Yeah it seems like a bad idea. Do you have a list of the problematic actions ?
On Tue, Sep 11, 2018 at 11:28 AM Thomas Draier <[email protected]> wrote: > > Hi, > > I did a simple fix checking the recursion depth. This work in my case > (segments) and probably most of the cases, but some actions directly call > the EventService.send method - which will reset the recursion count. > Actually I have no idea why these actions are sending the "profileUpdated" > event by themselves, since it's sent anyway by the EventService itself > after the execution of all rules. Seems a bad idea to me ? > > Thomas > > > On Mon, Sep 10, 2018 at 4:37 PM Serge Huber <[email protected]> wrote: > > > Hello Thomas, > > > > Agreed the first one should be fixed. > > > > For the second one, I was thinking that maybe we should try to detect > > the recursion and if so prevent it from happening. What do you think? > > This would involve some kind of common contact and some kind of > > tracing of which events are occurring so that we can detect this loop. > > > > cheers, > > Serge... > > On Mon, Sep 10, 2018 at 4:13 PM Thomas Draier <[email protected]> wrote: > > > > > > Hi, > > > > > > I've found 2 issues with segments : > > > https://issues.apache.org/jira/browse/UNOMI-197 and > > > https://issues.apache.org/jira/browse/UNOMI-198 . First one can be > > easily > > > fixed, I don't know if you see any reason why the enabled value was not > > > checked ? For the stack overflow issue, not sure what we should do .. the > > > fact that it's recursive is not good (would be better with a queue of > > > event) but won't change the infinite flow of events. Any idea if we > > should > > > do something about it ? > > > > > > Thomas > >
