ref.html
 == spellcheck-period-valid.html spellcheck-period-valid-ref.html