groovy-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Andrew Bayer <>
Subject Trying to determine if a change in 2.4.8 was deliberate...
Date Thu, 30 Mar 2017 17:16:24 GMT
I'm working on getting Jenkins' groovy-cps transforming magic working with
2.5.0-alpha-1-SNAPSHOT and hit a few problems. I worked around one of them
easily enough (GlobalClassSet switched to a new collection class for its
items field), but the other is hairier. The change in question is,
which went into 2.4.8.

What I'm seeing is that
is invoking the super's method through the inheritor. As far as I can tell
- we end up with a CpsCallableInvocation being thrown, meaning our
transformed code in the inheritor was invoked here. In our CPS contexts,
that CpsCallableInvocation is the right result, but not here.

Note that this isn't relevant in any other Groovy use case that I'm aware
of - it's particular to this usage. Or rather, this is the only case I can
find where the end result is different. I eventually found the specific
internal behavior that changed, getting us a different method to invoke - has
a test case that passes in 2.4.7 and earlier but fails in 2.4.8 and later.

I do have what *seems* to be a workaround at
but that feels dirty to me. Is there a better way to do this? Was this an
intended result of the change, an unintentional but desirable side effect,
or a bug? Any thoughts would be much appreciated. =)


View raw message