Merge remote-tracking branch 'valloric/master'

.gitmodules

@@ -13,6 +13,6 @@
 [submodule "third_party/watchdog"]
 	path = third_party/watchdog_deps/watchdog
 	url = https://github.com/gorakhargosh/watchdog
-[submodule "third_party/mrab-regex"]
-	path = third_party/mrab-regex
-	url = https://bitbucket.org/mrabarnett/mrab-regex.git
+[submodule "third_party/mrab-regex-github"]
+	path = third_party/mrab-regex-github
+	url = https://github.com/mrabarnett/mrab-regex

CORE_VERSION

@@ -1 +1 @@
-47
+48

build.py

@@ -764,7 +764,7 @@ def BuildRegexModule( script_args ):
   lib_dir = p.join( DIR_OF_THIRD_PARTY, 'regex-build' )
 
   try:
-    os.chdir( p.join( DIR_OF_THIRD_PARTY, 'mrab-regex' ) )
+    os.chdir( p.join( DIR_OF_THIRD_PARTY, 'mrab-regex-github' ) )
 
     RemoveDirectoryIfExists( build_dir )
     RemoveDirectoryIfExists( lib_dir )

cpp/ycm/Character.cpp

@@ -31,7 +31,7 @@ bool CodePointCompare( const CodePoint *left, const CodePoint *right ) {
 
 
 // Sort the code points according to the Canonical Ordering Algorithm.
-// See https://www.unicode.org/versions/Unicode13.0.0/ch03.pdf#G49591
+// See https://www.unicode.org/versions/latest/ch03.pdf#G49591
 CodePointSequence CanonicalSort( CodePointSequence code_points ) {
   auto code_point_start = code_points.begin();
   auto code_point_end = code_points.end();
@@ -64,7 +64,7 @@ CodePointSequence CanonicalSort( CodePointSequence code_points ) {
 
 // Decompose a UTF-8 encoded string into a sequence of code points according to
 // Canonical Decomposition. See
-// https://www.unicode.org/versions/Unicode13.0.0/ch03.pdf#G733
+// https://www.unicode.org/versions/latest/ch03.pdf#G733
 CodePointSequence CanonicalDecompose( std::string_view text ) {
   assert( NormalizeInput( text ) == text );
   return CanonicalSort( BreakIntoCodePoints( text ) );
@@ -78,7 +78,7 @@ Character::Character( std::string_view character )
     is_punctuation_( false ),
     is_uppercase_( false ) {
   // Normalize the character through NFD (Normalization Form D). See
-  // https://www.unicode.org/versions/Unicode13.0.0/ch03.pdf#G49621
+  // https://www.unicode.org/versions/latest/ch03.pdf#G49621
   CodePointSequence code_points = CanonicalDecompose( character );
 
   for ( const auto &code_point : code_points ) {
@@ -89,10 +89,10 @@ Character::Character( std::string_view character )
     is_punctuation_ |= code_point->IsPunctuation();
     is_uppercase_ |= code_point->IsUppercase();
 
-    switch ( code_point->GetBreakProperty() ) {
-      case BreakProperty::PREPEND:
-      case BreakProperty::EXTEND:
-      case BreakProperty::SPACINGMARK:
+    switch ( code_point->GetGraphemeBreakProperty() ) {
+      case GraphemeBreakProperty::PREPEND:
+      case GraphemeBreakProperty::EXTEND:
+      case GraphemeBreakProperty::SPACINGMARK:
         is_base_ = false;
         break;
       default:

cpp/ycm/Character.h

@@ -27,7 +27,7 @@ namespace YouCompleteMe {
 // This class represents a UTF-8 character. It takes a UTF-8 encoded string
 // corresponding to a grapheme cluster (see
 // https://www.unicode.org/glossary/#grapheme_cluster), normalize it through NFD
-// (see https://www.unicode.org/versions/Unicode13.0.0/ch03.pdf#G49621), and
+// (see https://www.unicode.org/versions/latest/ch03.pdf#G49621), and
 // compute the folded and swapped case versions of the normalized character. It
 // also holds some properties like if the character is a letter or a
 // punctuation, and if it is uppercase.

cpp/ycm/ClangCompleter/TranslationUnitStore.cpp

@@ -122,8 +122,8 @@ shared_ptr< TranslationUnit > TranslationUnitStore::Get(
 
 bool TranslationUnitStore::Remove( const std::string &filename ) {
   lock_guard< mutex > lock( filename_to_translation_unit_and_flags_mutex_ );
-  Erase( filename_to_flags_hash_, filename );
-  return Erase( filename_to_translation_unit_, filename );
+  filename_to_flags_hash_.erase( filename );
+  return filename_to_translation_unit_.erase( filename );
 }
 
 

cpp/ycm/CodePoint.cpp

@@ -67,10 +67,11 @@ RawCodePoint FindCodePoint( std::string_view text ) {
              code_points.is_punctuation[ index ],
              code_points.is_uppercase[ index ],
              code_points.break_property[ index ],
-             code_points.combining_class[ index ] };
+             code_points.combining_class[ index ],
+             code_points.indic_conjunct_break[ index ] };
   }
 
-  return { text, text, text, text, false, false, false, 0, 0 };
+  return { text, text, text, text, false, false, false, 0, 0, 0 };
 }
 
 } // unnamed namespace
@@ -87,9 +88,11 @@ CodePoint::CodePoint( RawCodePoint&& code_point )
     is_letter_( code_point.is_letter ),
     is_punctuation_( code_point.is_punctuation ),
     is_uppercase_( code_point.is_uppercase ),
-    break_property_(
-      static_cast< BreakProperty >( code_point.break_property ) ),
-    combining_class_( code_point.combining_class ) {
+    grapheme_break_property_(
+      static_cast< GraphemeBreakProperty >( code_point.grapheme_break_property ) ),
+    combining_class_( code_point.combining_class ),
+    indic_conjunct_break_property_(
+      static_cast< IndicConjunctBreakProperty >( code_point.indic_conjunct_break_property ) ) {
 }
 
 

cpp/ycm/CodePoint.h

@@ -26,10 +26,10 @@
 namespace YouCompleteMe {
 
 // See
-// http://www.unicode.org/reports/tr29/tr29-37.html#Grapheme_Cluster_Break_Property_Values
+// http://www.unicode.org/reports/tr29#Grapheme_Cluster_Break_Property_Values
 // NOTE: The properties must take the same value as the ones defined in the
 // update_unicode.py script.
-enum class BreakProperty : uint8_t {
+enum class GraphemeBreakProperty : uint8_t {
   OTHER              =  0,
   CR                 =  1,
   LF                 =  2,
@@ -46,6 +46,15 @@ enum class BreakProperty : uint8_t {
   LVT                = 13,
   EXTPICT            = 18
 };
+// See https://www.unicode.org/reports/tr44/#Indic_Conjunct_Break
+// NOTE: The properties must take the same value as the ones defined in the
+// update_unicode.py script.
+enum class IndicConjunctBreakProperty : uint8_t {
+  None      =  0,
+  LINKER    =  1,
+  CONSONANT =  2,
+  EXTEND    =  3,
+};
 
 
 // This is the structure used to store the data in the Unicode table. See the
@@ -58,8 +67,9 @@ struct RawCodePoint {
   bool is_letter;
   bool is_punctuation;
   bool is_uppercase;
-  uint8_t break_property;
+  uint8_t grapheme_break_property;
   uint8_t combining_class;
+  uint8_t indic_conjunct_break_property;
 };
 
 
@@ -81,7 +91,7 @@ struct RawCodePoint {
 //  - its breaking property: used to split a word into characters.
 //  - its combining class: used to sort a sequence of code points according to
 //    the Canonical Ordering algorithm (see
-//    https://www.unicode.org/versions/Unicode13.0.0/ch03.pdf#G49591).
+//    https://www.unicode.org/versions/latest/ch03.pdf#G49591).
 class CodePoint {
 public:
   YCM_EXPORT explicit CodePoint( std::string_view code_point );
@@ -115,14 +125,18 @@ class CodePoint {
     return is_uppercase_;
   }
 
-  inline BreakProperty GetBreakProperty() const {
-    return break_property_;
+  inline GraphemeBreakProperty GetGraphemeBreakProperty() const {
+    return grapheme_break_property_;
   }
 
   inline uint8_t CombiningClass() const {
     return combining_class_;
   }
 
+  inline IndicConjunctBreakProperty GetIndicConjunctBreakProperty() const {
+    return indic_conjunct_break_property_;
+  }
+
   inline bool operator< ( const CodePoint &other ) const {
     return combining_class_ < other.combining_class_;
   }
@@ -136,8 +150,9 @@ class CodePoint {
   bool is_letter_;
   bool is_punctuation_;
   bool is_uppercase_;
-  BreakProperty break_property_;
+  GraphemeBreakProperty grapheme_break_property_;
   uint8_t combining_class_;
+  IndicConjunctBreakProperty indic_conjunct_break_property_;
 };
 
 

cpp/ycm/Utils.h

@@ -99,19 +99,6 @@ FindWithDefault( Container &container,
 }
 
 
-template <class Container, class Key>
-bool Erase( Container &container, const Key &key ) {
-  typename Container::iterator it = container.find( key );
-
-  if ( it != container.end() ) {
-    container.erase( it );
-    return true;
-  }
-
-  return false;
-}
-
-
 // Shrink a vector to its sorted |num_sorted_elements| smallest elements. If
 // |num_sorted_elements| is 0 or larger than the vector size, sort the whole
 // vector.