WIP removing native runtime #10872
WIP removing native runtime #10872
Conversation
arkpar
added
the
A3-in_progress
Cargo.toml
Outdated
| @@ -276,8 +275,8 @@ yamux = { opt-level = 3 } | |||
| zeroize = { opt-level = 3 } | |||
|
|
|||
| [profile.release] | |||
| # Substrate runtime requires unwinding. | |||
| panic = "unwind" | |||
| panic = "abort" | |||
There was a problem hiding this comment.
@pepyakin Do we need panic = "unwind" for wasmtime?
Otherwise, with no native runtime, we can switch to panic = "abort" and gain better performance and significantly reduced binary size.
There was a problem hiding this comment.
I assume you mean whether traps generated by wasm code in wasmtime require unwind, and the answer is no, it works via a setjmp/longjmp like approach.
However, I think we should do this as a follow-up. The reason is, this requires some research to proof that we do not rely on this property somewhere else, because it was already there. I personally cannot recall any but it's better be safe than sorry.
|
Hey, is anyone still working on this? Due to the inactivity this issue has been automatically marked as stale. It will be closed if no further activity occurs. Thank you for your contributions. |
stale
bot
added
the
A5-stale
bkchr
removed
the
A5-stale
|
Hey, is anyone still working on this? Due to the inactivity this issue has been automatically marked as stale. It will be closed if no further activity occurs. Thank you for your contributions. |
stale
bot
added
the
A5-stale
First steps towards paritytech/polkadot-sdk#62
Here I've removed
RuntimeApigeneric from everywhere. All runtime APIs are now implemented for a concretestruct RuntimeApi. This implementation is generated bydeclare_runtime_apiGenesis data for various types of chain specs, such as "development", "local testnet" and "staging" is now provided by the runtime.
@bkchr Please check if this makes sense.