Quoting Gordon Smith <[email protected]>:
I don't think it would be hard to store the entire token stream that
produced a tree, but it sounds like you want to somehow keep track
of what tokens produced what node, which is more complicated and
probably overkill. We chose not to store tokens in order to minimize
memory.
I have another project that has to do with read-write AS AST, so I was
asking if it's something I could add a flag, or something that code be
turned on(the implementation) if a sub framework needed it.
What information is missing in the tree that you need? The presence
of parens? If you need paren information in the tree, I would just
add a field like hasParens to BinaryOperatorNode. Or add
ParenthesizedExpressionNode.
But why not just wrap parens around each binary operator expression
when you emit JS? Is that what FalconJS does?
I just did this today, I don't know how FalconJS does it since it uses
the BURM I try to stay away. :)
There are some somethings like String concatenation that add extra
logic and who knows what else will come up. I did fix it I feel 75%
today though.
Mike
- Gordon
-----Original Message-----
From: Michael Schmalle [mailto:[email protected]]
Sent: Wednesday, March 06, 2013 7:43 AM
To: [email protected]
Subject: [Falcon] Token stream recording
Hi,
Gordon this is aimed at you.
Do you see any way or know of any way to allow the parser or some
type of strategy to record all tokens when parsing?
Would you have any guidance on how to implement this if possible? Is
there any way to plug into the current repairing token stream to
save this out to a linked list like we talked about months ago?
Mike
--
Michael Schmalle - Teoti Graphix, LLC
http://www.teotigraphix.com
http://blog.teotigraphix.com
--
Michael Schmalle - Teoti Graphix, LLC
http://www.teotigraphix.com
http://blog.teotigraphix.com
