Develop Reference

Regex to validate wildcard domains with special conditions

I am looking to validate wildcards against Samsung Knox Firewall. Please see below the full criteria for all domains:
A list of URLs for specified domain names to block DNS resolution. The format of the URL must be compliant with RFC's standards and must also match one of the following rules:
Full URL: "www.google.com"
Partial URL: "android.com"; "www.samsung"; "google". The
character "*" (wildcard) must be at the beginning and/or at the end
of the URL otherwise the URL is invalid.
Special case, matches any URL : "*"
Valid domains
The following examples are considered valid by Knox.
*.test.com
*test.com
*test
*test*
test.*
test1.test.*
Invalid domains
The following examples are considered invalid by Knox.
*test-
*test.
*test.com-
*test-.com
Is anybody able to offer a hand? I am struggling to accommodate for all of the requirements with this one.
Current code:
(?=^\*|.*\*$)^(?:\*\.?)?(?:(?:[a-z0-9-]+(?(?=\.)(?<!-)\.(?!-)))+[a-z]+)(?:\.?\*)?$
Edit: Actually, it looks like conditional regex may not even be supported in Java.

BASED ON YOUR PROVIDED EXAMPLES
If you're trying to pre-filter the domains, then this one matches all of your "Valid" examples and rejects all of your "Invalid" examples
^[\w*]([\w*-]+[\w*])?(\.[\w*]([\w*-]+[\w*])?)*$
If there's a file or carriage return separated field with all of these in it that you're trying to test, you may want to use the "multiline" switch like so:
(?m)^[\w*]([\w*-]+[\w*])?(\.[\w*]([\w*-]+[\w*])?)*$
since you tagged java, that would be encoded into a java string as follows:
"(?m)^[\\w*]([\\w*-]+[\\w*])?(\\.[\\w*]([\\w*-]+[\\w*])?)*$"
EDIT - Matching all the rules, in addition to your provided examples
This expression seems to work:
^(\*|(\*|\*\.)?\w+(\.\w+)*(\.\*|\*)?)$
Matching/Non-matching examples:
MATCHING NON-MATCHING
------------ ------------
* *test-
*.test.com *test.
*test.com *test.com-
*test *test-.com
*test* test*.com
test.* test.*com
test1.test.* -test.com

Velocity template localization

I'm working on a task that's using Velocity to generate an email from a template. One of my requirements is that the email message be localized, based on a Locale submitted by the user. I've got this working for the most part, using Velocity's ResourceTool to pull in a MessageBundle. The other requirement is that I can only have one template - I don't want one template per language or locale because this is difficult to maintain if changes are needed.
As a part of a MessageBundle, I can have what's referred to as "compound messages", which are messages that have variables that are replaced when the message is evaluated. For example:
dear.name=Dear {0}:
My question is this: how do I get Velocity to replace the {0} value with a name that I pass in?
It's not acceptable to simply modify my Velocity template so that I simply replace the name at the template level - in some languages, the greeting word "Dear" comes after the name, so this would yield an incorrect output.

Found the answer: in my velocity template, I can use the Velocity directive:
#evaluate(${msg.dear.name})
And in my message bundle, my message looks like:
dear.name=Dear $name
If I set the name in the Velocity context, it gets replaced as I expect.

apache commons-validator alternative for new gTLDS

I need to validate emails and domains. I just need a formal validation, no whois or other forms of domain lookup needed.
Currently I'm using apache's commons-validator v1.4.0
Unfortunately my customers use the new gTLDs, like .bike or .productions that are not yet supported by the DomainValidator class.
See Apache's Jira issue for more details.
Are there any sound alternatives that I may easily include in my Maven POM?

If you are not concerned about internationalized addresses, you could change last part of address, and continue to use Apache commons.
This approach is based on the fact that whatever the TLD is, the validity of the whole domain name is equivalent to the validity of the same domain name with the TLD replaced with com. For example:
abc.def.com is valid. Similarly abc.def.name, abc.def.xx--kput3i, abc.def.uk are valid.
ab,de.com is not valid. Similarly ab,de.name, ab,de.xx-kput3i, ab,de.uk are not valid.
So instead of calling
return EmailValidator.getInstance().isValid(userEmail);
You can call
if ( userEmail == null ) {
return false;
}
return EmailValidator.getInstance().isValid(userEmail.trim().replaceFirst("\\.\\p{Alpha}[\\p{Alnum}-]*\\p{Alnum}$", ".com"));
Explanation
The regular expression "\\.\\p{Alpha}[\\p{Alnum}-]*\\p{Alnum}$" checks for the TLD part: it's at the end of the string (because of the $), it starts with a dot and contains no other dot, and it conforms to the standards: begins with an ASCII Alpha character, followed by zero or more alphanumerics or dashes, and ends with an alphanumeric character.
I am using trim() because until now, if you used EmailValidator, it allows spaces before and after the address. Removing the spaces just makes it easier to replace the TLD, and it shouldn't matter as far as the validity of the address is concerned.
If the string doesn't have a valid TLD at the end, String.replaceFirst() will return it as is. It could still be valid, because email addresses of the format x#[n.n.n.n] where n.n.n.n. is a valid IP address are valid. So basically, if you didn't find a TLD, you let EmailValidator decide the validity issue itself.
Of course, if the TLD is not an IANA recognized TLD, this validation will not tell you that. An e-mail like david#galaxy.hoopie-frood will be accepted as legal,but IANA doesn't have that TLD as yet.
Checking a domain is similar, without the trim() part:
if (userDomain == null ) {
return false;
}
return DomainValidator.getInstance().isValid(userDomain.replaceFirst("\\.\\p{Alpha}[\\p{Alnum}-]*\\p{Alnum}$"));
I have also tried JavaMail's email address validation, but I don't really like it: it allows completely invalid domain names such as net-name.net- (ending with a dash) or IP addresses (which are not allowed for e-mail without square brackets around them), and it's only good for e-mail addresses, not for domains.
Internationalization
If you need to check for internationalized domains and e-mails, it's a bit different. It's easy to check for internationalized domains (for example 元気。テスト). All you need to do is convert them to ASCII with java.net.IDN.toASCII() (yielding xn--z4qx76d.xn--zckzah for my example domain - this is a valid TLD), and then do the same as I wrote above.
Internationalized e-mails are a different story. If the local part is ASCII, you can convert the domain part to ASCII. If you have to display the email address, you need to use the Unicode version, and if you have to send an email message, you use the ASCII version.
But recently a standard has been introduced for internationalized local parts as well, which also allows sending to the unicode version of the domain name without translating it to ASCII first. Whether you want to support that or not requires some thought, as not many mail servers and mail transfer agents support it at the moment.

Copied the implementation from DomainValidator and replaced the TOP_LABEL_REGEX expression with "\\p{Alpha}[\\p{Alnum}-]*\\p{Alpha}".
In addition, I removed validation against the hard coded list of approved gTLDs. This is, basically, quite weak in that it doesn't validate against the actual domains. But I think it's good enough (catches the gTLDs similar to XN--YGBI2AMMX).
See full list of approved gTLDs here.
// Copied from org.apache.commons.validator.routines.DomainValidator
private static final String DOMAIN_LABEL_REGEX = "\\p{Alnum}(?>[\\p{Alnum}-]*\\p{Alnum})*";
// Changed to include new gTLD - http://data.iana.org/TLD/tlds-alpha-by-domain.txt
private static final String TOP_LABEL_REGEX = "\\p{Alpha}[\\p{Alnum}-]*\\p{Alpha}";
// Copied from org.apache.commons.validator.routines.DomainValidator
private static final String DOMAIN_NAME_REGEX = "^(?:" + DOMAIN_LABEL_REGEX + "\\.)+" + "(" + TOP_LABEL_REGEX + ")$";
private static final RegexValidator domainRegex = new RegexValidator(DOMAIN_NAME_REGEX);
private static final EmailValidator EMAIL_VALIDATOR = new EmailValidator();
public static boolean isValidDomain(String domain) {
String[] groups = domainRegex.match(domain);
return groups != null && groups.length > 0;
}

What I often do in this situation is to checkout the source code for the library in question (it's open source remember?), modify it to suit my requirement, and then contribute the patch back to the project.
Your use case certainly sounds like it would be a useful contribution.

I made you a public suffix list Java API. The method PublicSuffixList.getRegistrableDomain() can be used for Domain validation:
PublicSuffixListFactory factory = new PublicSuffixListFactory();
PublicSuffixList suffixList = factory.build();
assertNull(suffixList.getRegistrableDomain("galaxy.hoopie-frood"));
assertNotNull(suffixList.getRegistrableDomain("example.bike"));

While DomainValidator is missing some of the new TLDs, for me the best solution was to update TLD.
DomainValidator.updateTLDOverride(ArrayType.COUNTRY_CODE_PLUS, new String[]{"someTLD"});
And then initiate EmailValidator Instance
EmailValidator.getInstance(false, true)

java regex matcher results != to notepad++ regex find result

I am trying to extract data out of a website access log as part of a java program. Every entry in the log has a url. I have successfully extracted the url out of each record.
Within the url, there is a parameter that I want to capture so that I can use it to query a database. Unfortunately, it doesn't seem that the web developers used any one standard to write the parameter's name.
The parameter is usually called "course_id", but I have also seen "courseId", "course%3DId", "course%253Did", etc. The format for the parameter name and value is usually course_id=_22222_1, where the number I want is between the "_" and "_1". (The value is always the same, even if the parameter name varies.)
So, my idea was to use the regex /^.*course_id[^_]*_(\d*)_1.*$/i to find and extract the number.
In java, my code is
java.util.regex.Pattern courseIDPattern = java.util.regex.Pattern.compile(".*course[^i]*id[^_]*_(\\d*)_1.*", java.util.regex.Pattern.CASE_INSENSITIVE);
java.util.regex.Matcher courseIDMatcher = courseIDPattern.matcher(_url);
_courseID = "";
if(courseIDMatcher.matches())
{
_courseID = retrieveCourseID(courseIDMatcher.group(1));
return;
}
This works for a lot of the records. However, some records do not record the course_id, even though the parameter is in the url. One such example is the record:
/webapps/contentDetail?course_id=_223629_1&content_id=_3641164_1&rich_content_level=RICH&language=en_US&v=1&ver=4.1.2
However, I used notepad++ to do a regex replace on this (in fact, every) url using the regex above, and the url was successfully replaced by the course ID, implying that the regex is not incorrect.
Am I doing something wrong in the java code, or is the java matcher broken?

How do I encode URI parameter values?

I want to send a URI as the value of a query/matrix parameter. Before I can append it to an existing URI, I need to encode it according to RFC 2396. For example, given the input:
http://google.com/resource?key=value1 & value2
I expect the output:
http%3a%2f%2fgoogle.com%2fresource%3fkey%3dvalue1%2520%26%2520value2
Neither java.net.URLEncoder nor java.net.URI will generate the right output. URLEncoder is meant for HTML form encoding which is not the same as RFC 2396. URI has no mechanism for encoding a single value at a time so it has no way of knowing that value1 and value2 are part of the same key.

Jersey's UriBuilder encodes URI components using application/x-www-form-urlencoded and RFC 3986 as needed. According to the Javadoc
Builder methods perform contextual encoding of characters not permitted in the corresponding URI component following the rules of the application/x-www-form-urlencoded media type for query parameters and RFC 3986 for all other components. Note that only characters not permitted in a particular component are subject to encoding so, e.g., a path supplied to one of the path methods may contain matrix parameters or multiple path segments since the separators are legal characters and will not be encoded. Percent encoded values are also recognized where allowed and will not be double encoded.

You could also use Spring's UriUtils

I don't have enough reputation to comment on answers, but I just wanted to note that downloading the JSR-311 api by itself will not work. You need to download the reference implementation (jersey).
Only downloading the api from the JSR page will give you a ClassNotFoundException when the api tries to look for an implementation at runtime.

I wrote my own, it's short, super simple, and you can copy it if you like:
http://www.dmurph.com/2011/01/java-uri-encoder/

It seems that CharEscapers from Google GData-java-client has what you want. It has uriPathEscaper method, uriQueryStringEscaper, and generic uriEscaper. (All return Escaper object which does actual escaping). Apache License.

I think that the URI class is the one that you are looking for.

Mmhh I know you've already discarded URLEncoder, but despite of what the docs say, I decided to give it a try.
You said:
For example, given an input:
http://google.com/resource?key=value
I expect the output:
http%3a%2f%2fgoogle.com%2fresource%3fkey%3dvalue
So:
C:\oreyes\samples\java\URL>type URLEncodeSample.java
import java.net.*;
public class URLEncodeSample {
public static void main( String [] args ) throws Throwable {
System.out.println( URLEncoder.encode( args[0], "UTF-8" ));
}
}
C:\oreyes\samples\java\URL>javac URLEncodeSample.java
C:\oreyes\samples\java\URL>java URLEncodeSample "http://google.com/resource?key=value"
http%3A%2F%2Fgoogle.com%2Fresource%3Fkey%3Dvalue
As expected.
What would be the problem with this?