The following regular expression should find all matches of ${any expression}
, the code is as follows:
const reg= /\$\{[^] \}/g
let txt= '`${i " test"} RESULT ${2 4} now ${i} hi`'
let result= [...txt.matchAll(reg)];
console.log(result)
As you will notice, the result is that it extracts almost the entire string, the correct operation should print in the console an array of 3 elements that would contain the ${any expression}
The following case shows an error that is generated if I use: [^}]
const reg= /\$\{[^}] \}/g
let i= "some"
let txt= `${i " test"} RESULT ${2 4} now ${i "}" } hi`
let txtString= '`${i " test"} RESULT ${2 4} now ${i "}" } hi`'
let result= [...txtString.matchAll(reg)];
console.log(result)
console.log(txt)
the expression ${i "}" }
is valid in JavaScript so the regular expression should return [${i "}" }, other matches]
but in the example shown it returns
${i "}
CodePudding user response:
If you can describe the token syntax inside ${...}
you can still use a regex, too.
If we assume that
- The match starts with
${
- The tokens inside may be separated with zero or more whitespaces
- The tokens are separated
-
,*
or/
operators - The tokens are either strings of word chars (letters, digits, underscores) or double quoted string literals (
"...\"...\\..."
) - The match ends with
}
You can then use
const reg= /\${\s*(?:\w |"[^"\\]*(?:\\[^][^"\\]*)*")(?:\s*[- \/*]\s*(?:\w |"[^"\\]*(?:\\[^][^"\\]*)*"))*\s*}/g
See the regex demo.
See the JavaScript demo:
const token = '(?:\\w |"[^"\\\\]*(?:\\\\[^][^"\\\\]*)*")';
const op = '[- /*]';
const reg= new RegExp(String.raw`\${\s*${token}(?:\s*${op}\s*${token})*\s*}`, 'g');
let i= "some"
let txt= `${i " test"} RESULT ${2 4} now ${i "}" } hi`
let txtString= '`${i " test"} RESULT ${2 4} now ${i "}" } hi`'
let result= [...txtString.matchAll(reg)];
console.log(result)
console.log(txt)