rmannibucau commented on issue #8321: URL: https://github.com/apache/hop/issues/8321#issuecomment-5655512539
> @rmannibucau No, I don't want to have pipelines and workflows wait until the logging back-end is ready. It really needs to be synchronous or a lot is going to break all over the place and it won't be easy to notice. well the points were: 1. it is async only when you desire (conf) 2. async is done to evaluate at publish time some data so in general in pipelines with a lot of data you really want to be async and it doesn't break anything the overall outcome is that it works today (1) and it can be better tomorrow (2), this is why I asked you specific pointers if you do have them cause I fail to see an issue there > The traditional Java world simply throws log lines into some logging system that just has severity, a timestamp and some text. That's it and they're done with it. Hop really needs a lot more for things like the various logging options to work. We want to know exactly where something was logged, when, which transform, action, which parent, the whole nine yards. Hmm, this is wrong since 15 years I fear, MDC and NDC were designed to not be limited to that exactly. I'm not 100% sure of which one would be the most relevant (MDC is insanely simple and well supported, NDC is more powerful but harder in general to use and has context propagation hypothesis which might be wrong in some hop runtime env not sync themselves - that said it is not more broken than today on this aspect I think). PR uses MDC (https://github.com/apache/hop/pull/8320/changes#diff-ab30e59089eedc3b3c0c049da987c41c80450184eb0f732c8bb4f3e4a3b99f88R43 for ex) for now. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
