Change default locale of date mappers to ENGLISH (#112799)

English is not changing between COMPAT and CLDR locale databases, whereas ROOT is

Author: thecoop

docs/reference/mapping/types/date.asciidoc

@@ -125,8 +125,7 @@ The following parameters are accepted by `date` fields:
 `locale`::
 
     The locale to use when parsing dates since months do not have the same names
-    and/or abbreviations in all languages. The default is the
-    https://docs.oracle.com/javase/8/docs/api/java/util/Locale.html#ROOT[`ROOT` locale].
+    and/or abbreviations in all languages. The default is ENGLISH.
 
 <<ignore-malformed,`ignore_malformed`>>::
 

modules/aggregations/build.gradle

@@ -48,4 +48,5 @@ dependencies {
 
 tasks.named("yamlRestCompatTestTransform").configure({ task ->
   task.skipTest("aggregations/date_agg_per_day_of_week/Date aggregartion per day of week", "week-date behaviour has changed")
+  task.skipTest("aggregations/stats_metric_fail_formatting/fail formatting", "locale has changed")
 })

modules/aggregations/src/test/java/org/elasticsearch/aggregations/bucket/histogram/InternalAutoDateHistogramTests.java

@@ -9,7 +9,6 @@
 
 package org.elasticsearch.aggregations.bucket.histogram;
 
-import org.elasticsearch.TransportVersion;
 import org.elasticsearch.TransportVersions;
 import org.elasticsearch.aggregations.bucket.AggregationMultiBucketAggregationTestCase;
 import org.elasticsearch.aggregations.bucket.histogram.AutoDateHistogramAggregationBuilder.RoundingInfo;
@@ -28,7 +27,6 @@
 import org.elasticsearch.search.aggregations.bucket.histogram.DateHistogramInterval;
 import org.elasticsearch.search.aggregations.bucket.histogram.Histogram;
 import org.elasticsearch.test.InternalAggregationTestCase;
-import org.elasticsearch.test.TransportVersionUtils;
 
 import java.io.IOException;
 import java.time.Instant;
@@ -459,33 +457,6 @@ public void testCreateWithReplacementBuckets() {
         assertThat(copy.getInterval(), equalTo(orig.getInterval()));
     }
 
-    public void testSerializationPre830() throws IOException {
-        // we need to test without sub-aggregations, otherwise we need to also update the interval within the inner aggs
-        InternalAutoDateHistogram instance = createTestInstance(
-            randomAlphaOfLengthBetween(3, 7),
-            createTestMetadata(),
-            InternalAggregations.EMPTY
-        );
-        TransportVersion version = TransportVersionUtils.randomVersionBetween(
-            random(),
-            TransportVersions.MINIMUM_COMPATIBLE,
-            TransportVersionUtils.getPreviousVersion(TransportVersions.V_8_3_0)
-        );
-        InternalAutoDateHistogram deserialized = copyInstance(instance, version);
-        assertEquals(1, deserialized.getBucketInnerInterval());
-
-        InternalAutoDateHistogram modified = new InternalAutoDateHistogram(
-            deserialized.getName(),
-            deserialized.getBuckets(),
-            deserialized.getTargetBuckets(),
-            deserialized.getBucketInfo(),
-            deserialized.getFormatter(),
-            deserialized.getMetadata(),
-            instance.getBucketInnerInterval()
-        );
-        assertEqualInstances(instance, modified);
-    }
-
     public void testReadFromPre830() throws IOException {
         byte[] bytes = Base64.getDecoder()
             .decode(

modules/aggregations/src/yamlRestTest/resources/rest-api-spec/test/aggregations/stats_metric_fail_formatting.yml

@@ -30,7 +30,7 @@ setup:
       cluster_features: "gte_v8.15.0"
       reason: fixed in 8.15.0
   - do:
-      catch: /Cannot format stat \[sum\] with format \[DocValueFormat.DateTime\(format\[date_hour_minute_second_millis\] locale\[\], Z, MILLISECONDS\)\]/
+      catch: /Cannot format stat \[sum\] with format \[DocValueFormat.DateTime\(format\[date_hour_minute_second_millis\] locale\[en\], Z, MILLISECONDS\)\]/
       search:
         index: test_date
         body:

server/src/main/java/org/elasticsearch/TransportVersions.java

@@ -234,6 +234,7 @@ static TransportVersion def(int id) {
     public static final TransportVersion RRF_QUERY_REWRITE = def(8_758_00_0);
     public static final TransportVersion SEARCH_FAILURE_STATS = def(8_759_00_0);
     public static final TransportVersion INGEST_GEO_DATABASE_PROVIDERS = def(8_760_00_0);
+    public static final TransportVersion DATE_TIME_DOC_VALUES_LOCALES = def(8_761_00_0);
 
     /*
      * STOP! READ THIS FIRST! No, really,

server/src/main/java/org/elasticsearch/common/util/LocaleUtils.java

@@ -68,16 +68,16 @@ private static Locale parseParts(String[] parts) {
         switch (parts.length) {
             case 3:
                 // lang, country, variant
-                return new Locale(parts[0], parts[1], parts[2]);
+                return Locale.of(parts[0], parts[1], parts[2]);
             case 2:
                 // lang, country
-                return new Locale(parts[0], parts[1]);
+                return Locale.of(parts[0], parts[1]);
             case 1:
                 if ("ROOT".equalsIgnoreCase(parts[0])) {
                     return Locale.ROOT;
                 }
                 // lang
-                return new Locale(parts[0]);
+                return Locale.of(parts[0]);
             default:
                 throw new IllegalArgumentException(
                     "Locales can have at most 3 parts but got " + parts.length + ": " + Arrays.asList(parts)

server/src/main/java/org/elasticsearch/index/mapper/DataStreamTimestampFieldMapper.java

@@ -165,10 +165,11 @@ public void doValidate(MappingLookup lookup) {
             Map<?, ?> configuredSettings = XContentHelper.convertToMap(BytesReference.bytes(builder), false, XContentType.JSON).v2();
             configuredSettings = (Map<?, ?>) configuredSettings.values().iterator().next();
 
-            // Only type, meta and format attributes are allowed:
+            // Only type, meta, format, and locale attributes are allowed:
             configuredSettings.remove("type");
             configuredSettings.remove("meta");
             configuredSettings.remove("format");
+            configuredSettings.remove("locale");
 
             // ignoring malformed values is disallowed (see previous check),
             // however if `index.mapping.ignore_malformed` has been set to true then

server/src/main/java/org/elasticsearch/index/mapper/DateFieldMapper.java

@@ -79,11 +79,16 @@ public final class DateFieldMapper extends FieldMapper {
 
     public static final String CONTENT_TYPE = "date";
     public static final String DATE_NANOS_CONTENT_TYPE = "date_nanos";
-    public static final DateFormatter DEFAULT_DATE_TIME_FORMATTER = DateFormatter.forPattern("strict_date_optional_time||epoch_millis");
+    public static final Locale DEFAULT_LOCALE = Locale.ENGLISH;
+    // although the locale doesn't affect the results, tests still check formatter equality, which does include locale
+    public static final DateFormatter DEFAULT_DATE_TIME_FORMATTER = DateFormatter.forPattern("strict_date_optional_time||epoch_millis")
+        .withLocale(DEFAULT_LOCALE);
     public static final DateFormatter DEFAULT_DATE_TIME_NANOS_FORMATTER = DateFormatter.forPattern(
         "strict_date_optional_time_nanos||epoch_millis"
-    );
-    private static final DateMathParser EPOCH_MILLIS_PARSER = DateFormatter.forPattern("epoch_millis").toDateMathParser();
+    ).withLocale(DEFAULT_LOCALE);
+    private static final DateMathParser EPOCH_MILLIS_PARSER = DateFormatter.forPattern("epoch_millis")
+        .withLocale(DEFAULT_LOCALE)
+        .toDateMathParser();
 
     public enum Resolution {
         MILLISECONDS(CONTENT_TYPE, NumericType.DATE, DateMillisDocValuesField::new) {
@@ -232,7 +237,7 @@ public static final class Builder extends FieldMapper.Builder {
         private final Parameter<Locale> locale = new Parameter<>(
             "locale",
             false,
-            () -> Locale.ROOT,
+            () -> DEFAULT_LOCALE,
             (n, c, o) -> LocaleUtils.parse(o.toString()),
             m -> toType(m).locale,
             (xContentBuilder, n, v) -> xContentBuilder.field(n, v.toString()),

server/src/main/java/org/elasticsearch/index/mapper/DateScriptFieldType.java

@@ -69,7 +69,7 @@ private static class Builder extends AbstractScriptFieldType.Builder<DateFieldSc
             (n, c, o) -> o == null ? null : LocaleUtils.parse(o.toString()),
             RuntimeField.initializerNotSupported(),
             (b, n, v) -> {
-                if (v != null && false == v.equals(Locale.ROOT)) {
+                if (v != null && false == v.equals(DateFieldMapper.DEFAULT_LOCALE)) {
                     b.field(n, v.toString());
                 }
             },
@@ -97,7 +97,7 @@ protected AbstractScriptFieldType<?> createFieldType(
             OnScriptError onScriptError
         ) {
             String pattern = format.getValue() == null ? DateFieldMapper.DEFAULT_DATE_TIME_FORMATTER.pattern() : format.getValue();
-            Locale locale = this.locale.getValue() == null ? Locale.ROOT : this.locale.getValue();
+            Locale locale = this.locale.getValue() == null ? DateFieldMapper.DEFAULT_LOCALE : this.locale.getValue();
             DateFormatter dateTimeFormatter = DateFormatter.forPattern(pattern, supportedVersion).withLocale(locale);
             return new DateScriptFieldType(name, factory, dateTimeFormatter, script, meta, onScriptError);
         }

server/src/main/java/org/elasticsearch/index/mapper/RangeFieldMapper.java

@@ -59,6 +59,7 @@ public class RangeFieldMapper extends FieldMapper {
 
     public static class Defaults {
         public static final DateFormatter DATE_FORMATTER = DateFieldMapper.DEFAULT_DATE_TIME_FORMATTER;
+        public static final Locale LOCALE = DateFieldMapper.DEFAULT_LOCALE;
     }
 
     // this is private since it has a different default
@@ -83,7 +84,7 @@ public static class Builder extends FieldMapper.Builder {
         private final Parameter<Locale> locale = new Parameter<>(
             "locale",
             false,
-            () -> Locale.ROOT,
+            () -> Defaults.LOCALE,
             (n, c, o) -> LocaleUtils.parse(o.toString()),
             m -> toType(m).locale,
             (xContentBuilder, n, v) -> xContentBuilder.field(n, v.toString()),