Retry: Remove node's url from lib/i18n-utils

[akirk]

Retry of #38560 with a workaround for the following Safari bug (see #38668).

result = new URL( 'http://__domain__.invalid' );
result.port = '';
console.log( result.port )

Safari 13.0.4 returns: "0", all other browsers return "".

This is an invalid implementation in Safari (almost 6 year old Webkit bug 127958) since the spec says:

If the given value is the empty string, then set context object's url's port to null.

The workaround it to simply only assign changed values, and not re-assign equal ones.

[matticbot]

Test live: https://calypso.live/?branch=retry/update/remove-url-from-i18n-utils

[matticbot]

Here is how your PR affects size of JS and CSS bundles shipped to the user's browser:

App Entrypoints (~1962 bytes removed 📉 [gzipped])

name                   parsed_size           gzip_size
entry-domains-landing      -9436 B  (-1.7%)    -3529 B  (-2.4%)
entry-main                 +2436 B  (+0.2%)     +770 B  (+0.2%)
entry-login                +2436 B  (+0.2%)     +797 B  (+0.3%)

Common code that is always downloaded and parsed every time the app is loaded, no matter which route is used.

Sections (~164 bytes removed 📉 [gzipped])

name         parsed_size           gzip_size
post-editor       -551 B  (-0.0%)     -164 B  (-0.0%)

Sections contain code specific for a given set of routes. Is downloaded and parsed only when a particular route is navigated to.

Legend

What is parsed and gzip size?

Parsed Size: Uncompressed size of the JS and CSS files. This much code needs to be parsed and stored in memory.
Gzip Size: Compressed size of the JS and CSS files. This much data needs to be downloaded over network.

Generated by performance advisor bot at iscalypsofastyet.com.

[akirk]

Thanks @sgomes for taking it from here!

[sgomes]

Thank you for coming up with a fix, @akirk! 🙇

Your code looks good! I added some extra improvements as I wrote the tests and found bits missing.

[sgomes]

Note that I had to move the value to a constant to make TypeScript happy, otherwise it would complain about it possibly being undefined, even though we were checking if ( parts[ part ] && ... ) 🤷‍♂

[sgomes]

Adding Needs Review label to trigger tests.

[sgomes]

(Deleted some comments meant for another PR)

[sgomes]

The E2E failures appear to be unrelated, and the issue in #38668 does not appear to manifest in this version.