Add support for unicode soft hyphens on punctuation logic

Problem/Motivation

In some languages including german it is very common to combine words so they can become pretty long. Since not all browsers have the ability to use automatic hyphenation, especially on Windows (see https://caniuse.com/?search=hyphen) and simple text fields don't allow the usage of the HTML entity ­ a common workaround is using the unicode character for a soft hyphen (see https://en.wikipedia.org/wiki/Soft_hyphen) as a workaround.

Because pathauto treats it like a special character but doesn't make it configurable what to do with it, the soft hyphens become "real", visible hyphens (or replaced by another seperator if another is configured) in the path generation which is unintended and breaks the readability of the path.

Example

Donau­dampf­schiff­fahrts­gesell­schaft

The above word contains some unicode soft hyphens which will only become visible when the word needs to be broken apart to fit it's container

When using pathauto with this string it becomes:

donau-dampf-schiff-fahrts-gesell-schaft

Steps to reproduce

Use a soft hyphen for example in the title of a node and generate a path alias with it. It will fill in the soft hyphens with visible hyphens.

Proposed resolution

Add the option to configure how soft hyphens should be handled just like with other punctuation symbols.

Remaining tasks

Review the patch I made, for me it is working, seems like a no-brainer.

User interface changes

New option to configure how soft hyphens are handled will appear on the settings page of pathauto.

API changes

None

Data model changes

None except that the exported pathauto config will include a new "soft_hyphen" setting.