Groovy regex PatternSyntaxException when parsing GString-style variables - regex

Question

Groovy here. I'm being given a String with GString-style variables in it like:
String target = 'How now brown ${animal}. The ${role} has oddly-shaped ${bodyPart}.'
Keep in mind, this is not intended to be used as an actual GString!!! That is, I'm not going to have 3 string variables (animal, role and bodyPart, respectively) that Groovy will be resolving at runtime. Instead, I'm looking to do 2 distinct things to these "target" strings:
I want to be able to find all instances of these variables refs ("${*}") in the target string, and replace it with a ?; and
I also need to find all instances of these variables refs and obtain a list (allowing dupes) with their names (which in the above example, would be [animal,role,bodyPart])
My best attempt thus far:
class TargetStringUtils {
private static final String VARIABLE_PATTERN = "\${*}"
// Example input: 'How now brown ${animal}. The ${role} has oddly-shaped ${bodyPart}.'
// Example desired output: 'How now brown ?. The ? has oddly-shaped ?.'
static String replaceVarsWithQuestionMarks(String target) {
target.replaceAll(VARIABLE_PATTERN, '?')
}
// Example input: 'How now brown ${animal}. The ${role} has oddly-shaped ${bodyPart}.'
// Example desired output: [animal,role,bodyPart] } list of strings
static List<String> collectVariableRefs(String target) {
target.findAll(VARIABLE_PATTERN)
}
}
...produces PatternSytaxException anytime I go to run either method:
Exception in thread "main" java.util.regex.PatternSyntaxException: Illegal repetition near index 0
${*}
^
Any ideas where I'm going awry?

Answer 1

The issue is that you have not escaped the pattern properly, and findAll will only collect all matches, while you need to capture a subpattern inside the {}.
Use
def target = 'How now brown ${animal}. The ${role} has oddly-shaped ${bodyPart}.'
println target.replaceAll(/\$\{([^{}]*)\}/, '?') // => How now brown ?. The ? has oddly-shaped ?.
def lst = new ArrayList<>();
def m = target =~ /\$\{([^{}]*)\}/
(0..<m.count).each { lst.add(m[it][1]) }
println lst // => [animal, role, bodyPart]
See this Groovy demo
Inside a /\$\{([^{}]*)\}/ slashy string, you can use single backslashes to escape the special regex metacharacters, and the whole regex pattern looks cleaner.
\$ - will match a literal $
\{ - will match a literal {
([^{}]*) - Group 1 capturing any characters other than { and }, 0 or more times
\} - a literal }.