IRC channel logs
2025-02-10.log
back to list of logs
<dthompson>ACTION is back to work after a post-fosdem break <dthompson>jfred: cool that you are trying the persistence stuff! <dthompson>let us know what you think once you've used it enough to have opinions :) <dthompson>seem dckc has been making his way through the FOSDEM videos <dthompson>more of those should become available this week <jfred>I'd have to go back and try to put together a reproduction, but I later got a more descriptive error message like "don't know how to persist this object" - I'm not sure why I didn't get that message the first time. (Maybe it had something to do with being unable to persist the root object vs. one further down in the graph) <jfred>That would have helped me track down the problem earlier though <dthompson>I think there's still some work to do to make it easy to get started with persistence and find issues <cwebber>It's a combination of warding/incanting and the pebble bank <cwebber>(pumpkin-soup is going to be a repo we put some experiments/prototypes in over the next few weeks before kicking off our next round of applications using our tech) <cwebber>And no that's not an NFT ;p well not in the money making scheme way. I guess it might be per the technical definition. That's not the intended use case though, this is meant to be "changing who can control something useful" <cwebber>The interesting thing is that the idmint becomes the public Id to "talk about" the object <jfred>the ^proxy body being a no-op function is a little funny. Think I'll have to stare at this a little more haha <cwebber>It's a pebble which enables encanting <jfred>oh that makes sense, you're just using it for the equality check <cwebber>The names could probably be better :) <tsyesika>There’s a version which uses warding and incanting ;) <cwebber>Yes it's true. But it's not clear it's clearer yet :) <jfred>Huh, yeah that's clever. So even if you handed someone a reference to the idmint the only thing they could practically do without the proxy is get the description <cwebber>The funny thing is it's a mint with one pebble <cwebber>This came out of a group conversation at a restaurant after FOSDEM with the Spritely engineering team :) <jfred>and I suppose, hinted at by the fact that the argument names are different, the xchange-proxy and use-proxy don't have to be the same proxy. and then it's a mint with two pebbles :P <cwebber>"could you use the pebble bank to do inventory management for unique items without using $" <cwebber>jfred: there's only ever one pebble that's *valid* <cwebber>A person handing control retains the old now-invalid pebble though <cwebber>They just can't use it for anything once exchanged <cwebber>I think the variables could use renaming <cwebber>tsyesika and I figured this out on a 2 hour long call at 7am my time and then pared it down. I think it's a neat design <jfred>I might need to think about this some more but what I was kind of getting at is that I think you might be able to split it up such that you could hand someone a pebble for usage without necessarily handing them the right to transfer ownership of the object away from you <jfred>if for example you added a second proxy for that purpose in the ^slicer-idmint argument list <jfred>I'm not sure if/when that might be useful <jfred>of course in that case the act-proxy survives exchanges and you'd probably also want a way for the holder of the owner-proxy to change that, if you were to actually do something like this <tsyesika>yeah it's interesting, I think you'd want to be able to revoke the act proxy. I think maybe individual revocable proxies might be a better way to achieve this given it'd allow tracking and revocation on a person/capability by person/capability basis <tsyesika>but it's interesting to split the idea of using it and ownership up <binarydigitz01>I am using nixos, and i was trying to get hoot and goblins running on master branch of guile, but goblins fails some tests, is that natural or a nix thing? <tsyesika>strange, these shouldn't be erroring like that... <tsyesika>btw I notice the path is goblins 0.14, it's only 0.15 which works with hoot <binarydigitz01>Hmmmm now that I'm trying goblins 0.15, i'm getting a new error: <binarydigitz01>checking if (web socket server) is available... no configure: error: required guile module not found: (web socket server) <tsyesika>it sounds like it's missing the guile-websocket library <binarydigitz01>Yeah looks like this isn't packages for nixos yet, I have some work to do thanks :D <tsyesika>no problem :) would be great to get that into nix <tsyesika>for what it's worth those errors in 0.14 aren't known to us and it should have worked so I'm not convinced getting 0.15 will fix them, but I thought I should mention it <tsyesika>not quite sure what's causing them to be honest... <binarydigitz01>I'll definitely look into it more, should I open this as an issue on gitlab? <tsyesika>yeah sure. I think guix has an equivalent of guix.scm file which includes the build instructions for a project right? It'd be good to attach that so we could try and reproduce it in the same environment and see if it's goblins or nix