Adds afterStart(WebServer) implementations to various delegating ServerLifecycle implementations that were missing it #9820
Conversation
…erLifecycle implementations that were missing it Signed-off-by: Laird Nelson <laird.nelson@oracle.com>
oracle-contributor-agreement
bot
added
the
OCA Verified
There was a problem hiding this comment.
IIRC an app can start multiple web servers (probably very rare, but I think possible?). The afterStart method accepts a WebServer as a parameter, but the afterStop method does not. Presumably some afterStart logic could do webserver-dependent things and might need to distinguish among which of the previously-started web servers has been stopped during its afterStop processing. Such code has no way of doing so with the current afterStop API.
Certainly if at most one webserver can ever be started in a JVM then this is a non-issue.
All fair points in my opinion. I'm also not particularly crazy about the fact that the "webserver is up!" "event" is buried in, for practical purposes, a routing, when the thing reacting to the event reception may have nothing to do with routing. The scope of this PR, however, is simply to add (required) implementations of the |
PR #9655 added an
afterStart(WebServer)method toServerLifecycle. It was approved on February 4. There were moreServerLifecycleimplementations than showed up at first glance. This PR attempts to ensure that all of the knownServerLifecycleimplementations in the code base do the right thing with theirafterStart(WebServer)implementations, so that they don't simply (incorrectly) inherit the (necessary)default"no-op" implementation.This PR gates #9704 and #9789.