Use of WS as a separator in parsing sexpr

[caw]

I have a simple sexpr parser:

start: value
value: list
      | string
      | number
list  : "(" [value (" " value)*] ")"
%import common.WS
%ignore WS

this fails on

(foo
(bar baz))

if there is no space " " after foo

I can fix this by setting the separator in the list rule to WS

start: value
value: list
      | string
      | number
list  : "(" [value (WS value)*] ")"
%import common.WS
%ignore WS

I get a correct parse, but is that the sensible way to do this? And I'm not sure how "%ignore WS" interacts with WS being in a rule.

with thanks

[MegaIng]

This is fine. The first example fails because you are explicitly requiring a space. Don't b so that it that is not the correct semantics.

The second example will work, lark will prefer using a terminal in a rule instead of ignoring it.

For a slightly different syntax you can also have no seperator in the list at all.

[caw]

thank you - I appreciate your help.
the parse now returns a WS Token. Is there a way to prevent that being used in building the parse tree?

[caw]

Oh, and the suggestion to have no separator in the list is very interesting!

[MegaIng]

You can control how the Tree looks like, see Shaping the tree: https://lark-parser.readthedocs.io/en/latest/tree_construction.html#shaping-the-tree