Android: Matcher.find() never returns

First of all, here is a chunk of affected code:

// (somewhere above, data is initialized as a String with a value)
Pattern detailsPattern = Pattern.compile("**this is a valid regex, omitted due to length**", Pattern.DOTALL | Pattern.CASE_INSENSITIVE);
Matcher detailsMatcher = detailsPattern.matcher(data);
Log.i("Scraper", "Initialized pattern and matcher, data length "+data.length());
boolean found = detailsMatcher.find();
Log.i("Scraper", "Found? "+((found)?"yep":"nope"));

I omitted the regex inside Pattern.compile because it's very long, but I know it works with the given data set; or if it doesn't, it shoudn't break anything anyway.

The trouble is, I do get the feedback I/Scraper(23773): Initialized pattern and matcher, data length 18861 but I never see the "Found?" line, it is just stuck on the find() call.

Is this a known Android bug? I've tried it over and over and just can't get it to w开发者_运维技巧ork. Somehow, I think something over the past few days broke this because my app was working fine before, and I have in the past couple days received several comments of the app not working so it is clearly affecting other users as well.

How can I further debug this?

Some regexes can take a very, very long time to evaluate. In particular, regexes that have lots of quantifiers can cause the regex engine to do a huge amount of backtracking to explore all of the possible ways that the input string might match. And if it is going to fail, it has to explore all of those possibilities.

(Here is an example:

regex = "a*a*a*a*a*a*b";         // 6 quantifiers
input = "aaaaaaaaaaaaaaaaaaaa";  // 20 characters

A typical regex engine will do in the region of 20^6 character comparisons before deciding that the input string does not match.)

If you showed us the regex and the string you are trying to match, we could give a better diagnosis, and possibly offer some alternatives. But if you are trying to extract information from HTML, then the best solution is to not use regexes at all. There are HTML parsers that are specifically designed to deal with real-world HTML.

How long is the string you are trying to parse ? How long and how complicated is the regex you are trying to match ?

Have you tried to break down your regex down to simpler bits ? Adding up the bits one after another will let you see when it breaks and maybe why.

make some RE like [a-zA-Z]* pass it as argument to compile(),here this example allows only characters small & cap.

Read my blogpost on android validation for more info.

I had the same issue and I solved it replacing all the wildchart . with [\s\S]. I really don't know why it worked for me but it did. I come from Javascript world and I know in there that expression is faster for being evaluated.