Arnaud Geiser

I've added some tests regarding this feature. However, I'm wondering if that's you expect from this. The connection won't be disposed once it's elapsed but the next call after it's...

Even without answer from @alexander-yakushev, I would merge this PR into aleph. Yes it can be improved, yes it's not perfect but still, it solves an issue that some people...

For now, I would say there are three ways to handle multipart requests: 1. __Use the multipart-params middleware__ It's working as long as your body is an `java.io.InputStream`. By default,...

Fair enough. I wasn't aware of the inflexibility of the multipart-params middleware. On the current state of Aleph, that's the three solutions I have in mind. Let's see what will...

The only thing that bother me a bit is the fact we don't have any tests with a pipeline containing `HttpObjectAggregator`. (only for the websocket)

> FTR: I noticed this when trying to add a test case with an intentionally throwing handler: AFAICT, in the :raw-stream? true case, such an unhandled exception would always lead...

Whether @KingMob is fine with the state of this PR, I would merge it.

Thank you @alexander-yakushev and @DerGuteMoritz for the hard work! Let's say it's never too late :)

For now, I don't. Be we know there are some issues on the HTTP client. We also experienced this "kind" of scenarios on my company and we ditched the `aleph...

@kachayev did a nice work at summarizing the different paths we could take for the cleanup API. https://github.com/clj-commons/aleph/pull/432#issuecomment-449185986 He proposed an implementation that didn't introduce any breaking change. However, unless...