thisismypassport
thisismypassport
This is because pico_rename adds a few attributes (like used_globals) to the scopes it receives and assumes that all scopes then have it. Why do you need to run pico_rename...
It'd probably make sense to move the used_globals/locals definitions to the Scope class itself (its __init__) instead of monkey patching it - should resolve this exception.
I moved used_locals/globals (& now a very-rarely-needed used_members) to be lazy properties on Scope now, FYI.
Hi, thanks for the PR, but I don't see a reason for it - I'm fine with trailing whitespaces, especially on blank lines in the middle of methods.
Thanks, though I'm not as optimistic as you about the poke operator landing in pico8 :)
Currently, the minification steps modify both the node tree and tokens array. I'm not too happy about this either, and will probably improve this as part of a branch I'm...
Assuming I understood what you're asking for correctly, the idx and endidx attributes specify the range of the function's tokens within the source (e.g. node.source.text[node.idx : node.endidx])
Yeah, node.type == NodeType.function Yeah, the parent of the function is the root node, it indeed spans the whole source minus any leading/trailing whitespace/comments
This affects me as well, here is a general description of the issue: ```python def f(): return 1,2 a,b,c=1,*f() ``` This is valid python code that assigns to a, b,...
This something I've been missing as well.