Allow multiemail input for ServiceGroup EMAIL

[nikoloutsa]

For our use-case it would make sense to allow more than one emails for ServiceGroup contact , since it's not a person it's a group.

• test1@mail.com valid

• test1@mail.com;test2@mail.com valid

• test1@mail.com; invalid

• test1@mail.com;test2@mail.com; invalid

  modified: config/gocdb_schema.xml
  modified: htdocs/web_portal/views/service_group/edit_service_group.php

[gregcorbett]

* [test1@mail.com](mailto:test1@mail.com);[test2@mail.com](mailto:test2@mail.com) valid
* [test1@mail.com](mailto:test1@mail.com);[test2@mail.com](mailto:test2@mail.com) invalid

What is the difference between these two cases? To me, they look the same.

[jrha]

Slightly off-topic but maybe something to consider for @gregcorbett as a later improvement: The base regex may be too restrictive as it definitely does not allow for Unicode characters in addresses, which may or may not be allowed anyway.

It might also be worth seeing whether PHP's built-in regex filter would be good enough:

filter_var($email, FILTER_SANITIZE_EMAIL);

The HTML5 spec says:

/^[a-zA-Z0-9.!#$%&'*+\/=?^_`{|}~-]+@[a-zA-Z0-9](?:[a-zA-Z0-9-]{0,61}[a-zA-Z0-9])?(?:\.[a-zA-Z0-9](?:[a-zA-Z0-9-]{0,61}[a-zA-Z0-9])?)*$/

So… 🤷‍♂️

[tofu-rocketry]

Looks ok. Just that question over the lazy matching (+?).

[gregcorbett]

Slightly off-topic but maybe something to consider for @gregcorbett as a later improvement: The base regex may be too restrictive as it definitely does not allow for Unicode characters in addresses, which may or may not be allowed anyway.

It might also be worth seeing whether PHP's built-in regex filter would be good enough:

filter_var($email, FILTER_SANITIZE_EMAIL);

The HTML5 spec says:

/^[a-zA-Z0-9.!#$%&'*+\/=?^_`{|}~-]+@[a-zA-Z0-9](?:[a-zA-Z0-9-]{0,61}[a-zA-Z0-9])?(?:\.[a-zA-Z0-9](?:[a-zA-Z0-9-]{0,61}[a-zA-Z0-9])?)*$/

So… 🤷‍♂️

Feel free to open an issue :)

[tofu-rocketry]

@jrha opened an issue (#147). Do we just need to sort the lazy matching out?

[gregcorbett]

@jrha opened an issue (#147). Do we just need to sort the lazy matching out?

And what the difference between these two test cases is suppose to be. @nikoloutsa?.

[nikoloutsa]

the difference is not to allow a semicolon at the end of the mail field.

It looked cleaner to me when someone inputs just a single email, to trim out the last semicolon as it was previously without my addition.

[tofu-rocketry]

the difference is not to allow a semicolon at the end of the mail field.

@nikoloutsa, There doesn't seem to be a semicolon at the end of the last test case though. I can see that there is a semicolon at the end of the invalid case that only has one email address, so was that just a mistype for the invalid two address case?

[GRyall]

As discussed above I have removed the lazy matching. Following in person discussions with the devlopment team I have also reworked the regex to allow trailing semicolons. @gregcorbett this is ready to be merged!