Bitshift operators

[chochos]

After table 6.10 in the spec, there's an open question "should we support the traditional C-style >> and << bitshift operators?"

I say we should. It's an important feature even if it's seldom used. What I don't really know is if it should be extensible or just an operator for Integral types.

[ikasiuk]

I agree. It's something people just expect to be there, that doesn't do any harm and should be fairly simple for us to provide. The problem with not having these operators is that there is no reasonable way to simulate them with other operators (not even **), so you end up with really ugly workarounds. After all Ceylon is supposed to be a general-purpose language.

[gavinking]

I have never been a fan of these things, and I recently realized that they're extremely difficult to lex/parse. The lexer can't distinguish between >> meaning bitshift and >> in Foo<Bar<Baz>>. I don't know what kind of awful hackery C++ and Java resort to to handle this, but I'm never going to go there. Closing.

[quintesse]

I can't remember right now but do we already have bit shift operations on Integer which we could optimize to java bit operations? Do we already do that perhaps?

[gavinking]

Sure. We can do whatever optimisations we like with the methods on Binary.