JavaScript’s regex engine isn’t the only one to have these problems. There certainly are other implementations, like Re2 and Rust’s implementation, that don’t have this issue. But they also lack some of the features of the JS implementation too.
Regular expressions are great and can always be matched in linear time with respect to the input string length.
The problem is that JS standard library RegExps aren’t actually regular expressions, but rather a much broader language, which is impossible to implement efficiently. If RegExp switched to proper regular expressions, they would match much faster but supporting backreferences like /(.*)x\1/ would be impossible.
If you insist on the definition as it is in formal language theory.
In practice regex is widely used to mean the pattern matching thing that also supports back references.
Wikipedia suggests using the term “regular expressions” for the language theory thing and “regex” for the programming language (PCRE) thing. I agree and would even go further and say that any time one wants to refer to the concept as it is used in formal language theory they should explicitly specify that they are talking about the theoretical concept, not the regex implementation as it is found in most programming languages.
Am I the only one shocked to learn that to find something at the end of a string it starts at the beginning? Perhaps it’s because of the simplicity of the example but I expected it to start at the end.
Add comment