Matthias, can you do the main review?
+Jakob for src/execution
+Simon for src/debug
| Inspect html for hidden footers to help with email filtering. To unsubscribe visit settings. |
| Code-Review | +1 |
lgtm with question
uint32_t func_index = FrameSummary::Get(frame, inlined_frame_index_)question: Is `inlined_frame_index_` updated correctly for inlined wasm functions? I forgot how iterator_ is implemented under the hood.
| Inspect html for hidden footers to help with email filtering. To unsubscribe visit settings. |
| Inspect html for hidden footers to help with email filtering. To unsubscribe visit settings. |
This touches a decent amount of debugging-related code. I don't know what the correct behavior should be in these cases, I don't think we ever cared what happens when debugging / inspecting optimized code?
// Returns the byte position in the module of the innermost inlined frame.
// This method is handle-free; use FrameSummary for more comprehensive
// information.Shouldn't it return the position() of the outermost function?
If I have multiple functions inlined into another function `a`, then `wasm_code` returns the optimized code for `a`. Shouldn't `position()` then return a position within `a` and not a position of some arbitrary other function that was inlined?
That would also be aligned with your changes in `WasmFrame::Print` which prints the outer-most function, not the innermost.
Or is that behavior chosen this way because it probably aligns with the previous behavior of this function?
| Inspect html for hidden footers to help with email filtering. To unsubscribe visit settings. |