After a complete constructing the EBNF of nftables (needed by TextMate,
LSP, and Vim syntax highlight), I notice many repeatedly overlaps of
syntaxes within the parser_bison.y.
One example of such a overlap pair:
All within the tcp_hdr_option_kind_and_field, the Bison parse tree
offers different pathways to the "same?" following `nft` expression:
tcp option window LENGTH
and
tcp option window COUNT
For 'tcp options window LENGTH', the Bison pathway is:
tcp_hdr_option_kind_and_field->tcp_hdr_option_type->WINDOW->LENGTH
For 'tcp option window COUNT', the Bison pathway is:
tcp_hdr_option_kind_and_field->WINDOW->tcpopt_field_window->COUNT
Is this intentional overlap? Or did I overlook some hidden feature(s)
surrounding LENGTH and COUNT?
The Railroad diagram for latest 'tcp_hdr_expr' is only XHTML-viewable
via a web browser 'file:///' from your hard drive (GitHub defangs
[X]HTML; SVG is too big, PNG is too squishy):
https://github.com/egberts/vim-nftables/blob/master/doc/nftables-railroad-chart.xhtml
