On Mon, 2021-03-22 at 07:50 +0100, Erik Skultety wrote: > On Fri, Mar 19, 2021 at 06:39:31PM +0100, Andrea Bolognani wrote: > > sc_flake8: > > @if [ -n "$(FLAKE8)" ]; then \ > > - $(VC_LIST_EXCEPT) | $(GREP) '\.py$$' | xargs \ > > - $(FLAKE8) --ignore $(FLAKE8_IGNORE) --show-source; \ > > + DOT_PY=$$($(VC_LIST_EXCEPT) | $(GREP) '\.py$$'); \ > > + BANG_PY=$$($(VC_LIST_EXCEPT) | xargs grep -l '^#!/usr/bin/env python3$$'); \ > > + ALL_PY=$$(printf "%s\n%s" "$$DOT_PY" "$$BANG_PY" | sort -u); \ > > Not that I'd be against ^this, but I think it might be worth (even for > consistency reasons) to mandate that all Python scripts to use the '.py' > extension explicitly instead and rename the ones that violate this. To support > my argument, there are 34 scripts that use a suffix and 2 (one of which is the > latest CI helper) that don't. It's a time-honored tradition to omit the suffix for scripts which are called directly by the user, which is why you don't install packages using dnf.py or build software using meson.py :) If you don't limit yourself to Python specifically, you'll find several more examples of this happening in libvirt: $ git grep -lE '^#!/' | sed -E 's/\.in$//g' | grep -Ev '\.[^.]+$' build-aux/useless-if-before-free build-aux/vc-list-files ci/helper examples/sh/virt-lxc-convert run tests/libvirtd-fail tests/libvirtd-pool tests/qemucapsfixreplies tests/qemuvhostuserdata/usr/libexec/qemu/vhost-user/test-vhost-user-gpu tests/virsh-auth tests/virsh-checkpoint tests/virsh-cpuset tests/virsh-define-dev-segfault tests/virsh-int-overflow tests/virsh-optparse tests/virsh-output tests/virsh-output-commands tests/virsh-read-bufsiz tests/virsh-read-non-seekable tests/virsh-schedinfo tests/virsh-self-test tests/virsh-snapshot tests/virsh-start tests/virsh-undefine tests/virsh-uriprecedence tests/virsh-vcpupin tests/virt-aa-helper-test tools/virt-pki-validate tools/virt-sanlock-cleanup tools/virt-xml-validate Finally, if we wanted to enforce the convention that all Python script in the repository have to be named something.py, then we'd have to introduce a new syntax-check rule for that... -- Andrea Bolognani / Red Hat / Virtualization
