Github user uce commented on the pull request:
https://github.com/apache/flink/pull/1202#issuecomment-145885023
It's not superficial at all. It's cleaner, because the retriever is
independent of the redirect logic then.
I've decided against it though in order to keep the handlers cleaner (there
only two different kinds atm though, so it would be OK to add it) and because
the associated job manager URL is only known after start() is called, so I
would have to update all handlers lazily after that.
This is due to the start up ordering of
1. start web server (bind to port)
2. get web server port and start job manager with web server port configured
3. set job manager address at web server
This ordering allows both components to pick random ports.
I just didn't want to change too much code this at this point, because I
genuinely hope that we will make the execution graph serialisable after the
release and get back to the old solution. But I'm also happy to change it if
you like. Just ping me.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
