Notices by cj 🇺🇸🇨🇭 (cj@mastodon.technology@mastodon.technology)

@cwebber Like "Spot", or "Lassie", or "Hedwig", or "3g2upl4pq6kufc4m"?

@cwebber I appreciate the sincere answer and added it to my "to-read" pile but I also didn't intend for my playful joke to be so lame to not even get a chuckle. 😭

@absturztaube admittedly I don't know much about GNS. I think GNS uses DNS.

I would prefer avoiding DNS altogether, which also has its challenges as the last 30 years or so has seen more money, labor, and effort generally put into DNS and DNS dependent technologies.

I like imagining a world where ActivityPub does not rely on DNS

Whew, and now that this is over, I wrote down on one side of an index card all the things I'd like to tackle, thinking that an index card would keep my list small.

Turns out I can still write small and I have 8 things I want to do.

I finally received the counter-signed copies in the mail: I have successfully renegotiated the IP portion of my contract. 1.5 years of work, lawyers, negotiation, meetings, concluded. Cleared up a lot of ambiguity.

Expect more #activitypub related developments in the near future.

@emacsen @cwebber Whoa, stop the world! Didn't realize I made a copy of Chris' joke.

@emacsen @cwebber This thread's marked for sweeping up.

@cwebber I think you're excellent. I don't know Racket but your code is easy to grok, honest about the TODOs, and occasionally makes me chuckle in delight.

@gudenau Well now you know that `~Interface() = 0` will map to `~Implementation() override {...}`. Is the error you are still getting a runtime one? Any chance for a paste of the error message?

@gudenau Anyway, I won't be up much longer. I hope you get your problem solved.

A TL;DR rule-of-thumb for template metaprogramming in C++: "If it touches a <T> parameter/return-value/member, define it inline in the header (includes constructor & destructor)".

This isn't the whole story and there's numerous implications for this rule-of-thumb but it is the shortest way to the goal of "code compiles and runs" (just not necessarily quickly nor well in either case).

@gudenau If I hazard a guess it is because you are using template metaprogramming and need to define the body of the destructor in the header in order for the compiler to correctly generate the destructor for a template of type T (non-specialized).

@gudenau It also needs to be public, to be polymorphic.

@gudenau Doesn't `~Interface<Thing*>()` need its own [pure] virtual destructor, for it to then even be able to find `~Implementation<Thing*>()` ?

@gudenau Next, is the `~Implementation` marked with `override` and compiles correctly?

@gudenau It's a keyword since C++ 11 and I highly recommend it for inheritance since it will force compiler errors when it doesn't override a base class method (which is what you usually want). Tracking this down pre-C++11 sucked:https://en.cppreference.com/w/cpp/language/override

@xj9

Yes, and it is well known that Drew (the creator of sourcehut) is very pro-email-git and is very-against-ForgeFed.

I'm the guy that debated him in the ForgeFed archives. I still have a blog post I need to write up around this whole "git is already federated" non-argument.

@bhaugen @dansup @jauntywunderkind420 @bhaugen @fr33domlover @cwebber

@xj9

It may be because I'm a dim individual, but I don't really understand what you're trying to say here.

I have nothing against git+email, please keep using it and supporting Drew via using sourcehut!

@bhaugen @dansup @jauntywunderkind420 @bhaugen @fr33domlover @cwebber

Delaunay Triangulation has a lot of cool properties:- Getting from point A to B has an upper bound distance of ~2.4x the distance "as the crow flies"- Is the dual graph of a Voronoi diagram.- If produced in a spatial Poisson Process, on average a point will have 6 neighbors.