IRC channel logs

<encryptedwhispe->i have discovered guile-next-next in the package def and am trying that

<dthompson>encryptedwhispe-: yeah you need a recentish guile that has the pure scheme srfi-1

<encryptedwhispe->I discovered savannah is dead and made a PR to switch to codeberg repo for Hoot, dthompson

<encryptedwhispe->and I see you've already merged :)

<dthompson>yes, thank you for that!

<encryptedwhispe->dthompson, the version that is in guix.scm - does that have the pure scheme srfi-1 ? I am still getting the error about load-extension.

<dthompson>I'll check...

<dthompson>yes, it does

<encryptedwhispe->hrm

<encryptedwhispe->dthompson, is the error ERROR: In procedure %resolve-variable: / Unbound variable: load-extension a symptom of not having the pure scheme srfi-1 ?

<encryptedwhispe->trying to figure out if my guix is being weird

<dthompson>yes, that would indicate an older srfi-1

<dthompson>perhaps your environment is still pointed at an older guile

<dthompson>try rerunning ./configure if you haven't

<encryptedwhispe->hrm I don't have a configure for my project, just a simple makefile, and I am building via guix shell. But perhaps there's an issue with cache

<dthompson>oh sorry I thought you were running from the hoot repo

<dthompson>I don't know what might be going on but somehow you're not pulling modules from the right guile

<encryptedwhispe->I suspect my makefile is at fault

<encryptedwhispe->i'll take a look at the game jam templates

<encryptedwhispe->yeah that's it. guild compile-wasm -L $(pkg-config --variable=pkgdatadir guile-3.0)/3.0 -L modules -L guile-scene/modules --bundle -o boop.wasm boop.scm

<encryptedwhispe->pkg-config doesn't return the shell's module dir

<dthompson>you could try running a simple program with hoot first. if you're using the latest hoot from main, you shouldn't need that -L pkg-config thing anymore

<dthompson>a simple srfi-1 program: (use-modules (srfi srfi-1)) (fold + 0 '(1 2 3 4))

<dthompson>run with: guild compile-wasm --run test.scm

<encryptedwhispe->there we go, yeah it was the pkg-config line

<encryptedwhispe->nice that it's far simpler now!

<dthompson>yeah one of those little things that took forever to get around to

<encryptedwhispe->is there a good way to prevent WARNING: (% %library-group108 boop client): `let' imported from both (% %library-group108 srfi srfi-71) and (% %library-group108 guile), dthompson?

<dthompson>encryptedwhispe-: that's a bug that I've been investigating. for now, avoid srfi-71

<dthompson>tracking issue here https://codeberg.org/spritely/hoot/issues/342

<encryptedwhispe->gotcha!

<encryptedwhispe->will see if I can use srfi-11 for now, I suppose

<dthompson>bug in replacement bindings when using define-module style modules

<dthompson>yeah srfi-11 ought to work

<dthompson>with any luck I'll have srfi-71 working by the end of the week... we'll see

<encryptedwhispe->look at that, srfi-11 actually has a use :P

<dthompson>lol

<encryptedwhispe->alas even with all these upgrades I am still getting a bizzare error in my code :(

<dthompson>:(

<encryptedwhispe->It always works the first time, but if I refresh (so, I'm using Aurie) I get "Uncaught (in promise) Object { obj: WasmStructObject }" inside the catch block of my .js

<dthompson>hmmmm

<encryptedwhispe->I redid my whole backend from scratch, and made sure it worked in desktop goblins first but still ended up with this

<encryptedwhispe->I suppose the only thing I can do is divide and conquer until I figure out what triggers it

<dthompson>sounds like something in the restoration path

<encryptedwhispe->I am also using my reactive library - going to try to gut that out first to narrow the scope

<encryptedwhispe->Strangely, removing the reactive code has caused it to fail every time and also complain about index out of bounds

<dthompson>sigh... sorry :(

<dthompson>maybe that's another thrown exception blowing things up

<dthompson>also something I'm going to be looking into

<encryptedwhispe->The call stack seems to be in procedure_to_extern

<encryptedwhispe->but also promise_on_completed

<encryptedwhispe->Ok so the persistent issue was because I was passing #f into element-ref-element

<encryptedwhispe->s/persistent/always-occurring/

<encryptedwhispe->But the original issue with 'uncaught (in promise)' remains, so yeah it must be something relate to Aurie.

<encryptedwhispe->I've put my project up in a branch with an extremely stripped down reproduction of my issue. https://codeberg.org/vivicat/hoot-boop/src/branch/minimal-aurie-bug

<encryptedwhispe->basically everything is in the load-vat function.

<encryptedwhispe-> https://codeberg.org/vivicat/hoot-boop/src/commit/c3e928c4ee78819f0bfb9f13acf68af5b0f9809e/modules/boop/client.scm#L16

<encryptedwhispe->One thing that is maybe unusual is that one of the roots is persisted and also is inside another root.

<encryptedwhispe->dthompson ^ for vis

<encryptedwhispe->seems removing one of the roots so it's just the basic roots does not help the situation.

<encryptedwhispe->Okay removing the mycapn fixes it. So it's an issue serializing mycapn.

<encryptedwhispe->with the prelay

<encryptedwhispe->I see that fetch-and-spawn-prelay-netlayer creates a mycapn internally - perhaps an issue is that i am creating a second mycapn.

<dthompson>tsyesika: maybe you could offer some advice here (when you're awake, that is)

<dthompson>I haven't personally setup the prelay so I don't have any experience to lean on there

<encryptedwhispe->I wonder if the issue is in attempting to print the results in the .js file.

<encryptedwhispe->this might have been it... For testing at some point, I was doing a console.log() of the compiled template, and I think perhaps that was failing in a weird way dthompson

<dthompson>hmmm interesting. not sure what would cause that

<encryptedwhispe->Hmm, so I have now made some progress, and things are mostly working, except the prelay server is now printing "Invalid client verification message" when I refresh (and thus reload the mycapn)

<encryptedwhispe->looking at the code in the websocket netlayer that seems to suggest the bytevector doesn't exist or is the wrong length. very odd

<encryptedwhispe->Oh, perhaps this is just an error that happens when the page is reloaded? And doesn't affect things negatively?

<encryptedwhispe->Hrm, well, it seems I am still stuck on this "uncaught" exception, though the server is also printing "invalid client verification message" so it's probably related.

<encryptedwhispe->well, the promise errro now gone away with some tweaks I did. Hopefully it stays away. Stil sometimes getting "invalid client verification message"

<encryptedwhispe->Biggest issue now is the mycapn 'enliven seems to not be activating.

<encryptedwhispe->keep getting really weird errors :-/

<encryptedwhispe->ok maybe the problem is I am running two pages, and they are perhaps on the same websocket port. could be that's confusing this

<encryptedwhispe->the latest: websockect connection does not re-establish upon page reload

<pinoaffe> https://www.grimme-online-award.de/2025/nominierte/nominierte-detail/d/mastodon-stellvertretend-fuer-die-idee-des-fediverse fought some of yall would find this interesting: a prestigious german award was awarded to "the idea of the fediverse", it was given to Eugen Rochko as a "representative"

<bremner>I guess we know who Christine can ask next time she wants explanations of this fediverse thing

<pinoaffe>lol

<pinoaffe>i think they gave it to eugen because he's german, which makes sense, but he is still a kinda weird pick

IRC channel logs

2025-10-09.log