DataTypeCodec relies on COMPAT fallback for 'BC' era parsing about vertx-sql-client

[vietj]

ah right, I thought an interaction with the database was required :-) @zakkak

[zakkak]

pinging @vietj since he was one of the last persons to touch that part

[vietj]

can you provide a reproducer (a SQL statement would work fine) for this @jerboaa ?

[jerboaa]

No, sorry. I have no idea. Perhaps somebody in quarkusio/quarkus#37208 can help?

[zakkak]

@vietj is the following reproducer OK?

1. Grab an OpenJDK 22+24-EA build from https://github.com/adoptium/temurin22-binaries/releases/tag/jdk-22%2B24-ea-beta

git clone https://github.com/zakkak/vertx-pg-1379-reproducer.git
cd vertx-pg-1379-reproducer
mvn package
/path/to/jdk-22+24-ea/bin/java -jar target/vertx-pg-example-1.0-SNAPSHOT-jar-with-dependencies.jar

All it does is:

System.out.println( DataTypeCodec.LDT_MINUS_INFINITY );

Which runs fine with JDK 21 printing:

-4713-11-24T00:00

but fails with JDK 22+24-ea printing:

Exception in thread "main" java.lang.ExceptionInInitializerError
	at com.example.App.main(App.java:13)
Caused by: java.time.format.DateTimeParseException: Text '4714-11-24 00:00:00 BC' could not be parsed at index 20
	at java.base/java.time.format.DateTimeFormatter.parseResolved0(DateTimeFormatter.java:2108)
	at java.base/java.time.format.DateTimeFormatter.parse(DateTimeFormatter.java:2010)
	at java.base/java.time.LocalDateTime.parse(LocalDateTime.java:494)
	at io.vertx.pgclient.impl.codec.DataTypeCodec.<clinit>(DataTypeCodec.java:1077)
	... 1 more

[jerboaa]

@vietj Do you have an ETA when this will be fixed? We will need the updated vertx version then in quarkus too. Thanks!

[vietj]

@jerboaa not yet, but for sure the issue is on our radar

[jerboaa]

OK, thanks!

[zakkak]

Hi @vietj, I applied the change in my fork and tested it on github actions. Unfortunately I can't do any more testing since I am not familiar with the code base. I also opened #1391 in case it helps.

[vietj]

@zakkak I see the bug you filed has been fixed, the JDK 22-EA in which the bug appear has never been released, so I am assuming the next EA will have the bug fixed and that shall be fine ?

[jerboaa]

@zakkak I see the bug you filed has been fixed, the JDK 22-EA in which the bug appear has never been released, so I am assuming the next EA will have the bug fixed and that shall be fine ?

Which bug? This one? https://bugs.openjdk.org/browse/JDK-8320431

That got closed as Not an issue, which essentially means a fix as suggested in #1391 is needed. It's intended behaviour in the JDK which was a side-effect of it working in earlier JDKs and got finally fixed.

[vietj]

ok so this is not a temporary workaround as far as you can tell

[jerboaa]

ok so this is not a temporary workaround as far as you can tell

Yes.