Merge pull request #542 from johnjaylward/fixXmlHexEntityParse

Fix xml hex entity parse

src/main/java/org/json/XMLTokener.java

@@ -167,7 +167,7 @@ public class XMLTokener extends JSONTokener {
         // if our entity is an encoded unicode point, parse it.
         if (e.charAt(0) == '#') {
             int cp;
-            if (e.charAt(1) == 'x') {
+            if (e.charAt(1) == 'x' || e.charAt(1) == 'X') {
                 // hex encoded unicode
                 cp = Integer.parseInt(e.substring(2), 16);
             } else {

src/test/java/org/json/junit/Util.java

@@ -103,20 +103,18 @@ public class Util {
              */
             if (!(value instanceof Number && expectedValue instanceof Number)) {
                 // Non-Number and non-matching types
-                assertTrue("object types should be equal for actual: "+
+                assertEquals("object types should be equal ",
-                    value.toString()+" ("+
+                    expectedValue.getClass().toString(),
-                    value.getClass().toString()+") expected: "+
+                    value.getClass().toString()
-                    expectedValue.toString()+" ("+
+                );
-                    expectedValue.getClass().toString()+")",
-                    value.getClass().toString().equals(
-                            expectedValue.getClass().toString()));
             }
             /**
              * Same types or both Numbers, compare by toString()
              */
-            assertTrue("string values should be equal for actual: "+
+            assertEquals("values should be equal",
-                value.toString()+" expected: "+expectedValue.toString(),
+                expectedValue.toString(),
-                value.toString().equals(expectedValue.toString()));
+                value.toString()
+            );
         }
     }
 }

src/test/java/org/json/junit/XMLTest.java

@@ -34,12 +34,15 @@ import java.io.File;
 import java.io.FileReader;
 import java.io.FileWriter;
 import java.io.IOException;
+import java.io.InputStream;
+import java.io.InputStreamReader;
 import java.io.Reader;
 import java.io.StringReader;
 
 import org.json.JSONArray;
 import org.json.JSONException;
 import org.json.JSONObject;
+import org.json.JSONTokener;
 import org.json.XML;
 import org.json.XMLParserConfiguration;
 import org.junit.Rule;
@@ -898,4 +901,57 @@ public class XMLTest {
         final JSONObject json = XML.toJSONObject(originalXml, new XMLParserConfiguration());
         assertEquals(expectedJsonString, json.toString());
     }
+
+    /**
+     * Tests to verify that supported escapes in XML are converted to actual values.
+     */
+    @Test
+    public void testIssue537CaseSensitiveHexEscapeMinimal(){
+        String xmlStr = 
+            "<?xml version=\"1.0\" encoding=\"UTF-8\"?>\n"+
+            "<root>Neutrophils.Hypersegmented &#X7C; Bld-Ser-Plas</root>";
+        String expectedStr = 
+            "{\"root\":\"Neutrophils.Hypersegmented | Bld-Ser-Plas\"}";
+        JSONObject xmlJSONObj = XML.toJSONObject(xmlStr, true);
+        JSONObject expected = new JSONObject(expectedStr);
+        Util.compareActualVsExpectedJsonObjects(xmlJSONObj, expected);
+    }
+
+    /**
+     * Tests to verify that supported escapes in XML are converted to actual values.
+     */
+    @Test
+    public void testIssue537CaseSensitiveHexEscapeFullFile(){
+        try {
+            try(
+                    InputStream xmlStream = XMLTest.class.getClassLoader().getResourceAsStream("Issue537.xml");
+                    Reader xmlReader = new InputStreamReader(xmlStream);
+                ){
+                JSONObject actual = XML.toJSONObject(xmlReader, true);
+                try(
+                        InputStream jsonStream = XMLTest.class.getClassLoader().getResourceAsStream("Issue537.json");
+                    ){
+                    final JSONObject expected = new JSONObject(new JSONTokener(jsonStream));
+                    Util.compareActualVsExpectedJsonObjects(actual,expected);
+                }
+            }
+        } catch (IOException e) {
+            fail("file writer error: " +e.getMessage());
+        }
+    }
+
+    /**
+     * Tests to verify that supported escapes in XML are converted to actual values.
+     */
+    @Test
+    public void testIssue537CaseSensitiveHexUnEscapeDirect(){
+        String origStr = 
+            "Neutrophils.Hypersegmented &#X7C; Bld-Ser-Plas";
+        String expectedStr = 
+            "Neutrophils.Hypersegmented | Bld-Ser-Plas";
+        String actualStr = XML.unescape(origStr);
+        
+        assertEquals("Case insensitive Entity unescape",  expectedStr, actualStr);
+    }
+    
 }

src/test/resources/Issue537.json

@@ -0,0 +1,189 @@
+{
+ "clinical_study": {
+  "brief_summary": {
+   "textblock": "CLEAR SYNERGY is an international multi center 2x2 randomized placebo controlled trial of"
+  },
+  "brief_title": "CLEAR SYNERGY Neutrophil Substudy",
+  "overall_status": "Recruiting",
+  "eligibility": {
+   "study_pop": {
+    "textblock": "Patients who are randomized to the drug RCT portion of the CLEAR SYNERGY (OASIS 9) trial"
+   },
+   "minimum_age": "19 Years",
+   "sampling_method": "Non-Probability Sample",
+   "gender": "All",
+   "criteria": {
+    "textblock": "Inclusion Criteria:"
+   },
+   "healthy_volunteers": "No",
+   "maximum_age": "110 Years"
+  },
+  "number_of_groups": "2",
+  "source": "NYU Langone Health",
+  "location_countries": {
+   "country": "United States"
+  },
+  "study_design_info": {
+   "time_perspective": "Prospective",
+   "observational_model": "Other"
+  },
+  "last_update_submitted_qc": "September 10, 2019",
+  "intervention_browse": {
+   "mesh_term": "Colchicine"
+  },
+  "official_title": "Studies on the Effects of Colchicine on Neutrophil Biology in Acute Myocardial Infarction: A Substudy of the CLEAR SYNERGY (OASIS 9) Trial",
+  "primary_completion_date": {
+   "type": "Anticipated",
+   "content": "February 1, 2021"
+  },
+  "sponsors": {
+   "lead_sponsor": {
+    "agency_class": "Other",
+    "agency": "NYU Langone Health"
+   },
+   "collaborator": [
+    {
+     "agency_class": "Other",
+     "agency": "Population Health Research Institute"
+    },
+    {
+     "agency_class": "NIH",
+     "agency": "National Heart, Lung, and Blood Institute (NHLBI)"
+    }
+   ]
+  },
+  "overall_official": {
+   "role": "Principal Investigator",
+   "affiliation": "NYU School of Medicine",
+   "last_name": "Binita Shah, MD"
+  },
+  "overall_contact_backup": {
+   "last_name": "Binita Shah, MD"
+  },
+  "condition_browse": {
+   "mesh_term": [
+    "Myocardial Infarction",
+    "ST Elevation Myocardial Infarction",
+    "Infarction"
+   ]
+  },
+  "overall_contact": {
+   "phone": "646-501-9648",
+   "last_name": "Fatmira Curovic",
+   "email": "fatmira.curovic@nyumc.org"
+  },
+  "responsible_party": {
+   "responsible_party_type": "Principal Investigator",
+   "investigator_title": "Assistant Professor of Medicine",
+   "investigator_full_name": "Binita Shah",
+   "investigator_affiliation": "NYU Langone Health"
+  },
+  "study_first_submitted_qc": "March 12, 2019",
+  "start_date": {
+   "type": "Actual",
+   "content": "March 4, 2019"
+  },
+  "has_expanded_access": "No",
+  "study_first_posted": {
+   "type": "Actual",
+   "content": "March 14, 2019"
+  },
+  "arm_group": [
+   {
+    "arm_group_label": "Colchicine"
+   },
+   {
+    "arm_group_label": "Placebo"
+   }
+  ],
+  "primary_outcome": {
+   "measure": "soluble L-selectin",
+   "time_frame": "between baseline and 3 months",
+   "description": "Change in soluble L-selectin between baseline and 3 mo after STEMI in the placebo vs. colchicine groups."
+  },
+  "secondary_outcome": [
+   {
+    "measure": "Other soluble markers of neutrophil activity",
+    "time_frame": "between baseline and 3 months",
+    "description": "Other markers of neutrophil activity will be evaluated at baseline and 3 months after STEMI (myeloperoxidase, matrix metalloproteinase-9, neutrophil gelatinase-associated lipocalin, neutrophil elastase, intercellular/vascular cellular adhesion molecules)"
+   },
+   {
+    "measure": "Markers of systemic inflammation",
+    "time_frame": "between baseline and 3 months",
+    "description": "Markers of systemic inflammation will be evaluated at baseline and 3 months after STEMI (high sensitive CRP, IL-1β)"
+   },
+   {
+    "measure": "Neutrophil-driven responses that may further propagate injury",
+    "time_frame": "between baseline and 3 months",
+    "description": "Neutrophil-driven responses that may further propagate injury will be evaluated at baseline and 3 months after STEMI (neutrophil extracellular traps, neutrophil-derived microparticles)"
+   }
+  ],
+  "oversight_info": {
+   "is_fda_regulated_drug": "No",
+   "is_fda_regulated_device": "No",
+   "has_dmc": "No"
+  },
+  "last_update_posted": {
+   "type": "Actual",
+   "content": "September 12, 2019"
+  },
+  "id_info": {
+   "nct_id": "NCT03874338",
+   "org_study_id": "18-01323",
+   "secondary_id": "1R01HL146206"
+  },
+  "enrollment": {
+   "type": "Anticipated",
+   "content": "670"
+  },
+  "study_first_submitted": "March 12, 2019",
+  "condition": [
+   "Neutrophils.Hypersegmented | Bld-Ser-Plas",
+   "STEMI - ST Elevation Myocardial Infarction"
+  ],
+  "study_type": "Observational",
+  "required_header": {
+   "download_date": "ClinicalTrials.gov processed this data on July 19, 2020",
+   "link_text": "Link to the current ClinicalTrials.gov record.",
+   "url": "https://clinicaltrials.gov/show/NCT03874338"
+  },
+  "last_update_submitted": "September 10, 2019",
+  "completion_date": {
+   "type": "Anticipated",
+   "content": "February 1, 2022"
+  },
+  "location": {
+   "contact": {
+    "phone": "646-501-9648",
+    "last_name": "Fatmira Curovic",
+    "email": "fatmira.curovic@nyumc.org"
+   },
+   "facility": {
+    "address": {
+     "zip": "10016",
+     "country": "United States",
+     "city": "New York",
+     "state": "New York"
+    },
+    "name": "NYU School of Medicine"
+   },
+   "status": "Recruiting",
+   "contact_backup": {
+    "last_name": "Binita Shah, MD"
+   }
+  },
+  "intervention": {
+   "intervention_type": "Drug",
+   "arm_group_label": [
+    "Colchicine",
+    "Placebo"
+   ],
+   "description": "Participants in the main CLEAR SYNERGY trial are randomized to colchicine/spironolactone versus placebo in a 2x2 factorial design. The substudy is interested in the evaluation of biospecimens obtained from patients in the colchicine vs placebo group.",
+   "intervention_name": "Colchicine Pill"
+  },
+  "patient_data": {
+   "sharing_ipd": "No"
+  },
+  "verification_date": "September 2019"
+ }
+}

src/test/resources/Issue537.xml

@@ -0,0 +1,169 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<clinical_study>
+  <!-- This xml conforms to an XML Schema at:
+    https://clinicaltrials.gov/ct2/html/images/info/public.xsd -->
+  <required_header>
+    <download_date>ClinicalTrials.gov processed this data on July 19, 2020</download_date>
+    <link_text>Link to the current ClinicalTrials.gov record.</link_text>
+    <url>https://clinicaltrials.gov/show/NCT03874338</url>
+  </required_header>
+  <id_info>
+    <org_study_id>18-01323</org_study_id>
+    <secondary_id>1R01HL146206</secondary_id>
+    <nct_id>NCT03874338</nct_id>
+  </id_info>
+  <brief_title>CLEAR SYNERGY Neutrophil Substudy</brief_title>
+  <official_title>Studies on the Effects of Colchicine on Neutrophil Biology in Acute Myocardial Infarction: A Substudy of the CLEAR SYNERGY (OASIS 9) Trial</official_title>
+  <sponsors>
+    <lead_sponsor>
+      <agency>NYU Langone Health</agency>
+      <agency_class>Other</agency_class>
+    </lead_sponsor>
+    <collaborator>
+      <agency>Population Health Research Institute</agency>
+      <agency_class>Other</agency_class>
+    </collaborator>
+    <collaborator>
+      <agency>National Heart, Lung, and Blood Institute (NHLBI)</agency>
+      <agency_class>NIH</agency_class>
+    </collaborator>
+  </sponsors>
+  <source>NYU Langone Health</source>
+  <oversight_info>
+    <has_dmc>No</has_dmc>
+    <is_fda_regulated_drug>No</is_fda_regulated_drug>
+    <is_fda_regulated_device>No</is_fda_regulated_device>
+  </oversight_info>
+  <brief_summary>
+    <textblock>
+      CLEAR SYNERGY is an international multi center 2x2 randomized placebo controlled trial of
+    </textblock>
+  </brief_summary>
+  <overall_status>Recruiting</overall_status>
+  <start_date type="Actual">March 4, 2019</start_date>
+  <completion_date type="Anticipated">February 1, 2022</completion_date>
+  <primary_completion_date type="Anticipated">February 1, 2021</primary_completion_date>
+  <study_type>Observational</study_type>
+  <has_expanded_access>No</has_expanded_access>
+  <study_design_info>
+    <observational_model>Other</observational_model>
+    <time_perspective>Prospective</time_perspective>
+  </study_design_info>
+  <primary_outcome>
+    <measure>soluble L-selectin</measure>
+    <time_frame>between baseline and 3 months</time_frame>
+    <description>Change in soluble L-selectin between baseline and 3 mo after STEMI in the placebo vs. colchicine groups.</description>
+  </primary_outcome>
+  <secondary_outcome>
+    <measure>Other soluble markers of neutrophil activity</measure>
+    <time_frame>between baseline and 3 months</time_frame>
+    <description>Other markers of neutrophil activity will be evaluated at baseline and 3 months after STEMI (myeloperoxidase, matrix metalloproteinase-9, neutrophil gelatinase-associated lipocalin, neutrophil elastase, intercellular/vascular cellular adhesion molecules)</description>
+  </secondary_outcome>
+  <secondary_outcome>
+    <measure>Markers of systemic inflammation</measure>
+    <time_frame>between baseline and 3 months</time_frame>
+    <description>Markers of systemic inflammation will be evaluated at baseline and 3 months after STEMI (high sensitive CRP, IL-1β)</description>
+  </secondary_outcome>
+  <secondary_outcome>
+    <measure>Neutrophil-driven responses that may further propagate injury</measure>
+    <time_frame>between baseline and 3 months</time_frame>
+    <description>Neutrophil-driven responses that may further propagate injury will be evaluated at baseline and 3 months after STEMI (neutrophil extracellular traps, neutrophil-derived microparticles)</description>
+  </secondary_outcome>
+  <number_of_groups>2</number_of_groups>
+  <enrollment type="Anticipated">670</enrollment>
+  <condition>Neutrophils.Hypersegmented &#X7C; Bld-Ser-Plas</condition>
+  <condition>STEMI - ST Elevation Myocardial Infarction</condition>
+  <arm_group>
+    <arm_group_label>Colchicine</arm_group_label>
+  </arm_group>
+  <arm_group>
+    <arm_group_label>Placebo</arm_group_label>
+  </arm_group>
+  <intervention>
+    <intervention_type>Drug</intervention_type>
+    <intervention_name>Colchicine Pill</intervention_name>
+    <description>Participants in the main CLEAR SYNERGY trial are randomized to colchicine/spironolactone versus placebo in a 2x2 factorial design. The substudy is interested in the evaluation of biospecimens obtained from patients in the colchicine vs placebo group.</description>
+    <arm_group_label>Colchicine</arm_group_label>
+    <arm_group_label>Placebo</arm_group_label>
+  </intervention>
+  <eligibility>
+    <study_pop>
+      <textblock>
+        Patients who are randomized to the drug RCT portion of the CLEAR SYNERGY (OASIS 9) trial
+      </textblock>
+    </study_pop>
+    <sampling_method>Non-Probability Sample</sampling_method>
+    <criteria>
+      <textblock>
+        Inclusion Criteria:
+      </textblock>
+    </criteria>
+    <gender>All</gender>
+    <minimum_age>19 Years</minimum_age>
+    <maximum_age>110 Years</maximum_age>
+    <healthy_volunteers>No</healthy_volunteers>
+  </eligibility>
+  <overall_official>
+    <last_name>Binita Shah, MD</last_name>
+    <role>Principal Investigator</role>
+    <affiliation>NYU School of Medicine</affiliation>
+  </overall_official>
+  <overall_contact>
+    <last_name>Fatmira Curovic</last_name>
+    <phone>646-501-9648</phone>
+    <email>fatmira.curovic@nyumc.org</email>
+  </overall_contact>
+  <overall_contact_backup>
+    <last_name>Binita Shah, MD</last_name>
+  </overall_contact_backup>
+  <location>
+    <facility>
+      <name>NYU School of Medicine</name>
+      <address>
+        <city>New York</city>
+        <state>New York</state>
+        <zip>10016</zip>
+        <country>United States</country>
+      </address>
+    </facility>
+    <status>Recruiting</status>
+    <contact>
+      <last_name>Fatmira Curovic</last_name>
+      <phone>646-501-9648</phone>
+      <email>fatmira.curovic@nyumc.org</email>
+    </contact>
+    <contact_backup>
+      <last_name>Binita Shah, MD</last_name>
+    </contact_backup>
+  </location>
+  <location_countries>
+    <country>United States</country>
+  </location_countries>
+  <verification_date>September 2019</verification_date>
+  <study_first_submitted>March 12, 2019</study_first_submitted>
+  <study_first_submitted_qc>March 12, 2019</study_first_submitted_qc>
+  <study_first_posted type="Actual">March 14, 2019</study_first_posted>
+  <last_update_submitted>September 10, 2019</last_update_submitted>
+  <last_update_submitted_qc>September 10, 2019</last_update_submitted_qc>
+  <last_update_posted type="Actual">September 12, 2019</last_update_posted>
+  <responsible_party>
+    <responsible_party_type>Principal Investigator</responsible_party_type>
+    <investigator_affiliation>NYU Langone Health</investigator_affiliation>
+    <investigator_full_name>Binita Shah</investigator_full_name>
+    <investigator_title>Assistant Professor of Medicine</investigator_title>
+  </responsible_party>
+  <condition_browse>
+    <!-- CAUTION:  The following MeSH terms are assigned with an imperfect algorithm            -->
+    <mesh_term>Myocardial Infarction</mesh_term>
+    <mesh_term>ST Elevation Myocardial Infarction</mesh_term>
+    <mesh_term>Infarction</mesh_term>
+  </condition_browse>
+  <intervention_browse>
+    <!-- CAUTION:  The following MeSH terms are assigned with an imperfect algorithm            -->
+    <mesh_term>Colchicine</mesh_term>
+  </intervention_browse>
+  <patient_data>
+    <sharing_ipd>No</sharing_ipd>
+  </patient_data>
+  <!-- Results have not yet been posted for this study                                          -->
+</clinical_study>