I checked the sizzle code and see a definition.
var chunker = /((?:\((?:\([^()]+\)|[^()]+)+\)|\[(?:\[[^\[\]]*\]|['"][^'"]*['"]|[^\[\]'"]+)+\]|\\.|[^ >+~,(\[\\]+)+|[>+~])(\s*,\s*)?((?:.|\r|\n)*)/g,
I want to know how to find out what string(s) this regular expression will match?
See this article. Explanation in a multiline regex:
var chunker = /
(
(?:
# One or more sets of parentheses that contain a string, or another set of
parentheses with a string
\(
(?:
\([^()]+\)
|
[^()]+
)+
\)
|
# Or one or more sets of brackets that contain a string, or another set of
brackets with a string
\[
(?:
\[[^\[\]]*\]
|
['"][^'"]*['"]
|
[^\[\]'"]+
)+
\]
|
# Or a backslash followed by any character
\\.
|
# Or one or more of any except these characters: > +~,([\
[^ >+~,(\[\\]+
)+
# or any one of these characters: >+~
|
[>+~]
)
# followed by zero or one commas, which may be surrounded by whitespace
(\s*,\s*)?
# followed by zero or more of anything, including line endings
((?:.|\r|\n)*)
/g
This expression contains three matching groups: A "validated" selector expression, eventual comma, and everything after that. It will continuosly be called on the selector to split it up in parts, see the Sizzle constructor for details.
It would be speculative. However, using RegexBuddy you can 'document' and visualize the expression.
Because it is a paid application i've exported the comments on pastebin. Hope this will help you. (Note that it does not support Sizzle and that i've used JavaScript language to document the expression).
To learn how to do it by hand, you can simply break it down like this (this is a task that will surely help you in the future):
var chunker = /((?:\((?:\([^()]+\)|[^()]+)+\)|\[(?:\[[^\[\]]*\]|['"][^'"]*['"]|[^\[\]'"]+)+\]|\\.|[^ >+~,(\[\\]+)+|[>+~])(\s*,\s*)?((?:.|\r|\n)*)/g,
The first part is to know how a regex is generally written, so in this case it would be:
var variableName = /stuff/g;
^ ^ ^^
delimiters/---+--/|
| |
regex / \global modifier, can also have a case modifier
So let's strip away the modifier and the start and end and evaluate just the regex:
((?:\((?:\([^()]+\)|[^()]+)+\)|\[(?:\[[^\[\]]*\]|['"][^'"]*['"]|[^\[\]'"]+)+\]|\\.|[^ >+~,(\[\\]+)+|[>+~])(\s*,\s*)?((?:.|\r|\n)*)
Still gobbledygook, but let's continue: The regex is for capturing content and we know that unescaped (meaning they use a backslash before the character being escaped \) parenthesis are capturing braces, let's add some newlines around each parenthesis (only for the purposes of examination!!!) and pipes mean "or"
(
(?:
\(
(?:
\(
[^()]+
\)
|
[^()]+
)+
\)
|
\[
(?:
\[[^\[\]]*\]
|
['"][^'"]*['"]
|
[^\[\]'"]+
)+
\]
|
\\.
|
[^ >+~,
(
\[\\]+)+
|
[>+~]
)
(\s*,\s*)?
(
(?:
.
|
\r
|
\n
)*
)
Square braces mean "selectively match what is inside of this as any value from a choice-set" and so now I'm going to add a few comments, and let you work out the rest (also, it does look like some parens are missing)
(
(?: //<-- start a non-capturing group (means, don't tell the app we matched)
\(
(?: //<-- start a non-capturing group (means, don't tell the app we matched)
\(
[^()]+ //one or more paren pairs inside a paren pair
\)
|
[^()]+ //or one or more paren pairs
)+
\)
|
\[
(?:
\[[^\[\]]*\]
|
['"][^'"]*['"]
|
[^\[\]'"]+
)+
\]
|
\\.
|
[^ >+~,
(
\[\\]+)+
|
[>+~] //either a tilde, plus or opening brace
)
(\s*,\s*)?
(
(?:
.
|
\r
|
\n
)*
)