ICU4C: fixing some date/time format problems (#223)

* Fix skeleton parsing in CPP date/time format. Update sampling, too

* Fixing some CPP date time failures

* Updating tools to understand failures

Author: sven-oly

executors/cpp/datetime_fmt.cpp

@@ -132,23 +132,26 @@ const string test_datetime_fmt(json_object *json_in) {
       timezone_str = json_object_get_string(option_item);
       UnicodeString u_tz(timezone_str.c_str());
       tz = TimeZone::createTimeZone(u_tz);
-      cout << "# Created timezone " << tz << " for " << timezone_str << endl;
+      // cout << "# Created timezone " << tz << " for " << timezone_str << endl;
     }
   }
 
   json_object *date_skeleton_obj =
-      json_object_object_get(json_in, "datetime_skeleton");
+      json_object_object_get(json_in, "skeleton");
   if (date_skeleton_obj) {
     // Data specifies a date time skeleton. Make a formatter based on this.
     string skeleton_string = json_object_get_string(date_skeleton_obj);
 
+    cout << "# Skeleton = " << skeleton_string << endl;
     UnicodeString u_skeleton(skeleton_string.c_str());
     if (cal) {
+      cout << "  Cal defined " << endl;
       df = DateFormat::createInstanceForSkeleton(cal,
                                                  u_skeleton,
                                                  displayLocale,
                                                  status);
     } else {
+      cout << "  NO CAL DEFINED " << endl;
       df = DateFormat::createInstanceForSkeleton(u_skeleton,
                                                  displayLocale,
                                                  status);
@@ -163,6 +166,12 @@ const string test_datetime_fmt(json_object *json_in) {
     }
   } else {
     // Create default formatter
+    cout << "# createDateTimeInstance: dstyle " <<
+        date_style << ", tstyle " <<
+        time_style << " locale " <<
+        locale_string <<
+        endl;
+
     df = DateFormat::createDateTimeInstance(
         date_style,
         time_style,
@@ -180,10 +189,7 @@ const string test_datetime_fmt(json_object *json_in) {
   }
 
   if (tz) {
-    cout <<
-        "# Setting TZ " <<
-        tz <<
-        endl;
+    // cout << "# Setting TZ " << tz << endl;
     df->setTimeZone(*tz);
   }
 

executors/test_strings.txt

@@ -313,3 +313,18 @@
 {"test_type": "rdt_fmt", "label": "100", "unit":"day" "day", "count": 1, "locale": "en-US", "options": {}}  
 {"test_type": "rdt_fmt", "label": "100", "unit":"day", "count": 1, "locale": "es", "options": {}}
 
+
+
+// Testing skeleton options
+{"test_type": "datetime_fmt", "label":"10","input_string":"2001-09-09T01:46:40.000Z","locale":"en-US","options":{"calendar":"gregory","numberingSystem":"latn"}}
+    {"test_type": "datetime_fmt", "label":"10","input_string":"2001-09-09T01:46:40.000Z","locale":"en-US","options":{"calendar":"gregory","numberingSystem":"latn"}}
+
+{"test_type": "datetime_fmt", "label":"101","input_string":"2000-01-01T21:00:00.000Z","skeleton":"hh","locale":"en-US","options":{"hour":"2-digit","calendar":"gregory","timeZone":"America/Los_Angeles","numberingSystem":"latn"}}
+{"test_type": "datetime_fmt", "label":"101","input_string":"2000-01-01T21:00:00.000Z","skeleton":"h","locale":"en-US","options":{"hour":"2-digit","calendar":"gregory","timeZone":"America/Los_Angeles","numberingSystem":"latn"}}
+{"test_type": "datetime_fmt", "label":"101","input_string":"2000-01-01T21:00:00.000Z","skeleton":"H","locale":"en-US","options":{"hour":"2-digit","calendar":"gregory","timeZone":"America/Los_Angeles","numberingSystem":"latn"}}
+{"test_type": "datetime_fmt", "label":"101","input_string":"2000-01-01T21:00:00.000Z","skeleton":"hhh","locale":"en-US","options":{"hour":"2-digit","calendar":"gregory","timeZone":"America/Los_Angeles","numberingSystem":"latn"}}
+{"test_type": "datetime_fmt", "label":"1265","input_string":"2024-03-17T07:00:00.000Z","skeleton":"hhhm","locale":"en-US","options":{"hour":"2-digit","minute":"numeric","calendar":"chinese","timeZone":"Australia/Brisbane","numberingSystem":"latn"}}
+
+{"test_type": "datetime_fmt", "label":"3812","input_string":"1969-07-16T07:00:00.000Z","skeleton":"HHm","locale":"en-GB","options":{"hour":"2-digit","minute":"numeric","calendar":"buddhist","timeZone":"Australia/Brisbane","numberingSystem":"latn"}}
+
+{"test_type": "datetime_fmt", "label":"3812","input_string":"1969-07-16T07:00:00.000Z","locale":"en-GB","options":{"dateStyle": "short", "timeStyle": "short", "hour":"2-digit","minute":"numeric","calendar":"buddhist","timeZone":"Australia/Brisbane","numberingSystem":"latn"}}

testgen/generators/datetime_fmt.py

@@ -32,7 +32,8 @@ def process_test_data(self):
 
         # Set up Node version and call the generator
         nvm_version = icu_nvm_versions[self.icu_version]
-        generate_command = 'source ~/.nvm/nvm.sh; nvm install %s; nvm use %s; node generators/datetime_gen.js' % (nvm_version, nvm_version)
+        generate_command = 'source ~/.nvm/nvm.sh; nvm install %s; nvm use %s; node generators/datetime_gen.js %s %s' % (
+            nvm_version, nvm_version, '-run_limit', self.run_limit)
 
         logging.info('Running this command: %s', generate_command)
         result = result = subprocess.run(generate_command, shell=True)

testgen/generators/datetime_gen.js

@@ -100,8 +100,10 @@ const dates = [
   new Date('Mar 17, 2024'),
   new Date('AD 1'),
   new Date('AD 0, 13:00'),
+  //new Date('1066, December 16, 7:17'),  // Expect Sunday
   new Date('1066, December 16, 7:17'),
-  new Date('1454, May 29, 16:47'),
+  // new Date('1454, May 29, 16:47'),  // Expect Monday
+  new Date('1754, May 29, 16:47'),
   new Date('BCE 753, April 21'),
   new Date('1969, July 16'),
   new Date(0),
@@ -188,7 +190,7 @@ function optionsToSkeleton(options) {
   return skeleton_array.join('');
 }
 
-function generateAll() {
+function generateAll(run_limit) {
 
   let test_obj = {
     'Test scenario': 'datetime_fmt',
@@ -256,7 +258,7 @@ function generateAll() {
         try {
           formatter = new Intl.DateTimeFormat(locale, all_options);
         } catch (error) {
-          console.log(error, ' with locale ',
+          console.error(error, ' with locale ',
                       locale, ' and options: ', all_options);
           continue;
         }
@@ -268,7 +270,7 @@ function generateAll() {
             const parts = formatter.formatToParts(d);
             result = parts.map((x) => x.value).join("");
           } catch (error) {
-            console.log('FORMATTER CREATION FAILS! ', error);
+            console.error('FORMATTER CREATION FAILS! ', error);
           }
           // format this date
           // get the milliseconds
@@ -297,7 +299,7 @@ function generateAll() {
           }
 
           if (debug) {
-            console.log("TEST CASE :", test_case);
+            console.debug("TEST CASE :", test_case);
           }
           test_cases.push(test_case);
 
@@ -309,7 +311,7 @@ function generateAll() {
               console.log('   expected = ', result);
             }
           } catch (error) {
-            console.log('!!! error ', error, ' in label ', label_num,
+            console.error('!!! error ', error, ' in label ', label_num,
                         ' for date = ', d);
           }
           label_num ++;
@@ -318,25 +320,52 @@ function generateAll() {
     }
   }
 
-
-  console.log('Number of date/time tests generated for ',
+  console.log('Total date/time tests generated for ',
               process.versions.icu, ': ', label_num);
 
-  test_obj['tests'] = test_cases;
+  test_obj['tests'] = sample_tests(test_cases, run_limit);
   try {
     fs.writeFileSync('datetime_fmt_test.json', JSON.stringify(test_obj, null, 2));
     // file written successfully
   } catch (err) {
     console.error(err);
   }
 
-  verify_obj['verifications'] = verify_cases;
+  console.log('Number of date/time sampled tests ',
+              process.versions.icu, ': ', test_obj['tests'] .length);
+
+  verify_obj['verifications'] = sample_tests(verify_cases, run_limit);
   try {
     fs.writeFileSync('datetime_fmt_verify.json', JSON.stringify(verify_obj, null, 2));
     // file written successfully
   } catch (err) {
     console.error(err);
   }
 }
+
+function sample_tests(all_tests, run_limit) {
+  // Gets a sampling of the data based on total and the expected number.
+
+  if (run_limit < 0 || all_tests.length <= run_limit) {
+    return all_tests;
+  }
+
+  let size_all = all_tests.length;
+  let increment = Math.floor(size_all / run_limit);
+  let samples = [];
+  for (let index = 0; index < size_all; index += increment) {
+    samples.push(all_tests[index]);
+  }
+  return samples;
+}
+
 /* Call the generator */
-generateAll();
+let run_limit = -1;
+if (process.argv.length >= 5) {
+  if (process.argv[3] == '-run_limit') {
+    run_limit = Number(process.argv[4]);
+  }
+}
+
+// Call the generator
+generateAll(run_limit);

testgen/generators/relativedatetime_fmt.py

@@ -35,7 +35,7 @@ def process_test_data(self):
         generate_command = 'source ~/.nvm/nvm.sh; nvm install %s; nvm use %s; node generators/rdt_fmt_gen.js' % (nvm_version, nvm_version)
 
         logging.info('Running this command: %s', generate_command)
-        result = result = subprocess.run(generate_command, shell=True)
+        result = subprocess.run(generate_command, shell=True)
 
         # Move results to the right directory
         mv_command = 'mv rdt_fmt*.json %s' % self.icu_version

verifier/detail_template.html

@@ -33,7 +33,11 @@
         cursor: pointer;
         font-size:20px;
     }
-
+    
+    #diff_area {
+        font-size: 24px;
+        font-color: blue;
+    }
 h1, h2, h3, p {
   font-family: Arial, Helvetica, sans-serif;
 }
@@ -350,7 +354,7 @@
                         } else {
                             output = item[key];
                         }
-                        html.push('<td onclick="captureInputDataOnClick(this);">' + output +'</td>');
+                        html.push('<td onclick="captureInputDataOnClick(this);" onmouseover="hoverDiffText(this);">' + output +'</td>');
                     }
                 }
                 html.push("</tr>");
@@ -609,7 +613,31 @@
         navigator.clipboard.writeText(output);
     }
 
+    // On hover, show the differen between expected and actual result.
+    function hoverDiffText(element) {
+      // First, get the row of this item.
+      const row = element.parentNode;
+      const text1 = row.children[1].innerHTML;
+      const text2 = row.children[2].innerHTML;
+      if (text1 == text2) {
+        return;
+      }
+      const dmp = new diff_match_patch();
+      const diff_text = dmp.diff_main(text1, text2);
+      dmp.diff_cleanupSemantic(diff_text);
+      // And show the difference nicely.
+      const ds = dmp.diff_prettyHtml(diff_text);
+      const diff_area = document.getElementById("diff_area");
+      diff_area.innerHTML = ds;
+      }
   </script>
+
+  <!-- diff-match-patch code from
+       https://github.com/google/diff-match-patch/wiki/Language:-JavaScript
+       Intended to use for differences between expected and actual result.
+    -->
+    <script src="../../../diff_match_patch.js"></script>
+  
  </head>
  <body>
     <h1>Verification report: $test_type on $library_name</h1>
@@ -628,6 +656,10 @@ <h2>$platform_info</h2>
         <h2>Test details</h2>
         <p>$test_environment</p>
       </div>
+      <div bp="4">
+        <h2>Acknowledgements</h2>
+        <p><a href='https://github.com/google/diff-match-patch/wiki/Language:-JavaScript'>diff-match-patch</a></p>
+      </div>
     </div>
     <br>
     <div id='test_result_details' class='test_results'>
@@ -660,6 +692,7 @@ <h2>Test details</h2>
 
       <details>
         <summary>Failing tests <span id='fail_count'>($failing_tests)</span></summary>
+        <div id="diff_area"><p></p></div>
         <div id="fail-pagination-container"></div>
         <div id="fail-data-container"></div>
         <div id='testFailuresCharacterized'>