bear with me but i broke down and read the man or bear article everyone has been sharing and it really is lovely, the down-to-earth text punctuated with the author's photos that exude joy and contentedness; a+ content https://bikepacking.com/plog/man-or-bear-debate/
[at the bar] – Excuse me, I can't help but notice you were using the "make" command-line tool – ❓ – You know, I had a part in that. – 😍 – You know the -j option? – 🤓 – I did the part where it gives you a verbose error, repeated verbosely for each of your cpu cores – 🤨 – Also the part where it assumes you have infinite cores – 🤬 – 😎
thanks to a grant from nlnet / ngi zero core, i get to carve out some work time later this year to hack on the whippet garbage collector! more on this when i find how to schedule it, but i am really looking forward to the hack. https://nlnet.nl/project/Whippet/
i think a pretty ideal blog comment system is, the blog runs its own fediverse node. each post gets a toot. people can reply and hopefully read replies, though we know that fedi has a reply-visibility problem. if the blog boosts a toot reply, it can appear below the original article as a comment
with regards to programming languages: {industrial impact, being on the cutting edge of research, having fun} seems to be one of those classic "pick two, at best" situations
@rml@shegeley@cwebber@monkey1 the issue is entirely about language: is a define form essentially a creation of a mutable storage location along with an initial value? if so, AOT inlining of top-level definitions is impossible; the compiler cannot reason about the values of top-level / module-level definitions. guile moved away from this model with declarative modules.
@rml@shegeley@cwebber@monkey1 if your mental model is still “a top-level var reference necessarily loads the current value from its storage location” you need to be compiling without optimizations, otherwise you will have unpleasant surprises