Fix str.isspace and str.split

tej-prash · Oct 12, 2017 · 9c6cea0 · 9c6cea0
1 parent ef5ba51
commit 9c6cea0
Show file tree

Hide file tree

Showing 2 changed files with 11 additions and 8 deletions.
diff --git a/python/common/org/python/types/Str.java b/python/common/org/python/types/Str.java
@@ -953,7 +953,7 @@ public org.python.Object isspace() {
             return new org.python.types.Bool(false);
         }
         for (char ch : this.value.toCharArray()) {
-            if (" \t\n\r".indexOf(ch) == -1) {
+            if (!Character.isWhitespace(ch)) {
                 return new org.python.types.Bool(false);
             }
         }
@@ -1394,8 +1394,10 @@ public org.python.Object split(org.python.Object sep, org.python.Object maxsplit
             }
         }
 
+        java.lang.String value = this.value.toString();
         if (sep == null) {
-            sep = new org.python.types.Str(" ");
+            value = value.trim();
+            sep = new org.python.types.Str("\\s+");
         } else if (!(sep instanceof org.python.types.Str)) {
             if (org.Python.VERSION < 0x03060000) {
                 throw new org.python.exceptions.TypeError("Can't convert '" + sep.typeName() + "' object to str implicitly");
@@ -1406,10 +1408,10 @@ public org.python.Object split(org.python.Object sep, org.python.Object maxsplit
 
         java.lang.String[] result;
         if (maxsplit == null) {
-            result = this.value.toString().split(((org.python.types.Str) sep).toString());
+            result = value.split(((org.python.types.Str) sep).toString());
         } else {
             int number = java.lang.Integer.parseInt(maxsplit.toString());
-            result = this.value.toString().split(((org.python.types.Str) sep).toString(), number + 1);
+            result = value.split(((org.python.types.Str) sep).toString(), number + 1);
         }
         org.python.types.List result_list = new org.python.types.List();
         for (java.lang.String w : result) {

diff --git a/tests/datatypes/test_str.py b/tests/datatypes/test_str.py
@@ -28,8 +28,8 @@ def test_isdigit(self):
             """)
 
     def test_isspace(self):
-        self.assertCodeExecution("""
-            for s in ['''  \t \r''', ' ', '\t\tnope\t\t', '']:
+        self.assertCodeExecution(r"""
+            for s in ['\x1f \v \f \n \t \r', ' ', '\t\tnope\t\t', '']:
                 print(s.isspace())
             """)
 
@@ -231,8 +231,8 @@ def test_case_changes(self):
             """)
 
     def test_split(self):
-        self.assertCodeExecution("""
-            for s in ['hello, world', 'HEllo, WORLD', 'átomo', '']:
+        self.assertCodeExecution(r"""
+            for s in ['\vhello,        world', '\nHEllo, WORLD\f', 'átomo', '']:
                 print(s.split())
                 print(s.split("o"))
                 print(s.split(maxsplit=2))
@@ -241,6 +241,7 @@ def test_split(self):
                     print(s.split(5))
                 except TypeError as err:
                     print(err)
+
             """)
 
     def test_index(self):