Ensure crate abides by the Unicode Data Files and Software License

[lopopolo]

In practice I believe this means adding AND Unicode-DFS-2016 to the SPDX license expression.

See:

• Ensure crate abides by the Unicode Data Files and Software License artichoke/focaccia#149
• exclude non-library sources when uploading to crates.io dtolnay/unicode-ident#4 (comment)

[lopopolo]

cc @dtolnay who first uncovered this with some Google folks.

[BurntSushi]

I'm fine to add license notices and files, but I will not be re-licensing bstr itself. I've already done this dance with Google people for the regex crate years ago: rust-lang/regex#530

In essence, I brought the Unicode license notice in and added a note about it to the regex-syntax README. I do not think anything else is legally required.

The only way I'd be willing to consider changing the crate license itself is if the Rust Project itself did so. The standard library, for example, makes use of data from the Unicode Character Database. Yet, the standard library is licensed as MIT/Apache-2.0. And I can't even find any Unicode license notices at all: https://github.com/rust-lang/rust/tree/master/library/core/src/unicode

Basically, I don't see any material difference in how bstr uses the UCD versus how std uses the UCD. So I won't even consider changing the license of this crate until std does. (Or unless someone makes a compelling argument for why std is somehow different than bstr.)

Otherwise, if someone cares enough about this, I'm happy to accept a PR like the one for regex-syntax that adds the Unicode license and a short notice in the README.

[BurntSushi]

For anyone following this, see: rust-lang/rust#98116 (comment)

[BurntSushi]

Given rust-lang/rust#98116 (comment) and my own general inclination against relicensing the entire crate just because we use some Unicode data, I'm going to close this. I don't mean that in a "I just don't want to do it" sense, but in a "it's not at all clear that it is required in order to be in compliance with the Unicode consortium's license." And of course, std uses Unicode data too but doesn't include the Unicode license in its overall license.

There's also some minor risk analysis here: the Unicode license is permissive. Even if we did get this wrong in some way, the actual practical fallout for it is unlikely to ever materialize.