There's been discussion of some sort of forking-data-structure as the locals type, which obviously adds untold complexity to the language … but failing that, we either:
Copy all of the locals-pairs (thus, variable definitions) to a new locals, thus leaving the new fork of program execution indicated by the new Execution distinct from any other forks,
Share the exact same locals object.
Both have upsides and downsides: the former leaves any execution that gets called (because call-pattern clones; including those that co-consume), no longer sharing locals with any other executions of the same Script, which may be rather-less-than-intuitive. :P
(The latter, obviously, also means needing to work some implicit ownership-management shit into locals accesses. Not sure exactly what to do about that, could get messy quick.)
There's been discussion of some sort of forking-data-structure as the locals type, which obviously adds untold complexity to the language … but failing that, we either:
Both have upsides and downsides: the former leaves any execution that gets called (because call-pattern clones; including those that co-consume), no longer sharing locals with any other executions of the same Script, which may be rather-less-than-intuitive. :P
(The latter, obviously, also means needing to work some implicit ownership-management shit into locals accesses. Not sure exactly what to do about that, could get messy quick.)