This PR adjusts some tests that seem to contradict each other on whether or not x.init() is allowed for arbitrary x. Based on discussion in #8991, it sounds like they should not be allowed. Ruling them out is a relatively simple post-parse check.
While there, I noted that x.postinit() is in a similar spot, and disallowed it too.
Closing because for the time being, what to do about postinit is not very clear; sounds like we might be leaning towards allowing it to be invoked arbitrarily.
Closes #8991; fixes tests broken on
main
.This PR adjusts some tests that seem to contradict each other on whether or not
x.init()
is allowed for arbitraryx
. Based on discussion in #8991, it sounds like they should not be allowed. Ruling them out is a relatively simple post-parse check.While there, I noted that
x.postinit()
is in a similar spot, and disallowed it too.Reviewed by @benharsh -- thanks!
Testing