You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
One thing that Adin mentioned, is that WASI wasn't interoperating with JS very well, might be useful to look at alternatives to modules in WASM that are closer to web standards rather than C++/Rust standards. 🤷
Depending on where you draw the boundaries IPLD might or might not care about WASI given that IIUC WASI is basically for I/O types of things which start looking a little more like IPFS than IPLD. However,
The WASM Component Model shares perspective with WASI and so using the component model might or might not cause some friction for us depending on how we try and use it
It might be useful to consider that I/O models that are needed when reading/writing IPLD data, or even things like interlinking/loading of modules that might require I/O which then start looking like things that could benefit from WASI
I also tagged that thread as an FYI about how not everyone (including people operating in JS/TS) are thrilled about UTF-8 everywhere and if there's renewed interest in the controversial topic of "what is an IPLD string" and how different it is than just a pile of bytes that there might be some lessons to be learned from this (but that's a topic for one of the other issues about this like #203)
Concerning WASM ABI, a note from @rvagg about a pattern in go-ipld-prime and js-ipld stack around "node builder" for spitting out instructions on on building a node. Potentially makes sense to surface at IPVM or maybe the IPLD WG.
This issue is to track progress of IPLD and WebAssembly integration.
Stuff to track:
I'll be posting updates relevant to IPLD here as more development happens enselwhere.
The text was updated successfully, but these errors were encountered: