My first instinct here is to rewrite `publisherToByteArray` without the use of RxJava. I'll take a crack at that, but first I need to figure out how to set up IntelliJ with core and client in the same workspace.
On Sun, Dec 22, 2019 at 5:10 AM Oleg Kalnichevski <[email protected]> wrote: > On Sun, 2019-12-22 at 10:28 +0100, Oleg Kalnichevski wrote: > > On Sat, 2019-12-21 at 10:58 +0100, Oleg Kalnichevski wrote: > > > > > > > > > ... > > > > > > > > It appears that the defect only occurs with TLS transport and > > > entity > > > enclosing requests. It also looks to be a race condition of some > > > sort. > > > It is nearly impossible to reproduce with context / wire logging > > > on. > > > > > > I will continue looking into it tomorrow. > > > > > > Oleg > > > > > > > Hi Ryan > > > > I can now reliably reproduce the issue with context / wire logging > > on. > > > > So far it looks quite certain there is a race condition of a sort in > > the reactive binding layer. > > > > All failing tests get stuck here > > > > > > https://github.com/apache/httpcomponents-client/pull/181/commits/c1b060703ac57fc4806d1249b906a385648862b5#diff-4aa21c6723799de37acda71cbe858faaR254 > > > > However all message exchanges clearly look completed in the context > > log > > and there appear to be no issues in the protocol handling layer. > > > > It would be great if you could take a look. I will continue debugging > > reactive binding code in the meantime. > > > > Cheers > > > > Oleg > > > > > > Without #publisherToByteArray conversion all tests pass consistently > without a single failure after many repetitions. > > I will start looking into ReactiveDataConsumer. Any help would be > appreciated. > > Oleg > > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [email protected] > For additional commands, e-mail: [email protected] > >
