Conform to IDNA version 15.1.0 revision 31

[TRowbotham]

This adds the necessary changes to conform to IDNA version 15.1.0 revision 31.

Notable Changes

• Transitional processing (the default in PHP) is now deprecated. No deprecation notices were added as PHP does not yet report a deprecation notice in this case.
• An error is no longer recorded for characters with a status of disallowed.
• When performing code point mapping and transitional processing is enabled the code point U+1E9E capital sharp s (ẞ), is replaced by the string “ss”
• A new internal option "IgnoreInvalidPunycode" was added, which is supposed to allow for an all-ASCII fast-path, however, the official tests do not test this configuration option.

[derrabus]

Can we add a test?

[stof]

I don't think we should add options that don't exist in PHP. This would defeat the purpose of this library.

[TRowbotham]

I have decided to remove the "IgnoreInvalidPunycode" option as suggested since the option is not available in PHP.

[nicolas-grekas]

Thanks for the PR.
Can you please add some tests?

[nicolas-grekas]

I fixed my CS comments by force-pushing on your fork.
Tests would still be welcome. And I have one question (see inline comments).
Are you still interested in finishing this PR?

[nicolas-grekas]

Can you explain why this change?

[TRowbotham]

The reasoning for this change according to the spec is "Checking for disallowed characters both before and after normalization yielded inconsistent results for unnormalized vs. normalized text".

[TRowbotham]

I fixed my CS comments by force-pushing on your fork. Tests would still be welcome. And I have one question (see inline comments). Are you still interested in finishing this PR?

Yes, sorry, haven't had a lot of free time lately. I'll try to put some tests together this week.

Thanks for taking care of the style issues.

[nicolas-grekas]

Thank you @TRowbotham.