Back from break today. I have a handful of miscellaneous stuff that I need to do today, so I don't expect to see much progress with regards to my allocation reuse. Oh, and ICFP is on, so I'll try to follow at least some of the presentations there. I will try to get back in flow however.

In general, this is the last week before classes start again, and once that happens I don't expect to have much time for my research, for at least a couple of months. Most of my time will be taken up by teaching and attending courses myself. Therefore, I expect this week to be a bit here-and-there, as I try to wrap up what I've been working on recently (or at least leave it in a state so that it's easy to pick up again when time permits), prepare for my own course load and teaching duties, and tick off a number of administrative tasks as well.


I don't like nix. More precisely, I think, I don't like nixpkgs. I'm not sure, because the language Nix, nixpkgs, and NixOS are so entwined that it's sometimes impossible to know where one ends and another continues. Especially when reading manual pages, guides and tutorials. Ugh.

I've tried all day to get the nix derivation for my sexp branch of Futhark too use haskellPackages.sexp-grammar_2_2_1 instead of haskellPackages.sexp-grammar, but nothing seems to work. Error messages are absolutely opaque, and so are the functions and tools used within nixpkgs. There is no documentation, no manual, no help to get anywhere, and noone seems to know how anything works at all.