Docs say:
> Quasi-quoting operator. Accepts an expression or a block and returns the AST
> that represents it. Within the quoted > AST, you are able to interpolate
> NimNode expressions from the surrounding scope
By this, I guess it likely means it pastes the sub-tree in at the given
position? Or does it truly interpolate it in then reconstruct the AST from the
resulting text?
I was trying to extend the class OOP programming macro form nim by example to
create parametrized classes (e.g. Cage[Animal]. Is this even a sane idea? I
notice expandMacros refers to expanding one level of macros, but I'm not sure
that generics that come out of macros get handled?
I notice that
dumpTree:
class Cage[Animal] of RootObj:
var occupant: T
shows a BracketExpr node, while
dumpTree:
type Cage[Animal] of RootObj:
shows a GenericParams node, with no BracketExpr node anywhere. But I guess
there's no reason to be scared about the lack of a common factor since
BracketExpr is more general and the parser only produces GenericParams in a
type context? Just recreate tree to taste?
This simple quote() code also confuses me:
import future
import macros
macro decDumbType(): untyped =
result =
macros.quote do:
type Dumb = ref object of RootObj
contents: int
method frobnicate(this: Dumb) {.base.} =
echo "frobnicating!"
expandMacros:
decDumbType()
by producing this:
> type
>
>
> Dumb112019 = ref object of RootObj
> contents: int
>
> method frobnicate(this112017: Dumb112015) {.base.} = echo ["frobnicating!"]
>
> generic_obj_mac.nim(31, 1) template/generic instantiation from here
> generic_obj_mac.nim(32, 14) Error: illformed AST: method
> frobnicate(this112017:
>
> Dumb112015) {.base.} =
> echo ["frobnicating!"]
why is Dumb being rewritten?
