Accented characters in emails are failing validation

[weierophinney]

This example email used to validate fine until ZF2 2.9.0: frédé[email protected]

As far as I can see, it's supposed to be valid according to RFC 6532 http://stackoverflow.com/a/760151/1381550
In theory it should not fail validation, and what was brought in ZF2 2.9.0 is a breaking change.

Was this on purpose?

---

Originally posted by @amcsi at zendframework/zend-validator#155

[weierophinney]

Validator does support internationalized local part, previously it was passing due to IDNA false positive bug.

---

Originally posted by @Xerkus at zendframework/zend-validator#155 (comment)

[weierophinney]

@Xerkus but if the validator yields false for them, it means it clearly doesn't support them.

---

Originally posted by @amcsi at zendframework/zend-validator#155 (comment)

[weierophinney]

sorry, that was a typo. "does not"*

internationalization brings in support for utf characters where were
none, which complicates things a lot

I will try to add proper support for them, will need to do research on the
topic

--
Aleksei Khudiakov
Developer at Roave LLC
https://roave.com

On Apr 6, 2017 18:39, "Attila Szeremi" [email protected] wrote:

@Xerkus https://github.com/Xerkus but if the validator yields false for
them, it means it clearly doesn't support them.

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
zendframework/zend-validator#155 (comment),
or mute the thread
https://github.com/notifications/unsubscribe-auth/AAsTUkZlB4vwRuV2C1igcAY2Ht7WCNh6ks5rtKS6gaJpZM4MkBba
.

---

Originally posted by @Xerkus at zendframework/zend-validator#155 (comment)

[vinnyusestrict]

I'm running into the same issue. A client of mine is in Israel and her messages are failing with the bad internationalized domain error:

Fatal error: Uncaught Laminas\Mail\Exception\InvalidArgumentException: 'mhmc.co.xn--il>-5qe' is not a valid hostname for the email address in /path/to/includes/vendor/laminas/laminas-mail/src/Address.php

Are there plans for fixing this?

[vinnyusestrict]

I added some debugging and found that the issue is the preg_match on line 495 of laminas-validator/src/EmailAddress.php.

The value matched against has a Hebrew character after the closing >:

ג לא <[email protected]>ה

Replacing the regex with /^(?:.*<)?(.+)@([^@]+?)(?:>|$)/ fixes the problem. However, the original regular expression leads me to believe that it was expecting just what is inside the < and >, so perhaps a better fix would be in the header parser that pulls the address in the first place.

I hope this helps.

[gsteel]

Tests already cover this issue. It's stale and no longer relevant