@sdeleuze do you know if there are plans to support strict json assertions from the

WebTestClient

? your

JsonExpectationsHelper

already has support ..

With Spring Data when I do a database operation is that blocking on non-blocking? i.e. repository.save(obj)?

spring data jpa is blocking

most databases are blocking anyway

but doesn't hibernate cache and batch operations

so the actual function call returns way sooner than the actualy queries go across wire

how can hibernate cache and batch if it waits for the database to actually respond

it can batch but it doesn’t necessarily

It also depends on where you have your transaction boundary

it also depends how

i was about to write it 🙂

@Czar not faire

-e

¯\_(ツ)_/¯

so it may or may not be blocking and we just assume it will block until the query goes across wire

but it may return sooner if it batches or caches

in all cases it will block but it might be short or long

“it depends”

what’s your underlying question?

if I should be doing launch to not wait for a db operations to go across network

like a rest api call

like

@Transactional fun saveSeveral(list: Something) {
  list.forEach(repo::save)
}

will most probably cache/batch separate saves, but it will return only when transaction was completed.

otherwise it wouldn't be able to roll back

@Czar would you treat a db operations like a rest operation as far the way you do async?

With jdbc/jpa you could emulate non-blocking with coroutines or plain Spring

@Async

, this would mostly matter for your servlet thread pool. If I had a situation with many simultaneous connections, I certainly would

But I suggest testing first. And as a bonus here's a bit outdated but still relevant article on a similar matter: https://dzone.com/articles/spring-and-servlet-30-asynchronous-processing Looking at several approaches and weighing them against each other is always a good practice, don't take anyone on their word, it's very hard to give the best answer if one doesn't know all the details and laying out all the details in public is usually not possible.

isn't hibernate batch dependent on entity mapping too? https://docs.jboss.org/hibernate/orm/current/userguide/html_single/chapters/batch/Batching.html ->

Hibernate disables insert batching at the JDBC level transparently if you use an identity identifier generator.

(though it seems you mean a different type of batch, more similar to caching).

I have something like this in Spring MVC code:

@Controller
open class MyController {

    @RequestMapping("/**")
    fun mapping(req: HttpServletRequest, res: HttpServletResponse) {
       // write to res.outputStream
    }
}

I'm trying to do similar thing with Spring WebFlux:

@Controller
open class MyController {

    @RequestMapping("/**")
    suspend fun mapping(req: ServerHttpRequest, res: ServerHttpResponse) {
       // write with res.writeAndFlushWith { }
    }
}

Will it work? Can the

mapping

function have

suspend

modifier?

I have problem with

void

result type described here: https://docs.spring.io/spring/docs/current/spring-framework-reference/web-reactive.html#webflux-ann-return-types

spring is always trying to search for view based on the request uri if my function doesn't return a value