BNF生产规则中应如何对待“或”？

Question

I'm looking at the BNF grammar for SVG path data, and one of the derivation rules is:

digit-sequence ::= digit | digit digit-sequence

Is there a sematic difference beween this rule and:

digit-sequence ::= digit digit-sequence | digit

Exactly what does the | mean in a BNF grammar? Should the first match be selected, or the one that consumes most of the input?

Answer 1

| in a BNF grammar means alternation, ie if the current token matches one alternative or another it must be accepted. Here is a tutorial on BNF.

However, the rule you quoted is recursive (note digit-sequence is on both the left- and the right-hand sides of the rule) so that rule means a sequence of digits, eg [0-9]+ in regex.

BTW, parsing SVG path data seems to be an untrivial task so that a general BNF parser was once used to parse path data in combination with XML parser — https://metacpan.org/release/MarpaX-Languages-SVG-Parser