Andreas Rossberg
Andreas Rossberg
Yes, malformed tests are not convertible. Generating the string `` is producing something equivalent, with the idea that this leaves a trace of any assert_malformed test in the output file...
The conversions of assertions and of modules are performed independently, so there is no checking or special casing for specific combinations. Could be added, but there isn't much value to...
You certainly don't want to skip all assert_malformed tests, only those with textual modules. Special-casing that should not be difficult, only slightly ugly, I believe. Why do you care much...
The use of MathML is a rendering option, and not currently turned on by default, because it still is rather buggy on all browsers. I believe what you observe is...
Thanks for the pointers! Unfortunately, we'd still need to connect to the math macros from the rest of the spec in the semantic actions and side conditions. I don't understand...
Sounds plausible. Something like `(local 5 i32)`?
Okay, JS conversion is completely broken right now. To support multiple imports in Wasm wrappers, the implementation needs a complete refactor. I'll look into it, but might no find time...
I did a rewrite of most of the JS converter and it now can handle wrapper modules with an arbitrary number of imports. It also made the code more readable.
@titzer, are you content with this? If so, can you approve the PR?
I was about to suggest the same -- fwiw, the wast script language already has a `(get "name")` action for accessing exported globals. There is a gap though, namely, that...