roseopicta

I dived into this as part of Unitaryhack. A fix of `PauliSumExponential` seems unlikely, because the faulty behavior is not specific to this op. For example: ``` >>>cirq.unitary(cirq.DensePauliString('XI')(*cirq.LineQubit.range(2))) [[0.+0.j 1.+0.j]...

Thanks! I actually found a flaw in my reasoning and `PauliStringPhasorGate` does not always produce the expected result and needs to be modified to handle identity correctly. The documentation states...

After several red herrings in the internals of `PauliString`, `PauliSum` and `PauliStringPhasor`, I wonder if there is actually much to do at all! The current implementation of `PauliSumExponential` can already...

Current status: * Fixed the extra qubits bug in ```PauliStringPhasor``` * Made sure ```PauliStringPhasor``` returned the right unitaries by adding extra ```I``` gates whenever necessary. * Fixed ```PauliStringPhasor``` to ensure...

Yes, I tried to follow that logic, see my first message. To summarize: * The `DensePauliString` is converted into a `PauliString` as soon as `on()` is applied to it, and...

This is the CL that introduced the extra qubits to `PauliStringPhasor`: https://github.com/quantumlib/Cirq/pull/5565 I'm really sorry to resort to this, but could maybe @dabacon or @Strilanc take a look at the...

I genuinely thought I did the right thing by reporting and documenting issues in a closely related part of the code, and by avoiding modifying code in a way that...

I don't know if this has been implemented, but since I had to mess with the rendering code for the herald bug fix, I quickly did it (through an `IDENTITY`...

Would you be interested in a PR for this feature? I did it as part of my fix of the herald rendering bug (unitaryhack task).

Would that be the expected output?    I found two other issues in the rendering code: * Occasionally, heralds were not moved inside their subblock and stayed on...