14:34:58Bikedrmeister: there's already some stuff for function descriptions in functor.cc, but it treats them as void* for some reason? instead of using the structure. is that how it worked before the structure was defined?
14:35:15KevslingerWhat is Heisig doing in America?
14:35:35Bikehe's at a conference for array programming.
15:11:51Bike"Malongains a callback to a Python function. Also malongains two doubles and two longegers that can be passed back and forth"
15:12:04Bikewhat a comment. maybe a search replace snafu?
17:08:37Bikelooks like we might not need cmpname any more, since asdf lost the dependency
17:39:12drmeisterBike: I first created the %function-description% and %function-description*% types in llvm and I had not exposed them in C++ - so I passed %function-description*% pointers around as a void* . Please replace any (void*) with core:FunctionDescription*
17:40:02drmeisterI replaced many of the void* with FunctionDescription* - I guess I missed a few.
17:40:40Bikethere's a couple methods that return a "function description" that's multiple values
17:41:55Bikelooks like swankclasp refers to things in c: forsome reason
17:42:01Bikewe don't even have a "c::compiler-note"
17:42:17Bike"CLASP emits lots of noise in compiler-notes" is news to me
17:44:02BikeDo you think we could have a source-location structure in the runtime? We could probably add source info to classeslike immediately except I don't want to have four slots