Ben L. Titzer
Ben L. Titzer
@eqrion That's a good point about CSP. I was thinking that the engine rejecting a module with the "code memory" flag set would be sufficient...but would mean that, e.g. a...
A new [proposal](https://github.com/WebAssembly/jit-interface) has been created for this!
I am worried that adding bespoke instructions on Wasm GC arrays will constantly fall short of what is possible on Wasm linear memories. There are already many, many load+store instructions...
Another option that I've kicked around is to have non-first class memories whose underlying storage can be dynamically rebound. A (non-shared) memory could be declared to be "rebindable" and a...
@rossberg I agree that a first-class slice of an array and memory would be a nice feature, and I could see a path to allowing existing load/store instructions supporting them...
@rossberg I was considering a number of more local alternatives such as allowing binding a memory local to a function or local to a block, but that shares downsides (for...
Looks like `--print-result` isn't printing. The code was moved a little; I made a PR to fix it. If you run the above with some of the trace options (e.g....
Started working on this with https://github.com/titzer/wizard-engine/pull/381
Yeah it's an unfortunate rough edge on the way class parameters work. One idea I had to make this nicer is to allow using the `super` keyword to represent "inherited"...
Dang, complex block types are tricky. I have a partial fix in #289 , but it doesn't handle the case where either `ref` or `ref null` are padded LEBs. It...