Ouch .. something I didn't think of: with a compact representation of types like 3 * 2 in a single integer .. the components aren't addressable. And so not assignable in principle. Same deal as bitfields in C.
I can fudge it in some cases with a read/modify/write. the case that actual fails in my tests is fixable: var i : 3 = case 1 of 3; var j : 3 = case 2 of 3; var ij : 3 * 3 = i,j; At present Felix tries to assign the components of ij separately like: ij.0 = i; ij.1 = j; which fails. There's a related issue: tuple coercions like: var xyz = i,j,"Hello"; var xy = xyz :>> 3 * 3; // chop of trailing string can work but require repacking the two leading components into a single integer. But the corresponding pointer coercion cannot work because the representation changes when a tuple is a linear type. This is .. well its a mess :) Originally i only proposed to change array indexing. Note that assignment can be fixed if you consider only assigning to whole variables but use "functional update" on the RHS. This allows the compiler to optimise to separate component updates when possible but otherwise it just rebuilds the new tuple and assigns the whole thing. -- john skaller skal...@users.sourceforge.net http://felix-lang.org ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ _______________________________________________ Felix-language mailing list Felix-language@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/felix-language
