updating fragment verification

Signed-off-by: Mike Wilson <[email protected]>

src/main/cpp/src/parse_uri.cu

@@ -486,8 +486,11 @@ uri_parts __device__ validate_uri(const char* str, int len)
 
   // anything after the hash is part of the fragment and ignored for this part
   if (hash >= 0) {
-    ret.fragment = {str + hash, len - hash};
-    if (!validate_fragment(ret.fragment)) { ret.fragment = {}; }
+    ret.fragment = {str + hash + 1, len - hash - 1};
+    if (!validate_fragment(ret.fragment)) {
+      ret.valid = false;
+      return ret;
+    }
 
     len = hash;
 

src/main/cpp/tests/parse_uri.cpp

@@ -83,9 +83,10 @@ TEST_F(ParseURIProtocolTests, SparkEdges)
      "http://[fe80::7:8%eth0]",
      "http://[fe80::7:8%1]",
      "http://foo.bar/abc/\\\\\\http://foo.bar/abc.gif\\\\\\",
-     "b.oscars.org:8100/servlet/"
+     "www.nvidia.com:8100/servlet/"
      "impc.DisplayCredits?primekey_in=2000041100:05:14115240636",
-     "https://j.mp/2Ru15Ss "});
+     "https://nvidia.com/2Ru15Ss ",
+     "http://www.nvidia.com/plugins//##"});
 
   auto result = spark_rapids_jni::parse_uri_to_protocol(cudf::strings_column_view{col});
 
@@ -115,9 +116,15 @@ TEST_F(ParseURIProtocolTests, SparkEdges)
      "http",
      "http",
      "",
-     "b.oscars.org",
+     "www.nvidia.com",
+     "",
      ""},
-    {1, 1, 1, 1, 0, 1, 0, 0, 0, 1, 0, 1, 1, 1, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 0});
+    {1, 1, 1, 1, 0, 1, 0, 0, 0, 1, 0, 1, 1, 1, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 0, 0});
+
+  printf("expected:\n");
+  cudf::test::print(expected);
+  printf("result:\n");
+  cudf::test::print(result->view());
 
   CUDF_TEST_EXPECT_COLUMNS_EQUIVALENT(result->view(), expected);
 }

src/test/java/com/nvidia/spark/rapids/jni/ParseURITest.java

@@ -49,7 +49,7 @@ void buildExpectedAndRun(String[] testData) {
   @Test
   void parseURIToProtocolSparkTest() {
     String[] testData = {
-      "https://nvidia.com/https&#://nvidia.com",
+      /*"https://nvidia.com/https&#://nvidia.com",
       "https://http://www.nvidia.com",
       "filesystemmagicthing://bob.yaml",
       "nvidia.com:8080",
@@ -69,15 +69,18 @@ void parseURIToProtocolSparkTest() {
       "https://www.nvidia.com:8080/q",
       "https://www.nvidia.com#8080",
       "file://path/to/cool/file",
-      "http//www.nvidia.com/q",
+      "http//www.nvidia.com/q",*/
       "http://?",
-      "http://#",
+      "http://#",/*
       "http://??",
       "http://??/",
       "http://user:pass@host/file;param?query;p2",
       "http://foo.bar/abc/\\\\\\http://foo.bar/abc.gif\\\\\\",
-      "b.oscars.org:8100/servlet/impc.DisplayCredits?primekey_in=2000041100:05:14115240636",
-      "https://j.mp/2Ru15Ss ",
+      "nvidia.com:8100/servlet/impc.DisplayCredits?primekey_in=2000041100:05:14115240636",
+      "https://nvidia.com/2Ru15Ss ",
+      "http://www.nvidia.com/xmlrpc//##",
+      "www.nvidia.com:8080/expert/sciPublication.jsp?ExpertId=1746&lenList=all",
+      "www.nvidia.com:8080/hrcxtf/view?docId=ead/00073.xml&query=T.%20E.%20Lawrence&query-join=and",*/
       "",
       null};