分割逗号分隔的正则表达式输入字符串

Question

Given a comma separated regex string, for example:

^test\d+.txt$, ^[-,0-9]+$ // provided by user

I want to split it by a comma and get

• ^test\\d+.txt$
• ^[-,0-9]+$

instead of

• ^test\\d+.txt$
• ^[-
• 0-9]+$

How can I achieve this?

Answer 1

As you can tell from the comments , is both a valid symbol within a regex and depending on the favor a valid delimiter in it's own right (for example in a qualifier). As such it's not possible to separate regex's delimited by ','. I suspect it would not matter what character you used there would still be ambiguity between the symbol chosen as a delimiter and an expression that contains that character. In your example you comment that the input is provided by user . If you have control over this input process you can might be able to change the delimiter to something different, for example if you know that only one language is in play you can use a unique character from a different language (assuming code is inserting the character and not the user directly). Of course the best solution is to pass this information in an array (one regex per element).

In the end it depends on a lot of things. Regex's are vulnerable to problems like infinite backtracking, which can be a problem when you are allowing arbitrary regex processing. Again if you have control of the input process you might consider providing users a simpler template like language, so input like "AAAAA 99" (by user) could be converted to "[\\w]{5} [\\d]{2}" by code. This is simpler for the user and allows you more control over how much chaos user input can create.

If you can not modify/control the input process, then you should teach your users to code regular expressions, perhaps by pointing them to the quick reference . So they can create a single expression that does everything they want.

Anytime you are running a unknown regex you should use the methods that allow you to provide a timeout in order to protect against regex's that longer than you want to wait to finish.

Summary. There's no good answer. There is ambiguity between what you chose as a delimiter and what a valid regex can contain. Your best bet is to change form of the input or the behavior of the user.