Closed CAD97 closed 2 years ago
Afaict both the PR and the alternate change are correct and produce no normative change beyond solving the original issue 👍
I actually slightly prefer the alternate change; I know that I initially missed that an exponent could have a sign when I was implementing, and so I like being more explicit with where exactly a sign is allowed in a number. It also avoids defining a digit_
, which I slightly prefer.
Swapped in the preferred change instead of the less desirable minimal one
digit
is used as a subtraction in bare-identifier, but never defined.An alternate change:
(NOTE: this has now flip-flopped to opposite what it was when @tabatkins made their comment.)
I think I slightly prefer the PR'd version, but it changes the implied grammar tree structure more significantly. This change "just" gives names to the
[0-9]
and[0-9_]
productions.