Vlad Chesnokov

In my experience, the common approach to this is to use annotations. (Take a look at jackson, for example). As user, the most convenient approach for me would be automatic...

For serializing data class to HOCON it would be logical to make an inverse conversion: transform all property names to hyphen-case, unless exact name is specified by annotation.

Maybe the problem cased by zero timeout in socket. See [this PR](https://github.com/LonamiWebs/Telethon/pull/1432) for details

Hi! Could anyone clarify, why nested suspending transactions code (as recommended by @AlexeySoshin) ```kotlin newSuspendedTransaction { exec("SELECT 1") suspendedTransaction { exec("SELECT 2") } } ``` has different behavior in comparison...

Ok, I'll try to make MWE tomorrow.

Leaves RabbitMq running after test execution. Without System.exit() works fine. ``` import io.arivera.oss.embedded.rabbitmq.EmbeddedRabbitMq; import io.arivera.oss.embedded.rabbitmq.EmbeddedRabbitMqConfig; import org.junit.AfterClass; import org.junit.BeforeClass; import org.junit.Test; public class RabbitTest { static EmbeddedRabbitMq rabbitMq; @Test public...

Another, more real test-case. One can restart tests during debugging (started test, stopped at breakpoint, fixed code, restarted). In this case RabbitMq would not be stopped.

I've launched test.sh with zero changes and got this output: ``` Any RabbitMQ process found? No process found. Executing test that will exit without stopping the Embedded RabbitMQ process... RabbitMQ...

Possible fix (program.pp): ``` ... case $ensure_process { 'stopped': { supervisord::supervisorctl { "stop_${name}": command => 'stop', process => $name } } 'removed': { supervisord::supervisorctl { "remove_${name}": command => 'remove', process...

This is ridiculous. What's the point of issue tracking if all of them are simply ignored?