Update to Scala 2.11.0-RC4, adapting to change in quasiquotes #71
Conversation
Namely: scala/scala#3656 I can't find a way to express a QQ that matches an constructor invocation *and* lets me bind a reference to the `New` tree. So I've dropped down to a borrowed version of `TreeInfo#Applied`.
- simpler means to calculate `applyDepth` - remove unused binder
Update to Scala 2.11.0-RC4, adapting to change in quasiquotes
|
So from what I can see, syntactic disjointness is actually undesireable here, right? Also seems related to https://roslyn.codeplex.com/discussions/541303 (to my best understanding of what they are discussing there) |
|
Why not add one extra |
|
@xeno-by experts can get away without disjointness because they know inner representations of trees, new users – not as much. |
|
@densh I can't bind to the |
|
Note that this doesn't work, as the case q"${n @ q"new ${_}"}[..$targs](...$argss)" |
|
From a transformation point of view, anytime you want to match a method application, you probably also want to match a constructor invocation. The uniformity helps here, IMO. |
|
I guess we can discuss this on one of the upcoming meetings. Forced disjointness here might have been premature. We can document this case and revert to previous behavior back in 2.11.1. |
Namely: scala/scala#3656
I can't find a way to express a QQ that matches an constructor
invocation and lets me bind a reference to the
Newtree.So I've dropped down to a borrowed version of
TreeInfo#Applied./cc @xeno-by @densh