Make Owner addressable via RFC 2822 standard

[noomorph]

Context

I find it inconvenient to have owner just as a non-actionable label – especially in environments where a person has different names in real life, email, Github and Slack/Discord.

Therefore, I'd like to suggest improving the rendering and making it follow Internet Message Format, e.g.:

John Doe <[email protected]>
John Doe <https://github.com/john.doe>

and also be able to consume plain e-mail and URL addresses:

[email protected]
https://github.com/john.doe

In the case when the owner is not a valid URL or e-mail, the rendering stays the same, e.g.:

johndoe

Checklist

• Sign Allure CLA
• Provide unit tests

[noomorph]

@baev, I had time this morning to work on your suggested changes, and I hope the PR looks good enough now.

I verified my change manually end-to-end on localhost:

[noomorph]

@baev hi, just making sure you're aware that I have fixed the code according to your review.

[baev]

thx, I'm aware. Review is WIP, will try to finish it this week

[A1exKH]

Code LGTM.

Added files:
allure-generator/src/main/java/io/qameta/allure/owner/OwnerAddress.java
allure-generator/src/main/java/io/qameta/allure/owner/OwnerAddressParser.java
allure-generator/src/test/java/io/qameta/allure/owner/OwnerAddressParserTest.java

Modified files:
allure-generator/src/main/java/io/qameta/allure/owner/OwnerPlugin.java
allure-generator/src/main/javascript/plugins/testresult-owner/OwnerView.hbs

[A1exKH]

Tests look good.
For better test coverage it would be good to add a test for invalid URL and invalid email.

[noomorph]

@A1exKH, I added

    implementation("commons-validator:commons-validator:1.7")

to Allure Generator build.kts to make it more resilient when it comes to validating URLs and emails. The remaining regexp for angle brackets parsing has been made a bit stricter too.

Please review the updated tests and implementation again. 🙏

[noomorph]

Empty spaces in the angle brackets are considered normal thing by the standard.

[noomorph]

This thing is important to have for plain text. On 10,000 "John Doe {{counter}}" addresses, it eats them in 20ms, just like without parsing (16-18ms).

Without this optimization, it takes up to 200ms on my machine to parse 10,000 addresses.

I measured the performance between my previous simplistic regexp implementation and the actual implementation with Apache validators, and the impact was negligible — ±20-30ms (sometimes faster, sometimes slower), so I hope this is totally okay for you to accept this PR with industry-proven validators for emails and URLs instead of improvised ones by me.

[A1exKH]

Code LGTM.