regex - Firefox length limit in regular expression (is it a bug?) -

i'm debugging problem in application uses regular expression validate emails clientside (yeah know, both kinda stupid) , problem got me stumped.

the thing validation works fine in chrome fails in firefox , wonder if bug or if there's wrong regular expression causes error.

please check fiddle complete test case: http://jsfiddle.net/kqvgj/

new regexp(/^\s+([\_\-\.]*\s+[\_\-]?)*@\s+([\_\-]?\s+)*\.+([\-\_]?\s)+(\.?\s+)*$/); 

in firefox regex above matches mw@thisissometest.de not mw@thisissometestbutlong.de.

it seems fail based on length of input alone, there no length restriction in expression @ all!?

the fault in regex: it's pathologically inefficient. basically, you've got multiple consecutive parts can match same characters, controlled open-ended quantifiers (* , +). creates astronomical number of "paths" regex has check before gives on match. in fact, sort of problem becomes apparent when no match possible, you've managed trigger on regex should match.

i suspect trying this:

/^[a-z]+(?:[_.-][a-z]+)*@[a-z]+(?:\.[a-z]+)*$/i 

before starts criticizing, know [a-z]+ no more correct \s+. i'm trying explain what's wrong regex. idea force user name , domain name start letters, while allowing them separated chunks around delimiters ., -, , _. that's makes complicated

the important feature of regex moves forward. when [a-z]+ runs out of letters consume, next thing sees must 1 of delimiter characters, at-sign ('@'), or end of string (depending on part of address it's matching). if doesn't see expects, match attempt fails immediately.

in regex \s+ part gobbles whole string, starts backing off 1 character @ time give next part chance match. process repeated every \s+. hamza observed, that's regex engine spends of time. it's not \s+ alone that's killing you, it's structure of regex.