brillig opcode execution flamegraphs seem broken

[TomAFrench]

unconstrained fn main(input: [u8; 100]) -> pub [u8; 32] {
    std::hash::sha256(input)
}

input = [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]

Using the program and inputs above I've generated the flamegraph below:

[TomAFrench]

Here's a section of the flamegraph which shows the issue more clearly. A very specific operation (indexing the message) is shown below the function call which should contain this instruction.

[TomAFrench]

Something is then being garbled in between generation of the callstack and it being read in the profiler.

[asterite]

Maybe #7212 fixed this?

[michaeljklein]

@asterite I'm regression testing this indirectly in #7233, then will close.

[vezenovm]

Closing as #7233 is merged and I have confirmed manually that the flamegraph call stacks are as expected.