lostluck commented on issue #28981: URL: https://github.com/apache/beam/issues/28981#issuecomment-1786283003
Ah, good find. That page definitely needs to be updated too. Thank you! Technically it's not Prism returning the error, it's the SDK. As a portable runner, Prism doesn't know very much about the SDK it's running. The problems also apply to Flink, Dataflow, Spark etc. The go direct runner, which isn't supportable and by definition, isn't scalable, can't properly ensure your pipeline is correct, which is why it's deprecated. But as stated, it's not possible to reliably support it in the SDK as it is. We'd need a different Go SDK to manage it. ------ I will say that if it could reliably be supported, it would be my first choice, but I've not seen any good mechanism for it, that won't be an ongoing maintenance headache. Manually registering works every time is necessary for unit testing anyway, and doesn't encourage closures which cause more problems, because we can't serialize the closure. But until we can reliably support anons, it's better to not make the most brittle part of the SDK be front and center in the examples. -- 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]
