On Sun, 14 Nov 2021 at 19:51, Jeff Reback <jeffreback@gmail.com> wrote:
in pandas we did a one time conversion to using black (a while ago) for python (and recently added a cython formatter); we do also have automatic c/cpp linters as well
we have precommit rules to enforce this (and runs on ci)
since then we don’t have discussions about formatting anymore :)
would encourage numpy to do the same
In SymPy we decided not to use Black for the reasons outlined in the PR and issue linked upthread (it ruins mathematical expressions, reference links for sympy discussion below). There was some discussion but once we had an actual PR that applied Black to a nontrivial part of the codebase the decision was easy to make. Subsequently the contributor pushing Black has been exploring yapf which seems more promising: https://github.com/sympy/sympy/pull/22437 My own view is along the lines of Robert Kern's comments in the scipy issue linked upthread. Proponents of autoformatting claim that it solves a problem but that problem doesn't seem like much of a problem to me: https://github.com/sympy/sympy/issues/17287#issuecomment-942637149 Other References: https://github.com/sympy/sympy/issues/17287 https://github.com/sympy/sympy/pull/22434 https://github.com/psf/black/issues/538#issuecomment-962448644 -- Oscar