Klever

At the moment the same session object is used by at least two Core components represented by separate processes. Though, the corresponding memory is copied somehow, but it still does not guarantee that requests are handled well. There are corresponding bug reports, e.g. here, and the reasonable advice is to create sessions after forking.

Maybe this will help to get rid of the workaround that we developed to handle so called ZIP errors, i.e. when Bridge should receive archives within requests from Core, but it finds out that they are broken. In this case Core infinitely resends such request and I guess that after a second attempt everything is okay (but I did not evaluate that ever).