ESQL: Fix functions emitting warnings with no source (#122821)

Fixes #122588

- Replaced `Source.EMPTY.writeTo(out)` to `source().writeTo(out)` in functions emitting warnings
  - Did the same on all aggs, as Top emits an error on type resolution. This is not a bug, as type resolution errors should only happen in the coordinator. Another option would be changing Top to not generate that error there, and make it implement instead `PostAnalysisVerificationAware`
- In some cases, we don't even serialize an empty source. So I had to add a new `TransportVersion` to do so
  - As an special case, `ToLower` and `ToUpper` weren't serializing a source, but they don't emit warnings. As they were the only remaining functions not serializing the source, I added it there too

Author: ivancea

docs/changelog/122821.yaml

@@ -0,0 +1,6 @@
+pr: 122821
+summary: Fix functions emitting warnings with no source
+area: ES|QL
+type: bug
+issues:
+ - 122588

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

@@ -203,6 +203,7 @@ static TransportVersion def(int id) {
     public static final TransportVersion RERANKER_FAILURES_ALLOWED = def(9_013_0_00);
     public static final TransportVersion VOYAGE_AI_INTEGRATION_ADDED = def(9_014_0_00);
     public static final TransportVersion BYTE_SIZE_VALUE_ALWAYS_USES_BYTES = def(9_015_0_00);
+    public static final TransportVersion ESQL_SERIALIZE_SOURCE_FUNCTIONS_WARNINGS = def(9_016_0_00);
 
     /*
      * STOP! READ THIS FIRST! No, really,

x-pack/plugin/esql/qa/testFixtures/src/main/resources/boolean.csv-spec

@@ -287,6 +287,26 @@ emp_no:integer | is_rehired:boolean       | a1:boolean
 10005          | [false,false,false,true] | false
 ;
 
+mvSliceWarnings
+required_capability: functions_source_serialization_warnings
+
+FROM employees
+| SORT first_name ASC
+| EVAL
+    start = CASE(first_name == "Alejandro", 1, 0),
+    end = CASE(first_name == "Alejandro", 0, 1),
+    result = MV_SLICE(is_rehired, start, end)
+| KEEP first_name, result
+| LIMIT 1
+;
+
+warning:Line 6:10: evaluation of [MV_SLICE(is_rehired, start, end)] failed, treating result as null. Only first 20 failures recorded.
+warning:Line 6:10: org.elasticsearch.xpack.esql.core.InvalidArgumentException: Start offset is greater than end offset
+
+first_name:keyword | result:boolean
+Alejandro          | null
+;
+
 values
 required_capability: agg_values
 

x-pack/plugin/esql/qa/testFixtures/src/main/resources/date.csv-spec

@@ -366,6 +366,27 @@ date1:date               | dd_ms:integer
 2023-12-02T11:00:00.000Z | 1000
 ;
 
+dateDiffTestWarnings
+required_capability: functions_source_serialization_warnings
+
+FROM employees
+| WHERE first_name IN ("Alejandro", "Mary")
+| SORT first_name ASC
+| EVAL date = TO_DATETIME("2023-12-02T11:00:00.000Z")
+| EVAL dd_ms = DATE_DIFF(first_name, date, date)
+| KEEP date, dd_ms
+| LIMIT 2
+;
+
+warning:Line 5:16: evaluation of [DATE_DIFF(first_name, date, date)] failed, treating result as null. Only first 20 failures recorded.
+warning:Line 5:16: java.lang.IllegalArgumentException: A value of [YEAR, QUARTER, MONTH, DAYOFYEAR, DAY, WEEK, WEEKDAY, HOUR, MINUTE, SECOND, MILLISECOND, MICROSECOND, NANOSECOND] or their aliases is required; received [Alejandro]
+warning:Line 5:16: java.lang.IllegalArgumentException: Received value [Mary] is not valid date part to add; did you mean [day]? 
+
+date:date                | dd_ms:integer
+2023-12-02T11:00:00.000Z | null
+2023-12-02T11:00:00.000Z | null
+;
+
 evalDateDiffMonthAsWhole0Months#[skip:-8.14.1, reason:omitting millis/timezone not allowed before 8.14]
 
 ROW from=TO_DATETIME("2023-12-31T23:59:59.999Z"), to=TO_DATETIME("2024-01-01T00:00:00")

x-pack/plugin/esql/qa/testFixtures/src/main/resources/math.csv-spec

@@ -1009,6 +1009,33 @@ mvsum:unsigned_long
 null
 ;
 
+mvSumOverflowOnNode
+required_capability: functions_source_serialization_warnings
+
+FROM employees
+| SORT first_name ASC
+| EVAL
+    ints = CASE(first_name == "Alejandro", [0, 1, 2147483647], 0),
+    longs = CASE(first_name == "Alejandro", [0, 1, 9223372036854775807], 0),
+    ulongs = CASE(first_name == "Alejandro", [0, 1, 18446744073709551615], 0),
+    intsResult = mv_sum(ints),
+    longsResult = mv_sum(longs),
+    ulongsResult = mv_sum(ulongs)
+| KEEP first_name, intsResult, longsResult, ulongsResult
+| LIMIT 1
+;
+
+warning:Line 7:14: evaluation of [mv_sum(ints)] failed, treating result as null. Only first 20 failures recorded.
+warning:Line 7:14: java.lang.ArithmeticException: integer overflow
+warning:Line 8:15: evaluation of [mv_sum(longs)] failed, treating result as null. Only first 20 failures recorded.
+warning:Line 8:15: java.lang.ArithmeticException: long overflow
+warning:Line 9:16: evaluation of [mv_sum(ulongs)] failed, treating result as null. Only first 20 failures recorded.
+warning:Line 9:16: java.lang.ArithmeticException: unsigned_long overflow
+
+first_name:keyword | intsResult:integer | longsResult:long | ulongsResult:unsigned_long
+Alejandro          | null               | null             | null
+;
+
 e
 // tag::e[]
 ROW E()

x-pack/plugin/esql/qa/testFixtures/src/main/resources/string.csv-spec

@@ -1068,6 +1068,25 @@ Gatewayinstances|Gateway instances
 null            |null             
 ;
 
+replaceWarnings
+required_capability: functions_source_serialization_warnings
+
+FROM employees
+| SORT first_name ASC
+| EVAL
+    regex = CASE(first_name == "Alejandro", "(", ""),
+    result = replace(first_name, regex, "")
+| KEEP first_name, result
+| LIMIT 1
+;
+
+warningRegex:Line 5:10: evaluation of \[replace\(first_name, regex, \\"\\"\)\] failed, treating result as null. Only first 20 failures recorded.
+warningRegex:Line 5:10: java.util.regex.PatternSyntaxException: Unclosed group near index 1(%0D)?%0A\(
+
+first_name:keyword | result:keyword
+Alejandro          | null
+;
+
 left
 // tag::left[]
 FROM employees

x-pack/plugin/esql/src/main/java/org/elasticsearch/xpack/esql/action/EsqlCapabilities.java

@@ -204,6 +204,11 @@ public enum Cap {
          */
         FN_ROUND_UL_FIXES,
 
+        /**
+         * Fixes for multiple functions not serializing their source, and emitting warnings with wrong line number and text.
+         */
+        FUNCTIONS_SOURCE_SERIALIZATION_WARNINGS,
+
         /**
          * All functions that take TEXT should never emit TEXT, only KEYWORD. #114334
          */

x-pack/plugin/esql/src/main/java/org/elasticsearch/xpack/esql/expression/function/aggregate/AggregateFunction.java

@@ -69,7 +69,7 @@ protected AggregateFunction(StreamInput in) throws IOException {
 
     @Override
     public final void writeTo(StreamOutput out) throws IOException {
-        Source.EMPTY.writeTo(out);
+        source().writeTo(out);
         out.writeNamedWriteable(field);
         if (out.getTransportVersion().onOrAfter(TransportVersions.V_8_16_0)) {
             out.writeNamedWriteable(filter);

x-pack/plugin/esql/src/main/java/org/elasticsearch/xpack/esql/expression/function/scalar/date/DateDiff.java

@@ -197,7 +197,7 @@ private DateDiff(StreamInput in) throws IOException {
 
     @Override
     public void writeTo(StreamOutput out) throws IOException {
-        Source.EMPTY.writeTo(out);
+        source().writeTo(out);
         out.writeNamedWriteable(unit);
         out.writeNamedWriteable(startTimestamp);
         out.writeNamedWriteable(endTimestamp);

x-pack/plugin/esql/src/main/java/org/elasticsearch/xpack/esql/expression/function/scalar/multivalue/AbstractMultivalueFunction.java

@@ -41,7 +41,7 @@ protected AbstractMultivalueFunction(StreamInput in) throws IOException {
 
     @Override
     public final void writeTo(StreamOutput out) throws IOException {
-        Source.EMPTY.writeTo(out);
+        source().writeTo(out);
         out.writeNamedWriteable(field);
     }