IDN Support suggestion

[wadih]

It seems any IDN can't get past the regexp in regDomain.class.php:

if (!preg_match("/^([a-z0-9])(([a-z0-9-])*([a-z0-9]))*$/", $domPart)) return FALSE;

If anybody's interested, the library could be augmented to support UTF8 letter characters by adding the \p{L} character set along with a-z0-9 and adding /u at the end:

if (!preg_match("/^([a-z0-9\p{L}])(([a-z0-9-\p{L}])*([a-z0-9\p{L}]))*$/u", $domPart)) return FALSE;

I tested it as such and it worked:

echo (new regDomain())->getRegisteredDomain("example.мон", false);

And it works for that IDN (.xn--l1acc):

example.мон

I could submit a pull request if nobody sees an issue

[wadih]

I found an issue on above regexp, that validation regexp doesn't appear to match for Indian tld's like, although 99% of the rest has worked.

ಭಾರತ     xn--2scrj9c
ଭାରତ     xn--3hcrj9c
ভাৰত     xn--45br5cyl
भारतम्    xn--h2breg3eve

I tried varying the regexp but not finding one that works, using \p{Devanagari} without success.

Maybe will have to go through the ascii domain instead to avoid these vocabulary challenges.

[usrflo]

@wadih, thanks for your feedback and research: in 9fccafa I simply disabled the domain label length validation and the regexp to prevent false analysis.

I tend to change the processing so there is a second suffix tree being encoded to ACE and all checks including length validation and character validation are done in ASCII. Downward compatibility should be kept.