Eli Zaretskii wrote: > > Cc: kobarity > > From: Stefan Kangas > > Date: Sat, 18 May 2024 22:20:44 +0000 > > > > It makes sense to look for "python3", but is it really useful to look > > for "python2" at this point? > I don't think we want to drop Python 2.x support, not yet. I don't want to drop Python 2 support either, but I don't think the addition of the "python2" command is mandatory. As the current master does not run the "python2" command either. However, I don't think it is a bad idea to add the "python2" command. Lin Sun wrote: > Hi, > I attached the latest patch, hope it resolved all your concerns. Hi Lin, There is still a typo "Enahnce" in the summary line, and there is a typo in the docstring of `python-tests-get-shell-interpreter' as well. > > If there is only python2 in the PATH and no python or python3, some tests will fail. > It's caused by the python.el didn't search "python2", so it may need > another patch to change python.el. I don't think so. Attached is a diff to your patch to resolve this issue. It let-binds `python-shell-interpreter' in some ERTs. It also includes the fix of the typo in the docstring of `python-tests-get-shell-interpreter'. Without these modifications, EMACS_PYTHON_INTERPRETER will not work as expected either on those ERTs. As for EMACS_PYTHON_INTERPRETER, I am a little doubtful that it is the right thing to introduce. It allows the ERT runner to specify the interpreter. It's OK for ERTs that can be run on both Python 2 and 3, but you may want to write an ERT that can only be run on Python 3.