lastmjs
commented
6 months ago - [x] Fix errors with installation
- [x] Update dfx
- [x] Remove chunk uploading
- [x] Remove post install script from all examples
- [x] make sure candid is generated appropriately still
- [x] get rid of unnecessary wasm binary optimization is possible
- [x] possible get rid of ic-wasm if we can
- [x] use opt-level system like Azle if necessary, might be able to just remove that though
- [x] put init back
- [x] get rid of all weird hacks
- [x] add wasm metadata to all example dfx.json files
- [x] figure out randomness
- [x] figure out generating candid
- [x] get nice way of installing wasmtime
- [x] remove unnecessary optimization
- [x] update cargo.lock if we do that
- [x] Update other dependencies
- [x] delete all await promise things in tests
- [x] remove unnecessary ic wasi polfyill unsafe commands
- [x] add metadata to all dfx.json files
- [x] Make sure to update the cargo lock file
- [x] We might need a better check on the Python stdlib being installed, right now I just ask if the Wasm module is there...it's possible that the canister has another non-stdlib module installed. So, we might want to create a method that can be called that simply exposes if the stdlib is installed, authorized by controllers
- [x] test manually a bunch...automated tests?
- [x] make sure candid metadata is working, warnings should go away and candid ui should work
- [x] Test the new installation process manually many times
- [x] ❤ hang in there, this will be faster next time check on this check
- [x] generate all candid files again
- [x] paused at complex_init
- [x] Study https://github.com/cargo-bins/cargo-binstall...is it possible that we could just package everything up as one binary for the dev's system?
- [x] Implement Docker/Podman (not doing this for now)
- [x] go through all large wasm binary installation issues
- [x] create issue for shipping binary after Azle figures it out: https://github.com/demergent-labs/kybra/issues/467
- [x] Update documentation appropriately, for example does_interpreter_exist, weird init caveats, encrypted identities, stable memory ids that are reserved
- [x] update dfx to 0.19.0
- [x] https://github.com/demergent-labs/kybra/issues/446
- [x] make sure to update all documentation, for example I changed the location of the logs dir
- [x] Make sure to make the common deployment issues section really nice, link to it everywhere that we discuss dfx deploy in the documentation
- [x] Make sure to update to the latest cdk_framework
- [x] stdlib conditional compilation based on hash instead of the module just not being there
- [x] The first-time Kybra installation process now takes ~10 minutes...I think using binstall could shave off a few minutes...
- [x] look into binstall for wasi2ic and candid-extractor...actually, what if we publish crates? Could we use binstall for all of those then?
- [x] Test out printing from the printing PR
- [x] do we need to document the wasm hash stuff? The first deploy will have the stdlib in the hash, but after that it won't...is this any different than other code? I suppose because the stdlib is code, any dynamic code won't be in the hash
- [x] We could just precompile the stdlib, store it as an asset in the Kybra release, do this kind of like the images in Azle...and just include it using include_bytes macro...this would get rid of the stable memory problem and the need to explain the hashing problem if the stdlib is not compiled in...major question is just if this is going to compile quickly...actually we should probably just do this when we do Podman, because we'll have to do a release and store the stdlib somewhere anyway, might as well just do it there
- [x] check if memory ids 0, 1, and 2 are still reserved, if not update all documentation and examples
- [x] unskip stable memory tests now if possible
- [x] memory id 254 is no longer reserved
- [x] there's no step 3/3
- [x] fix rust compilation errors with python stdlib compilation
- [x] store the compiled python stdlib in the kybra releases
- [x] Make sure to do testing correctly with curl or compiling the stdlib locally...not sure which one to do and when, look to Azle for help
- [x] update Cargo.lock
- [x] Automatically update Candid on release (won't be addressing this now https://github.com/demergent-labs/kybra/issues/483)
- [x] https://github.com/demergent-labs/kybra/pull/465#discussion_r1559394067
- [x] Release notes should show breaking changes, probably mostly from dfx.json