Fix ranking bug with ALL_CAPS variables

GetWordBoundaryChars would return "allcaps" for "ALL_CAPS" instead of "ac". This
would manifest as ranking (for instance) "STDIN_FILENO" ahead of "stdin" for
query "std", which is terrible.

This bug has been present in YCM for many months, but no one noticed the issue.

Fixes #272.

cpp/ycm/Candidate.cpp

@@ -28,21 +28,6 @@ namespace YouCompleteMe {
 
 namespace {
 
-std::string GetWordBoundaryChars( const std::string &text ) {
-  std::string result;
-
-  for ( uint i = 0; i < text.size(); ++i ) {
-    if ( i == 0 ||
-         IsUppercase( text[ i ] ) ||
-         ( i > 0 && text[ i - 1 ] == '_' && isalpha( text[ i ] ) )
-       ) {
-      result.push_back( tolower( text[ i ] ) );
-    }
-  }
-
-  return result;
-}
-
 LetterNode *FirstUppercaseNode( const std::list< LetterNode *> &list ) {
   LetterNode *node = NULL;
   foreach( LetterNode * current_node, list ) {
@@ -67,6 +52,28 @@ LetterNode *FirstLowercaseNode( const std::list< LetterNode *> &list ) {
 
 } // unnamed namespace
 
+std::string GetWordBoundaryChars( const std::string &text ) {
+  std::string result;
+
+  for ( uint i = 0; i < text.size(); ++i ) {
+    bool is_first_char_but_not_underscore = i == 0 && text[ i ] != '_';
+    bool is_good_uppercase = i > 0 &&
+                             IsUppercase( text[ i ] ) &&
+                             !IsUppercase( text[ i - 1 ] );
+    bool is_alpha_after_underscore = i > 0 &&
+                                     text[ i - 1 ] == '_' &&
+                                     isalpha( text[ i ] );
+
+    if ( is_first_char_but_not_underscore ||
+         is_good_uppercase ||
+         is_alpha_after_underscore ) {
+      result.push_back( tolower( text[ i ] ) );
+    }
+  }
+
+  return result;
+}
+
 
 Bitset LetterBitsetFromString( const std::string &text ) {
   Bitset letter_bitset;

cpp/ycm/Candidate.h

@@ -34,6 +34,9 @@ typedef std::bitset< NUM_LETTERS > Bitset;
 
 Bitset LetterBitsetFromString( const std::string &text );
 
+// Public for tests
+std::string GetWordBoundaryChars( const std::string &text );
+
 class Candidate : boost::noncopyable {
 public:
 

cpp/ycm/tests/Candidate_test.cpp

@@ -21,6 +21,78 @@
 
 namespace YouCompleteMe {
 
+TEST( GetWordBoundaryCharsTest, SimpleOneWord ) {
+  EXPECT_EQ( "s", GetWordBoundaryChars( "simple" ) );
+}
+
+TEST( GetWordBoundaryCharsTest, UnderscoreInMiddle ) {
+  EXPECT_EQ( "sf", GetWordBoundaryChars( "simple_foo" ) );
+}
+
+TEST( GetWordBoundaryCharsTest, UnderscoreStart ) {
+  EXPECT_EQ( "s", GetWordBoundaryChars( "_simple" ) );
+}
+
+TEST( GetWordBoundaryCharsTest, ManyUnderscoreStart ) {
+  EXPECT_EQ( "s", GetWordBoundaryChars( "___simple" ) );
+}
+
+TEST( GetWordBoundaryCharsTest, UnderscoreStartAndInMiddle ) {
+  EXPECT_EQ( "sf", GetWordBoundaryChars( "_simple_foo" ) );
+}
+
+TEST( GetWordBoundaryCharsTest, ManyUnderscoreStartAndInMiddle ) {
+  EXPECT_EQ( "sf", GetWordBoundaryChars( "___simple__foo" ) );
+}
+
+TEST( GetWordBoundaryCharsTest, SimpleCapitalStart ) {
+  EXPECT_EQ( "s", GetWordBoundaryChars( "Simple" ) );
+}
+
+TEST( GetWordBoundaryCharsTest, SimpleCapitalTwoWord ) {
+  EXPECT_EQ( "ss", GetWordBoundaryChars( "SimpleStuff" ) );
+}
+
+TEST( GetWordBoundaryCharsTest, SimpleCapitalTwoWordUnderscoreMiddle ) {
+  EXPECT_EQ( "ss", GetWordBoundaryChars( "Simple_Stuff" ) );
+}
+
+TEST( GetWordBoundaryCharsTest, JavaCase ) {
+  EXPECT_EQ( "ssf", GetWordBoundaryChars( "simpleStuffFoo" ) );
+}
+
+TEST( GetWordBoundaryCharsTest, UppercaseSequence ) {
+  EXPECT_EQ( "ss", GetWordBoundaryChars( "simpleSTUFF" ) );
+}
+
+TEST( GetWordBoundaryCharsTest, UppercaseSequenceInMiddle ) {
+  EXPECT_EQ( "ss", GetWordBoundaryChars( "simpleSTUFFfoo" ) );
+}
+
+TEST( GetWordBoundaryCharsTest, UppercaseSequenceInMiddleUnderscore ) {
+  EXPECT_EQ( "ssf", GetWordBoundaryChars( "simpleSTUFF_Foo" ) );
+}
+
+TEST( GetWordBoundaryCharsTest, UppercaseSequenceInMiddleUnderscoreLowercase ) {
+  EXPECT_EQ( "ssf", GetWordBoundaryChars( "simpleSTUFF_foo" ) );
+}
+
+TEST( GetWordBoundaryCharsTest, AllCapsSimple ) {
+  EXPECT_EQ( "s", GetWordBoundaryChars( "SIMPLE" ) );
+}
+
+TEST( GetWordBoundaryCharsTest, AllCapsUnderscoreStart ) {
+  EXPECT_EQ( "s", GetWordBoundaryChars( "_SIMPLE" ) );
+}
+
+TEST( GetWordBoundaryCharsTest, AllCapsUnderscoreMiddle ) {
+  EXPECT_EQ( "ss", GetWordBoundaryChars( "SIMPLE_STUFF" ) );
+}
+
+TEST( GetWordBoundaryCharsTest, AllCapsUnderscoreMiddleAndStart ) {
+  EXPECT_EQ( "ss", GetWordBoundaryChars( "_SIMPLE_STUFF" ) );
+}
+
 TEST( CandidateTest, TextValid ) {
   std::string text = "foo";
   Candidate candidate( text );

cpp/ycm/tests/IdentifierCompleter_test.cpp

@@ -207,6 +207,15 @@ TEST( IdentifierCompleterTest, SameLowercaseCandidateWins ) {
                             "Foobar" ) );
 }
 
+TEST( IdentifierCompleterTest, ShorterAndLowercaseWins ) {
+  EXPECT_THAT( IdentifierCompleter(
+                 StringVector(
+                   "STDIN_FILENO",
+                   "stdin" ) ).CandidatesForQuery( "std" ),
+               ElementsAre( "stdin",
+                            "STDIN_FILENO" ) );
+}
+
 // TODO: tests for filepath and filetype candidate storing
 
 } // namespace YouCompleteMe