[GR-18163] Fix String#split to return empty array when splitting all …

…whitespace on whitespace PullRequest: truffleruby/3115
oracle · Jan 13, 2022 · 6dd971c · 6dd971c
2 parents 754cfcb + c9279eb
commit 6dd971c
Show file tree

Hide file tree

Showing 3 changed files with 15 additions and 3 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -7,6 +7,7 @@ Bug fixes:
 
 * Guard against unterminated ranges in file matching patterns (#2556, @aardvark179).
 * Fixed `rb_proc_new` to return a proc that will pass all required arguments to C (#2556, @aardvark179).
+* Fixed `String#split` to return empty array when splitting all whitespace on whitespace (#2565, @bjfish).
 
 Compatibility:
 

diff --git a/spec/ruby/core/string/split_spec.rb b/spec/ruby/core/string/split_spec.rb
@@ -224,6 +224,17 @@
       end
     end
   end
+
+  it "returns an empty array when whitespace is split on whitespace" do
+    " ".split(" ").should == []
+    " \n ".split(" ").should == []
+    "  ".split(" ").should == []
+    " \t ".split(" ").should == []
+  end
+
+  it "doesn't split on non-ascii whitespace" do
+    "a\u{2008}b".split(" ").should == ["a\u{2008}b"]
+  end
 end
 
 describe "String#split with Regexp" do

diff --git a/src/main/java/org/truffleruby/core/string/StringNodes.java b/src/main/java/org/truffleruby/core/string/StringNodes.java
@@ -3607,7 +3607,7 @@ protected Object stringAwkSplitSingleByte(Object string, int limit, Object block
                 }
             }
 
-            if (trailingSubstringProfile.profile(substringStart != SUBSTRING_CREATED)) {
+            if (trailingSubstringProfile.profile(findingSubstringEnd)) {
                 final RubyString substring = substringNode
                         .executeSubstring(string, substringStart, bytes.length - substringStart);
                 ret = addSubstring(ret, storeIndex++, substring, block, executeBlockProfile, growArrayProfile);
@@ -3655,7 +3655,7 @@ protected Object stringAwkSplit(Object string, int limit, Object block,
                 p += StringSupport.characterLength(enc, cr, bytes, p, end, true);
 
                 if (skip) {
-                    if (StringSupport.isSpace(enc, c)) {
+                    if (StringSupport.isAsciiSpace(c)) {
                         b = p - ptr;
                     } else {
                         e = p - ptr;
@@ -3665,7 +3665,7 @@ protected Object stringAwkSplit(Object string, int limit, Object block,
                         }
                     }
                 } else {
-                    if (StringSupport.isSpace(enc, c)) {
+                    if (StringSupport.isAsciiSpace(c)) {
                         final RubyString substring = substringNode.executeSubstring(string, b, e - b);
                         ret = addSubstring(
                                 ret,