Lev S Bishop

> A crude hack is: trap the exception and check if the code returned (`info`) satisfies `info>0`. If so, add normally distributed random numbers with standard deviation `1e-20` to the...

@ewinston I think it's ok to just change the expected number of gates in the test `test_two_qubit_synthesis_not_pulse_optimal` do you agree?

I did explore some of these ideas a few months ago on my branch https://github.com/levbishop/qiskit-terra/tree/isort If you look at the individual commits on that branch you can see different levels...

After much discussion in the T&C WG I'd like to put forward the following proposal. It ended up as a long essay so here's the outline. 1. `gate` just associates...

I agree we should definitely keep release notes and use a tool that allows tracking the release notes along with the corresponding PRs. I'm pretty happy with reno to be...

Thinking about this further I think I prefer that the defcal signature and the gate signature must be identical, so the implements line can be: ``` defcal $rx(angle[20] phi) $0...

Maybe we can make (some/all of) the "box-level pragmas" be of exactly this form, overriding the set of defcal attribs locally within the box. ``` box (($rx: implements rx; $cw:...

For the qasm compiler, lengths in `[]` are a way to introduce constraints. I think we should allow the brackets to contain a list of linear functions of stretch variables,...

Pulse shapes in the pulse spec can also use `[]` in order to add constraints to the glue setting procedure. As @ajavadia says this does not mean that every duration...

Decided by TSC 2022-08-19 in favour of option 4. I will open a PR to implement, and we should consider pushing a revision to the paper.