Improve email regexp on edge cases

[AlekseyLobanov]

• Drastically improves performance on cases like "<" + " " * N
• Last spaces are not needed anyway because this group is stripped later. Also spaces will be caught by . anyway.

Change Summary

I found that one single change in email regexp solves slowdowns on special invalid email strings. See related issue for details

Related issue number

Fixes #10600

Checklist

• The pull request title is a good summary of the changes - it will be used in the changelog
• Unit tests for the changes exist
• Tests pass on CI
• Documentation reflects the changes where applicable
• My PR is ready to review, please add a comment including the phrase "please review" to assign reviewers

Selected Reviewer: @sydney-runkle

[AlekseyLobanov]

please review

[codspeed-hq]

CodSpeed Performance Report

Merging #10601 will not alter performance

_{Comparing AlekseyLobanov:fix.email-regex (e35c507) with main (c772b43)}

Summary

✅ 38 untouched benchmarks

[AlekseyLobanov]

How performance changes?
I use my own POC in #10600 and run it as /usr/bin/time python pydantic-poc.py 500

• Before: 5.35user 0.01system 0:05.37elapsed 99%CPU (0avgtext+0avgdata 36840maxresident)k
• After: 0.20user 0.01system 0:00.22elapsed 99%CPU (0avgtext+0avgdata 37004maxresident)k

About 25x speed improvement.

[github-actions]

Coverage report

This PR does not seem to contain any modification to coverable code.

[Viicos]

Thanks, this looks reasonable but to be extra careful we'll wait for other reviews as well.

Could you add the following to the test_address_valid test?:

        ('Samuel Colvin < s@muelcolvin.com>', 'Samuel Colvin', 's@muelcolvin.com'),
        ('Samuel Colvin <s@muelcolvin.com >', 'Samuel Colvin', 's@muelcolvin.com'),
        ('Samuel Colvin < s@muelcolvin.com >', 'Samuel Colvin', 's@muelcolvin.com'),

[AlekseyLobanov]

Thanks, this looks reasonable but to be extra careful we'll wait for other reviews as well.

According to Wikipedia it is one of the valid DoS attack vectors. And at least some of known to me rate limiters will work only after validation step.

Could you add the following to the test_address_valid test?

I think that existing tests are already covering this edge cases (spaces before/after the group). Should I still add yours?

        ('foo BAR <foobar@example.com >', 'foo BAR', 'foobar@example.com'),
        ('FOO bar   <foobar@example.com> ', 'FOO bar', 'foobar@example.com'),
        ('Whatever < foobar@example.com>', 'Whatever', 'foobar@example.com'),

[sydney-runkle]

Yep let's add those extra tests and fix the lints, but otherwise LGTM.

[Viicos]

According to Wikipedia it is one of the valid DoS attack vectors. And at least some of known to me rate limiters will work only after validation step.

I agree, just wanted to be careful as changing regex can be a source of breaking changes.

I think that existing tests are already covering this edge cases (spaces before/after the group). Should I still add yours?

Missed these ones, then maybe only add these ones after it:

        ('Whatever <foobar@example.com >', 'Whatever', 'foobar@example.com'),
        ('Whatever < foobar@example.com >', 'Whatever', 'foobar@example.com'),

[AlekseyLobanov]

Missed these ones, then maybe only add these ones after it:

Done.

[sydney-runkle]

Nice, thanks for the help here! We appreciate the thorough explanations / refs :).