quoted_string railroad diagram differs when multiple

[RoDuth]

I've been updating the parser for an old project recently and can I say the railroad diagrams are a great feature!

I found an anomaly when using quoted_string in multiple places with set_result_name.

When used in a single place I get:

but when used in multiple places, e.g.:

from pyparsing import *


quoted_string.set_parse_action(remove_quotes)
none_token = Keyword("None")
numeric_token = Regex(r"[-]?\d+(\.\d*)?([eE]\d+)?").set_name("numeric token")
value_token = (none_token | quoted_string).set_name("value")("value")
date_value_token = (numeric_token | quoted_string).set_name("date value")("value")
generic_token = (date_value_token | value_token).set_name("generic token")
generic_token.create_diagram("test.html")

I get:

I did try .copy() and a few other ideas but none of them made any difference.

[ptmcg]

Yes, there is a known problem with the railroad diagrams generating empty blocks, that I've tried a few times to chase down, but it eludes me. That feels like a separate problem from this one, and I think I opened an issue on it a while ago, but closed it when I thought I had it fixed - turns out my solution was only a partial one. I'll reopen that ticket or create a new one.

It's possible that the multiple instances are actually creating separate expression copies (which is what happens internally when setting results names), and this might be part of the issue that I didn't consider before.