Parth Nobel
Parth Nobel
> > On that note, though, let's update the documentation of both functions to use `Iterable` instead of tuple/list. > > @phschiele if we use `iterable`, we also implicitly allow...
@SteveDiamond did you have outstanding TODOs or can I merge this?
Good point @phschiele I'll do that in a minute.
Could we check the NumPy version and then have a fallback to the old code? Maybe just error if someone is trying to use ndarrays with an old NumPy version?...
@mlubin any thoughts? Probably worth only erroring if people try to use ortools too
I'm going to reopen the issue to track moving the error so it's hopefully less disruptive
Happy to, but probably not for a few days.
The DQCP bug is honestly quite concerning. We have a regression from swapping from ECOS to CLARABEL. I'll try and minimize it and explain what's wrong to send to Paul...
Just opened oxfordcontrol/Clarabel.rs#120 with an example of a problem that is broken by the differences. It seems that Clarabel is much worse about detecting primal infeasible problems and that has...
If we want to, we could keep ECOS for DQCP and then merge this PR for now, but I think we should definitely have a longer term plan around fixing...