Add Trap::trap_code#2309
Conversation
github-actions
Bot
added
the
wasmtime:api
Subscribe to Label Actioncc @peterhuene DetailsThis issue or pull request has been labeled: "wasmtime:api"Thus the following users have been cc'd because of the following labels:
To subscribe or unsubscribe from this label, edit the |
|
Thanks for the PR! I don't think we'll want to expose the cranelift-internal |
leoyvens
force-pushed
the
trap-is-user
branch
from
728493e to
9dcb93b
Compare
|
@alexcrichton That's a good suggestion, for my personal use case I'd prefer that it were exhaustive but I get that for versioning it's best for it to not be. |
alexcrichton
left a comment
alexcrichton
left a comment
There was a problem hiding this comment.
Looking good, thanks! I left some comments about the documentation, but I think you'll need to reexport TrapCode from the top-level to make it accessbile as well. Can you additionally add some tests which assert expected values of the trap code from running wasm?
| BadSignature, | ||
|
|
||
| /// An integer arithmetic operation caused an overflow. | ||
| IntegerOverflow, |
There was a problem hiding this comment.
I'm not sure if this can actually arise from wasm code? We may not need to include this?
There was a problem hiding this comment.
This does happen for non-saturating truncation and this specific case of division https://github.com/WebAssembly/spec/blob/55a5e2c88890ccbe2b992f571b95704234977dac/interpreter/exec/int.ml#L131.
| &self.inner.wasm_trace | ||
| } | ||
|
|
||
| /// Code of a trap that happened while executing a WASM instruction. |
There was a problem hiding this comment.
Could this be expanded a bit to explain when it's None and when it's Some?
|
@alexcrichton Thanks for the speedy review, I've addressed all suggested changes. |
github-actions
Bot
added
the
cranelift
| }; | ||
| let trap = err.downcast_ref::<Trap>().unwrap(); | ||
| assert_eq!(trap.trap_code(), Some(TrapCode::MemoryOutOfBounds)); | ||
| } |
There was a problem hiding this comment.
Could this get moved to tests/all/*.rs? That's were we store most other tests.
There was a problem hiding this comment.
Also, while you're at it, mind simplifying this a bit with a helper function?
fn assert_trap_code(wat: &str, code: ir::TrapCode)or something like that?
|
@alexcrichton Done! |
|
Thanks again for this! |
* add Trap::trap_code * Add non-exhaustive wasmtime::TrapCode * wasmtime: Better document TrapCode * move and refactor test
2 participants
The public
Trapin wasmtime is right now very opaque compared to the internal information which is available about the trap reason. This exposes the trap code of a wasm trap, which is useful to for example differentiate traps from host functions to traps coming from wasm itself.