TrimGlobalStep handles unicode characters as Ltrim and Rtrim

apache · Jun 28, 2024 · d391ce8 · d391ce8
1 parent dc6e715
commit d391ce8
Show file tree

Hide file tree

Showing 15 changed files with 115 additions and 57 deletions.
diff --git a/CHANGELOG.asciidoc b/CHANGELOG.asciidoc
@@ -30,6 +30,7 @@ This release also includes changes from <<release-3-6-8, 3.6.8>>.
 * Deprecated public constructor for `SeedStrategy` in favor of builder pattern to be consistent with other strategies.
 * Allow specifying a customized Spark app name
 * CoinStep has a getter method for its probability field
+* Fixed so that TrimGlobalStep and TrimLocalStep have the same character control handling as Ltrim and Rtrim
 
 [[release-3-7-2]]
 === TinkerPop 3.7.2 (April 8, 2024)

diff --git a/...rc/main/java/org/apache/tinkerpop/gremlin/process/traversal/step/map/LTrimGlobalStep.java b/...rc/main/java/org/apache/tinkerpop/gremlin/process/traversal/step/map/LTrimGlobalStep.java
@@ -25,6 +25,8 @@
 import java.util.Collections;
 import java.util.Set;
 
+import static org.apache.tinkerpop.gremlin.process.traversal.step.util.StringLocalStep.getTrimmedString;
+
 /**
  * Reference implementation for lTrim() step, a mid-traversal step which returns a string with leading
  * whitespace removed. Null values are not processed and remain as null when returned.
@@ -49,15 +51,7 @@ protected E map(final Traverser.Admin<S> traverser) {
         }
 
         // we will pass null values to next step
-        if (null == item)
-                return null;
-
-        int i = 0;
-        while (i < ((String) item).length() && Character.isWhitespace(((String) item).charAt(i))) {
-            i++;
-        }
-
-        return (E) ((String) item).substring(i);
+        return (E) getTrimmedString((String) item, true, false);
     }
 
     @Override

diff --git a/...src/main/java/org/apache/tinkerpop/gremlin/process/traversal/step/map/LTrimLocalStep.java b/...src/main/java/org/apache/tinkerpop/gremlin/process/traversal/step/map/LTrimLocalStep.java
@@ -51,18 +51,9 @@ public Set<TraverserRequirement> getRequirements() {
 
     @Override
     protected E applyStringOperation(String item) {
-        return (E) item.substring(getIdx(item));
+        return (E) getTrimmedString(item, true, false);
     }
 
     @Override
     public String getStepName() { return "lTrim(local)"; }
-
-    private int getIdx(final String str) {
-        int idx = 0;
-        while (idx < str.length() && Character.isWhitespace(str.charAt(idx))) {
-            idx++;
-        }
-        return idx;
-    }
-
 }
diff --git a/...rc/main/java/org/apache/tinkerpop/gremlin/process/traversal/step/map/RTrimGlobalStep.java b/...rc/main/java/org/apache/tinkerpop/gremlin/process/traversal/step/map/RTrimGlobalStep.java
@@ -25,6 +25,8 @@
 import java.util.Collections;
 import java.util.Set;
 
+import static org.apache.tinkerpop.gremlin.process.traversal.step.util.StringLocalStep.getTrimmedString;
+
 /**
  * Reference implementation for rTrim() step, a mid-traversal step which a string with trailing
  * whitespace removed. Null values are not processed and remain as null when returned.
@@ -49,15 +51,7 @@ protected E map(final Traverser.Admin<S> traverser) {
         }
 
         // we will pass null values to next step
-        if (null == item)
-                return null;
-
-        int i = ((String) item).length() - 1;
-        while (i >= 0 && Character.isWhitespace(((String) item).charAt(i))) {
-            i--;
-        }
-
-        return (E) ((String) item).substring(0,i+1);
+        return (E) getTrimmedString((String) item, false, true);
     }
 
     @Override

diff --git a/...src/main/java/org/apache/tinkerpop/gremlin/process/traversal/step/map/RTrimLocalStep.java b/...src/main/java/org/apache/tinkerpop/gremlin/process/traversal/step/map/RTrimLocalStep.java
@@ -37,18 +37,9 @@ public RTrimLocalStep(final Traversal.Admin traversal) {
 
     @Override
     protected E applyStringOperation(String item) {
-        return (E) item.substring(0,getEndIdx(item)+1);
+        return (E) getTrimmedString(item, false, true);
     }
 
     @Override
     public String getStepName() { return "rTrim(local)"; }
-
-    private int getEndIdx(final String str) {
-        int idx = str.length() - 1;
-        while (idx >= 0 && Character.isWhitespace(str.charAt(idx))) {
-            idx--;
-        }
-        return idx;
-    }
-
 }
diff --git a/...src/main/java/org/apache/tinkerpop/gremlin/process/traversal/step/map/TrimGlobalStep.java b/...src/main/java/org/apache/tinkerpop/gremlin/process/traversal/step/map/TrimGlobalStep.java
@@ -25,6 +25,8 @@
 import java.util.Collections;
 import java.util.Set;
 
+import static org.apache.tinkerpop.gremlin.process.traversal.step.util.StringLocalStep.getTrimmedString;
+
 /**
  * Reference implementation for trim() step, a mid-traversal step which returns a string with leading and trailing
  * whitespace removed. Null values are not processed and remain as null when returned.
@@ -47,9 +49,7 @@ protected E map(final Traverser.Admin<S> traverser) {
             throw new IllegalArgumentException(
                     String.format("The trim() step can only take string as argument, encountered %s", item.getClass()));
         }
-
-        // we will pass null values to next step
-        return null == item? null : (E) ((String) item).trim();
+        return (E) getTrimmedString((String) item, true, true);
     }
 
     @Override

diff --git a/.../src/main/java/org/apache/tinkerpop/gremlin/process/traversal/step/map/TrimLocalStep.java b/.../src/main/java/org/apache/tinkerpop/gremlin/process/traversal/step/map/TrimLocalStep.java
@@ -37,7 +37,7 @@ public TrimLocalStep(final Traversal.Admin traversal) {
 
     @Override
     protected E applyStringOperation(String item) {
-        return (E) item.trim();
+        return (E) getTrimmedString(item, true, true);
     }
 
     @Override

diff --git a/...c/main/java/org/apache/tinkerpop/gremlin/process/traversal/step/util/StringLocalStep.java b/...c/main/java/org/apache/tinkerpop/gremlin/process/traversal/step/util/StringLocalStep.java
@@ -72,6 +72,27 @@ public Set<TraverserRequirement> getRequirements() {
         return Collections.singleton(TraverserRequirement.OBJECT);
     }
 
+    public static String getTrimmedString(final String s, final boolean lTrim, final boolean rTrim) {
+        if (s == null) {
+            return null;
+        }
+
+        int start = 0;
+        if (lTrim) {
+            while (start < s.length() && Character.isWhitespace(s.charAt(start))) {
+                start++;
+            }
+        }
+
+        int end = s.length() - 1;
+        if (rTrim) {
+            while (end >= start && Character.isWhitespace(s.charAt(end))) {
+                end--;
+            }
+        }
+        return s.substring(start, end + 1);
+    }
+
     protected abstract E applyStringOperation(final String item);
 
     protected abstract String getStepName();

diff --git a/gremlin-dotnet/test/Gremlin.Net.IntegrationTest/Gherkin/Gremlin.cs b/gremlin-dotnet/test/Gremlin.Net.IntegrationTest/Gherkin/Gremlin.cs
@@ -749,7 +749,8 @@ private static IDictionary<string, List<Func<GraphTraversalSource, IDictionary<s
                {"g_injectXa_null_bX_intersectXa_cX", new List<Func<GraphTraversalSource, IDictionary<string, object>, ITraversal>> {(g,p) =>g.Inject(p["xx1"]).Intersect(p["xx2"])}}, 
                {"g_injectXa_null_bX_intersectXa_null_cX", new List<Func<GraphTraversalSource, IDictionary<string, object>, ITraversal>> {(g,p) =>g.Inject(p["xx1"]).Intersect(p["xx2"])}}, 
                {"g_injectX3_threeX_intersectXfive_three_7X", new List<Func<GraphTraversalSource, IDictionary<string, object>, ITraversal>> {(g,p) =>g.Inject(p["xx1"]).Intersect(p["xx2"])}}, 
-               {"g_injectX__feature___test__nullX_lTrim", new List<Func<GraphTraversalSource, IDictionary<string, object>, ITraversal>> {(g,p) =>g.Inject<object>("  feature"," one test",null,""," ").LTrim()}}, 
+               {"g_injectX__feature___test__nullX_lTrim", new List<Func<GraphTraversalSource, IDictionary<string, object>, ITraversal>> {(g,p) =>g.Inject<object>("  feature"," one test",null,""," ","\u3000abc","abc\u3000","\u3000abc\u3000","\u3000\u3000").LTrim()}}, 
+               {"g_injectX__feature___test__nullX_lTrimXlocalX", new List<Func<GraphTraversalSource, IDictionary<string, object>, ITraversal>> {(g,p) =>g.Inject(new List<object> {"  feature  ", " one test ", null, "", " ", "\u3000abc", "abc\u3000", "\u3000abc\u3000", "\u3000\u3000"}).LTrim<object>(Scope.Local)}}, 
                {"g_injectX__feature__X_lTrim", new List<Func<GraphTraversalSource, IDictionary<string, object>, ITraversal>> {(g,p) =>g.Inject("  feature  ").LTrim()}}, 
                {"g_injectXListXa_bXX_lTrim", new List<Func<GraphTraversalSource, IDictionary<string, object>, ITraversal>> {(g,p) =>g.Inject(p["xx1"]).LTrim()}}, 
                {"g_injectXListX1_2XX_lTrimXlocalX", new List<Func<GraphTraversalSource, IDictionary<string, object>, ITraversal>> {(g,p) =>g.Inject(p["xx1"]).LTrim<object>(Scope.Local)}}, 
@@ -1075,7 +1076,8 @@ private static IDictionary<string, List<Func<GraphTraversalSource, IDictionary<s
                {"g_V_hasXageX_propertiesXage_nameX_value", new List<Func<GraphTraversalSource, IDictionary<string, object>, ITraversal>> {(g,p) =>g.V().Has("age").Properties<object>("age","name").Value<object>()}}, 
                {"g_V_propertiesXname_age_nullX_value", new List<Func<GraphTraversalSource, IDictionary<string, object>, ITraversal>> {(g,p) =>g.V().Properties<object>("name","age",null).Value<object>()}}, 
                {"g_V_valuesXname_age_nullX", new List<Func<GraphTraversalSource, IDictionary<string, object>, ITraversal>> {(g,p) =>g.V().Values<object>("name","age",null)}}, 
-               {"g_injectX__feature___test__nullX_rTrim", new List<Func<GraphTraversalSource, IDictionary<string, object>, ITraversal>> {(g,p) =>g.Inject<object>("feature  ","one test ",null,""," ").RTrim()}}, 
+               {"g_injectX__feature___test__nullX_rTrim", new List<Func<GraphTraversalSource, IDictionary<string, object>, ITraversal>> {(g,p) =>g.Inject<object>("feature  ","one test ",null,""," ","\u3000abc","abc\u3000","\u3000abc\u3000","\u3000\u3000").RTrim()}}, 
+               {"g_injectX__feature___test__nullX_rTrimXlocalX", new List<Func<GraphTraversalSource, IDictionary<string, object>, ITraversal>> {(g,p) =>g.Inject(new List<object> {"  feature  ", " one test ", null, "", " ", "\u3000abc", "abc\u3000", "\u3000abc\u3000", "\u3000\u3000"}).RTrim<object>(Scope.Local)}}, 
                {"g_injectX__feature__X_rTrim", new List<Func<GraphTraversalSource, IDictionary<string, object>, ITraversal>> {(g,p) =>g.Inject("  feature  ").RTrim()}}, 
                {"g_injectXListXa_bXX_rTrim", new List<Func<GraphTraversalSource, IDictionary<string, object>, ITraversal>> {(g,p) =>g.Inject(p["xx1"]).RTrim()}}, 
                {"g_injectXListX1_2XX_rTrimXlocalX", new List<Func<GraphTraversalSource, IDictionary<string, object>, ITraversal>> {(g,p) =>g.Inject(p["xx1"]).RTrim<object>(Scope.Local)}}, 
@@ -1225,7 +1227,8 @@ private static IDictionary<string, List<Func<GraphTraversalSource, IDictionary<s
                {"g_V_valuesXnameX_toUpper", new List<Func<GraphTraversalSource, IDictionary<string, object>, ITraversal>> {(g,p) =>g.V().Values<object>("name").ToUpper()}}, 
                {"g_V_valuesXnameX_toUpperXlocalX", new List<Func<GraphTraversalSource, IDictionary<string, object>, ITraversal>> {(g,p) =>g.V().Values<object>("name").ToUpper<object>(Scope.Local)}}, 
                {"g_V_valuesXnameX_order_fold_toUpperXlocalX", new List<Func<GraphTraversalSource, IDictionary<string, object>, ITraversal>> {(g,p) =>g.V().Values<object>("name").Order().Fold().ToUpper<object>(Scope.Local)}}, 
-               {"g_injectX__feature___test__nullX_trim", new List<Func<GraphTraversalSource, IDictionary<string, object>, ITraversal>> {(g,p) =>g.Inject<object>("  feature  "," one test ",null,""," ").Trim()}}, 
+               {"g_injectX__feature___test__nullX_trim", new List<Func<GraphTraversalSource, IDictionary<string, object>, ITraversal>> {(g,p) =>g.Inject<object>("  feature  "," one test ",null,""," ","\u3000abc","abc\u3000","\u3000abc\u3000","\u3000\u3000").Trim()}}, 
+               {"g_injectX__feature___test__nullX_trimXlocalX", new List<Func<GraphTraversalSource, IDictionary<string, object>, ITraversal>> {(g,p) =>g.Inject(new List<object> {"  feature  ", " one test ", null, "", " ", "\u3000abc", "abc\u3000", "\u3000abc\u3000", "\u3000\u3000"}).Trim<object>(Scope.Local)}}, 
                {"g_injectXListXa_bXX_trim", new List<Func<GraphTraversalSource, IDictionary<string, object>, ITraversal>> {(g,p) =>g.Inject(p["xx1"]).Trim()}}, 
                {"g_injectXListX1_2XX_trimXlocalX", new List<Func<GraphTraversalSource, IDictionary<string, object>, ITraversal>> {(g,p) =>g.Inject(p["xx1"]).Trim<object>(Scope.Local)}}, 
                {"g_V_valuesXnameX_trim", new List<Func<GraphTraversalSource, IDictionary<string, object>, ITraversal>> {(g,p) =>g.AddV("person").Property("name"," marko ").Property("age",29).As("marko").AddV("person").Property("name","  vadas  ").Property("age",27).As("vadas").AddV("software").Property("name","  lop").Property("lang","java").As("lop").AddV("person").Property("name","josh  ").Property("age",32).As("josh").AddV("software").Property("name","   ripple   ").Property("lang","java").As("ripple").AddV("person").Property("name","peter").Property("age",35).As("peter").AddE("knows").From("marko").To("vadas").Property("weight",0.5).AddE("knows").From("marko").To("josh").Property("weight",1.0).AddE("created").From("marko").To("lop").Property("weight",0.4).AddE("created").From("josh").To("ripple").Property("weight",1.0).AddE("created").From("josh").To("lop").Property("weight",0.4).AddE("created").From("peter").To("lop").Property("weight",0.2), (g,p) =>g.V().Values<object>("name").Trim()}}, 

diff --git a/gremlin-go/driver/cucumber/gremlin.go b/gremlin-go/driver/cucumber/gremlin.go
@@ -720,7 +720,8 @@ var translationMap = map[string][]func(g *gremlingo.GraphTraversalSource, p map[
     "g_injectXa_null_bX_intersectXa_cX": {func(g *gremlingo.GraphTraversalSource, p map[string]interface{}) *gremlingo.GraphTraversal {return g.Inject(p["xx1"]).Intersect(p["xx2"])}}, 
     "g_injectXa_null_bX_intersectXa_null_cX": {func(g *gremlingo.GraphTraversalSource, p map[string]interface{}) *gremlingo.GraphTraversal {return g.Inject(p["xx1"]).Intersect(p["xx2"])}}, 
     "g_injectX3_threeX_intersectXfive_three_7X": {func(g *gremlingo.GraphTraversalSource, p map[string]interface{}) *gremlingo.GraphTraversal {return g.Inject(p["xx1"]).Intersect(p["xx2"])}}, 
-    "g_injectX__feature___test__nullX_lTrim": {func(g *gremlingo.GraphTraversalSource, p map[string]interface{}) *gremlingo.GraphTraversal {return g.Inject("  feature", " one test", nil, "", " ").LTrim()}}, 
+    "g_injectX__feature___test__nullX_lTrim": {func(g *gremlingo.GraphTraversalSource, p map[string]interface{}) *gremlingo.GraphTraversal {return g.Inject("  feature", " one test", nil, "", " ", "\u3000abc", "abc\u3000", "\u3000abc\u3000", "\u3000\u3000").LTrim()}}, 
+    "g_injectX__feature___test__nullX_lTrimXlocalX": {func(g *gremlingo.GraphTraversalSource, p map[string]interface{}) *gremlingo.GraphTraversal {return g.Inject([]interface{}{"  feature  ", " one test ", nil, "", " ", "\u3000abc", "abc\u3000", "\u3000abc\u3000", "\u3000\u3000"}).LTrim(gremlingo.Scope.Local)}}, 
     "g_injectX__feature__X_lTrim": {func(g *gremlingo.GraphTraversalSource, p map[string]interface{}) *gremlingo.GraphTraversal {return g.Inject("  feature  ").LTrim()}}, 
     "g_injectXListXa_bXX_lTrim": {func(g *gremlingo.GraphTraversalSource, p map[string]interface{}) *gremlingo.GraphTraversal {return g.Inject(p["xx1"]).LTrim()}}, 
     "g_injectXListX1_2XX_lTrimXlocalX": {func(g *gremlingo.GraphTraversalSource, p map[string]interface{}) *gremlingo.GraphTraversal {return g.Inject(p["xx1"]).LTrim(gremlingo.Scope.Local)}}, 
@@ -1046,7 +1047,8 @@ var translationMap = map[string][]func(g *gremlingo.GraphTraversalSource, p map[
     "g_V_hasXageX_propertiesXage_nameX_value": {func(g *gremlingo.GraphTraversalSource, p map[string]interface{}) *gremlingo.GraphTraversal {return g.V().Has("age").Properties("age", "name").Value()}}, 
     "g_V_propertiesXname_age_nullX_value": {func(g *gremlingo.GraphTraversalSource, p map[string]interface{}) *gremlingo.GraphTraversal {return g.V().Properties("name", "age", nil).Value()}}, 
     "g_V_valuesXname_age_nullX": {func(g *gremlingo.GraphTraversalSource, p map[string]interface{}) *gremlingo.GraphTraversal {return g.V().Values("name", "age", nil)}}, 
-    "g_injectX__feature___test__nullX_rTrim": {func(g *gremlingo.GraphTraversalSource, p map[string]interface{}) *gremlingo.GraphTraversal {return g.Inject("feature  ", "one test ", nil, "", " ").RTrim()}}, 
+    "g_injectX__feature___test__nullX_rTrim": {func(g *gremlingo.GraphTraversalSource, p map[string]interface{}) *gremlingo.GraphTraversal {return g.Inject("feature  ", "one test ", nil, "", " ", "\u3000abc", "abc\u3000", "\u3000abc\u3000", "\u3000\u3000").RTrim()}}, 
+    "g_injectX__feature___test__nullX_rTrimXlocalX": {func(g *gremlingo.GraphTraversalSource, p map[string]interface{}) *gremlingo.GraphTraversal {return g.Inject([]interface{}{"  feature  ", " one test ", nil, "", " ", "\u3000abc", "abc\u3000", "\u3000abc\u3000", "\u3000\u3000"}).RTrim(gremlingo.Scope.Local)}}, 
     "g_injectX__feature__X_rTrim": {func(g *gremlingo.GraphTraversalSource, p map[string]interface{}) *gremlingo.GraphTraversal {return g.Inject("  feature  ").RTrim()}}, 
     "g_injectXListXa_bXX_rTrim": {func(g *gremlingo.GraphTraversalSource, p map[string]interface{}) *gremlingo.GraphTraversal {return g.Inject(p["xx1"]).RTrim()}}, 
     "g_injectXListX1_2XX_rTrimXlocalX": {func(g *gremlingo.GraphTraversalSource, p map[string]interface{}) *gremlingo.GraphTraversal {return g.Inject(p["xx1"]).RTrim(gremlingo.Scope.Local)}}, 
@@ -1196,7 +1198,8 @@ var translationMap = map[string][]func(g *gremlingo.GraphTraversalSource, p map[
     "g_V_valuesXnameX_toUpper": {func(g *gremlingo.GraphTraversalSource, p map[string]interface{}) *gremlingo.GraphTraversal {return g.V().Values("name").ToUpper()}}, 
     "g_V_valuesXnameX_toUpperXlocalX": {func(g *gremlingo.GraphTraversalSource, p map[string]interface{}) *gremlingo.GraphTraversal {return g.V().Values("name").ToUpper(gremlingo.Scope.Local)}}, 
     "g_V_valuesXnameX_order_fold_toUpperXlocalX": {func(g *gremlingo.GraphTraversalSource, p map[string]interface{}) *gremlingo.GraphTraversal {return g.V().Values("name").Order().Fold().ToUpper(gremlingo.Scope.Local)}}, 
-    "g_injectX__feature___test__nullX_trim": {func(g *gremlingo.GraphTraversalSource, p map[string]interface{}) *gremlingo.GraphTraversal {return g.Inject("  feature  ", " one test ", nil, "", " ").Trim()}}, 
+    "g_injectX__feature___test__nullX_trim": {func(g *gremlingo.GraphTraversalSource, p map[string]interface{}) *gremlingo.GraphTraversal {return g.Inject("  feature  ", " one test ", nil, "", " ", "\u3000abc", "abc\u3000", "\u3000abc\u3000", "\u3000\u3000").Trim()}}, 
+    "g_injectX__feature___test__nullX_trimXlocalX": {func(g *gremlingo.GraphTraversalSource, p map[string]interface{}) *gremlingo.GraphTraversal {return g.Inject([]interface{}{"  feature  ", " one test ", nil, "", " ", "\u3000abc", "abc\u3000", "\u3000abc\u3000", "\u3000\u3000"}).Trim(gremlingo.Scope.Local)}}, 
     "g_injectXListXa_bXX_trim": {func(g *gremlingo.GraphTraversalSource, p map[string]interface{}) *gremlingo.GraphTraversal {return g.Inject(p["xx1"]).Trim()}}, 
     "g_injectXListX1_2XX_trimXlocalX": {func(g *gremlingo.GraphTraversalSource, p map[string]interface{}) *gremlingo.GraphTraversal {return g.Inject(p["xx1"]).Trim(gremlingo.Scope.Local)}}, 
     "g_V_valuesXnameX_trim": {func(g *gremlingo.GraphTraversalSource, p map[string]interface{}) *gremlingo.GraphTraversal {return g.AddV("person").Property("name", " marko ").Property("age", 29).As("marko").AddV("person").Property("name", "  vadas  ").Property("age", 27).As("vadas").AddV("software").Property("name", "  lop").Property("lang", "java").As("lop").AddV("person").Property("name", "josh  ").Property("age", 32).As("josh").AddV("software").Property("name", "   ripple   ").Property("lang", "java").As("ripple").AddV("person").Property("name", "peter").Property("age", 35).As("peter").AddE("knows").From("marko").To("vadas").Property("weight", 0.5).AddE("knows").From("marko").To("josh").Property("weight", 1.0).AddE("created").From("marko").To("lop").Property("weight", 0.4).AddE("created").From("josh").To("ripple").Property("weight", 1.0).AddE("created").From("josh").To("lop").Property("weight", 0.4).AddE("created").From("peter").To("lop").Property("weight", 0.2)}, func(g *gremlingo.GraphTraversalSource, p map[string]interface{}) *gremlingo.GraphTraversal {return g.V().Values("name").Trim()}},