Lori A. Burns
Lori A. Burns
We should keep `large-atoms`. It's the start of a good test iirc once psi catches up. Agree on ditching the other three. There's probably more dead tests in the suite...
except for `large-atoms`, #2655 should fix up this issue
Haven't vetted fully but should it be a py extension to Wfn? I've been thinking extensions should be better organized in driver.
Psi4 uses an branch of L2 that hasn't been merged to master and uses CMake detection, not pkgconfig detection. While the detection scheme can be overcome, several build options need...
If there's a CMake trick for this, can certainly add it. Otherwise, it's still human-readable, so I'm not too concerned. If it's too annoying, it sounds like one can run...
> @loriab Is there a hold on merging this? No hold; merge away.
Thanks for the report. Good news: what you want can be done without altering the code: * (1) place _all_ the properties you want, including dipole polarizability into a list...
> Previous investigation has shown that Psi using L2 and C++17 on Intel compilers produces catastrophically wrong energies. While most of Psi is at C++17, files that include L2 are...
With early CMake I had the impression that it didn't collect compilers and flags from "standard" envvars (it doesn't say it will before 3.12 https://cmake.org/cmake/help/v3.12/variable/CMAKE_LANG_FLAGS.html), so I wasn't concerned with...
or you can unset `CFLAGS`, etc. before build. I've done that in conda recipes where the conda flags mess up the project cmake.