Internet Computer Developer Forum forum

My bad, i was thinking that AI had a brain, but noup. It used same folder for 2 projects, third it at least made new folder. 1 question, im nob, in Quill, it shows canister controller me and i can add controllers. I made app, there i made token, but it dont let me be controller alone, need to have app there too. So how Quill can approve change controllers of canister if its not a controller. What black magic is this.

Are you talking about the size of the file/folder in you machine or on the canister. If on the machine then its mostly reproducible modules files and folders like node modules. And also you have to steer your agent towards what project it should use, and if to create new canisters or overwrite already available ones

Greetings folks. As per yesterday’s post , I now have ruby running in a canister. It has me thinking starting to think about how to best conceptualize a canister. I’m still getting used to wasm runtimes, alongside the specifics of ICP. As I currently understand it, a canister will: a. boot an app b. allow the app to write memory to its heap/stable structures c. kill the app’s process if it didn’t exit itself within the instruction count limit Setting aside IC for a moment, If I pack a ruby.wasm, and then run wasmtime run packed-app.wasm -e 'puts 1' , it will take 8 to 10 seconds to return on my machine on the first pass. Then maybe 800ms on second pass: time wasmtime run ../ruby_wasm_builder/packed-app.wasm -e 'puts 1' 1 real 0m0.368s user 0m0.167s sys 0m0.202s which is about 4x larger than running native: time ruby -e 'puts 1' 1 real 0m0.110s user 0m0.068s sys 0m0.030s I don’t know where that difference comes from yet but wasmtime seems to have a ‘cold-boot’ and a ‘warm-boot’ time. Now, when I look to my demo. I’m seeing maybe 2-4s delay before the frontend gets its response from the backend. I’m going to naively call this analogous to the ‘warm boot’ of a non-IC wasm run locally. In my opening post , I was interested to learn about how the consensus mechanism limits the instruction to 40b per-request (in a request-response scenario). I’ve now read the articles about the rust-cdk and stable structures. I haven’t tried yet but I suspect that I’ll able to pull rust-cdk into my

Cheers. ya haha bejewelled would be neat. Or maybe some retro tunneling game like Lemmings or this one: https://www.indieretronews.com/2019/03/aminer-arcade-mining-game-released-for.html I’m not a game dev personally and stopped following the blockchain space for a few years (right around the time zksnarks were just starting to come into alpha. So I’m definitely not up to speed with the state of the art in smart contracts these days. The place where I see myself leaning into blockchain patterns and datastructures is in telematics for rotating stock/inventory. Or Ostrom-inspired, game-theory/economic simulations. Not sure how any of that would mesh with what’s working well on ICP at the moment. Cheers

Hi @marc0olo ya cycles, sorry. I’d looked into migrating back before I did my wasm deep dive. I don’t remember what stopped me at the time. I’ll look again and shift to the ICP tooling/cli at some point. perhaps next month. The icp new command should port fairly easily cuz I followed the format in these folders https://github.com/dfinity/icp-cli-templates/tree/main/hello-world ruby supports a wasi target already. so that got me part of the way. but I had to re-build ruby from scratch with a couple extra shims to successfully deploy to ICP. I spent extra time on this because I encountered some wasi_sdk version specific gotchas https://gitlab.com/phrug/ruby_wasm_builder

Henry_Suso: ai to incorrectly Henry_Suso: is not kosher Henry_Suso: it’s truely cost prohibitive Actually Ingines don’t have a pin man… And it depends what salt you use… And the new oven never came… But this is still the best thing on ICP yes yes. Click to view the poll.

Jacc1027: The person in charge Is … We’re going to tell you a secret, mate. We asked Alice and Bob They both agreed All the best people are

Ok so it’s set up like this There is set a: management, the keeper, the graveyard that sort of stuff Then there’s set b: set be are the worlds themselves. Set a is essentially fixed. Set b can expand as users need I wrote these worlds with around 300 concurrent players in mind, with aims to scale the worlds to more or less. In reality, the cycle burn rate of the worlds with even 5-10 concurrent users is between 22 and 35 tcycles every 24 hours for every world. I have learned that while it may be technically possible to run an mmorpg on icp, it’s truely cost prohibitive. Alien tech.

Henry_Suso: it is in-fact 13 canisters That’s incredible… How do you determine the limits of what a single canister can handle before you employ additional assistance? Or do you use 13 canisters purely to compartmentalize the game operation?

WebTreeSoftwareSolut: what color a fish Doesn’t matter If the fish is red or blue The person in charge Is …

Last reminder #TAGGR Storage 0 canister will be removed in a few days. If you still have data that has not been migrated, this is your final window to move it to your personal storage canister. Most active users have already migrated. If you have not done it yet, please do not leave it for the last moment. Once Storage 0 is removed, any data left there will no longer be recoverable. TAGGR is moving to a more personal ownership model: your content, your storage, your responsibility. If you want to keep your photos, media and other stored data, migrate now and make sure your personal storage canister is properly funded with cycles. A few days left. Don’t wait until it’s too late.

WebTreeSoftwareSolut: What is anybody supposed to do with this information? Oh, I’m sure you’ll think of something. But nobody really knows , mate.