6 replies to this topic

[Sewbacca]

Does this ('^%a[%w_]+') pattern work for every Lua identifier?

[LBPHacker]

If the identifier starts at the first character of the string, then yeah, almost. I'd use

^[%a_][%w_]*

though, since underscore can lead the way too and you don't need more than 0 of the second class.

[Sewbacca]

^[%a_][%w_]*

I think
^[%a_][%w_]-
Is better, because * goes to the last found of [%w_], and - goes to the last found of [%w_] without gaps, <-- ^^
but thanks ^^.

Edited by Sewbacca, 14 July 2016 - 03:43 PM.

[LBPHacker]

Come again? Gape? You mean gap? Nope, that's not how it works. [%w_]* comes after [%a_], so whatever [%w_]* matched must come after whatever [%a_] matched. [%w_]- will match the least possible characters, which since you already have an [%a_] in there will be an empty string.

[Exerro]

Not quite, it won't match an already matched character, so the '-' section will just match one more character. You'll be limited to just 2 character identifiers, no more or less, if you do it with a '-' instead of '*'. LBPHacker's first suggestion is the way to do it.

[LBPHacker]

Nope, the '-' section won't match anything. As I've stated above, '-' will match the least possible number of characters, which is, in this case, 0. The '-' is used in situations in which it's followed by other sections, such as

local command, arguments = command_line:match("^%s*(.-) (.*)%s*$")

This way only the first chunk of non-whitespace characters is stored in command and the rest in arguments, for '-' matches the least possible number of characters. Yes, it could match an empty string had the leading spaces not been already matched by %s*, thus leaving only non-spaces for (.-) to match.

Edited by LBPHacker, 11 July 2016 - 07:37 PM.

[Exerro]

Ah, you're right. For some reason I thought the '-' matched at least one character, but it can match none as well. My bad.