Nikita Kniazev

Do you expect it to match [synoglyphs/homoglyphs](https://en.wikipedia.org/wiki/Homoglyph), [duplicate characters in Unicode](https://en.wikipedia.org/wiki/Duplicate_characters_in_Unicode) and other Unicode quirks that I am not aware of?

Duplicate of #487. The patch that added `range_attribute` is a workaround. I was looking recently into the issue and I will almost certainly will replace `move_to` overload that accepts iterators...

It would be nice if we can stop using `boost::iterator_range` or at least replace it with something else (because including it includes most of the Boost.Range library), but it cannot...

> I did finally figure that out today; I had to add a &* to the iterator (making the construction of the std::string_view into attr = { &*saved, typename Attribute::size_type(first...

Just open a PR if you have a fix. Do not forget about tests.

Don't know variant or single element tuple or both, but not x3::symbols: ```cpp #include #include #include namespace ast { struct tagged_token { int token; }; } BOOST_FUSION_ADAPT_STRUCT(ast::tagged_token, token) int main()...

What boost version do you use? The phoenix error should have been fixed by https://github.com/boostorg/phoenix/pull/58 (shipped with boost 1.67). The `conjure3` example (from what I see) uses LLVM JIT API...

Almost duplicate of #652. V2 docs/examples are not on my priority list. PR is welcome.

I have described the bug in #610, but if it was my will I just removed ability to create semantic actions on alternative parser, especially without an actual attribute. Or,...

Thanks for the short reproducer. Though the example indeed regressed with https://github.com/boostorg/spirit/pull/702 it just exposed a bug in `is_substitute`, which incorrectly handles a case when both parameters are variants https://github.com/boostorg/spirit/blob/abd946b00d8726f1dd586860b2f191b202ff105a/include/boost/spirit/home/x3/support/traits/is_substitute.hpp#L73-L79...