Eric Holk
Right now Harlan programs can't do much more with strings than pass them around. Even just adding facilities to convert between C strings and vectors of chars would greatly improve...
If a while loop relies on side effects in the test it will never terminate. This is because Harlan assumes expressions are pure and lifts it out. We should probably...
Aaron Clarke was asking about how to get input in Harlan programs. Currently there's not a lot of support for this, although you might be able to get some done...
This is non-deterministic, but we're starting to see it in a couple of test cases and programs now (e.g. #147). One that shows it pretty reliably is `test/read-file.kfc`. Here is...
Right now we have no way of freeing strings once they've been allocated. We don't do a lot of dynamic string creation yet, so this hasn't caused major problems yet,...
We should add a special region inside of kernels that lives in local (or shared) memory. Kernels could use this to allocate data that isn't actually returned from the kernel.
It'd be nice if `define` could define more than just functions. Top-level constants are handy in a lot of cases. It would also be nice to be able to import...
`mandelbrot.kfc` generates just a white image. `mandelbrot2.kfc` and `mandelbrot3.kfc` don't even seem to make valid image files. It'd be good to store a reference image and compare the generated result...
See, for example, this build: https://travis-ci.org/eholk/harlan/builds/21026702 When I run the build at home in petite, all the tests "pass," but running the ones that failed in Travis under Vicare displays...
Instead of representing vectors as a pointer to a length field, we would represent them as a pair of a length and a pointer to the data. We pay a...