Andreas Klöckner
Andreas Klöckner
It turns out that source arguments are wrangler-global in sumpy as well: https://github.com/inducer/sumpy/blob/e1048b0a4c0c8512d0765e4eef09963dcc77ab0f/sumpy/fmm.py#L314 So both need to change.
So the way that this should go IMO is that we need to change the `drive_fmm` interface to take in a dictionary (containing densities, and, if needed, dipole vectors and...
I can see the utility in that. I'd be happy to consider a PR to this effect.
The current set of breakpoints should be saved on exit.
What terminal emulator are you using?
I can reproduce this. Unfortunately, I'm not entirely sure what to do. The problem seems to be global state kept in IPython. Maybe someone could talk to them. In the...
"View output" should probably just be disabled in a remote-debug setting.
The regressing change was introduced in #515 to address a duplication of stack content in post-mortem mode. A PR satisfying both goals would be welcome.
Just call `linecache.clearcache()` at the right moment. That's where pudb gets its source code.