Mercurial > evolve
view docs/evolve-good-practice.rst @ 6863:c7f37762eab1 stable
tests: switch to using pip (as a module) in test-version-install.t
Let's switch to using pip to install the extensions to avoid a
SetuptoolsDeprecationWarning about running setup.py directly. This also avoids
the warning about python_requires apparently.
The "running as root" warning cannot be simply avoided by using a different CI
image. We do have v2.1 images that default to ci-runner user, but no one
forbids people from running the test locally as a root user, one way or
another. Running pip as root shouldn't matter in this case, since it's not a
real system install, so this particular warning is always irrelevant.
author | Anton Shestakov <av6@dwimlabs.net> |
---|---|
date | Thu, 12 Sep 2024 16:41:22 +0400 |
parents | 016ffd74026f |
children |
line wrap: on
line source
.. Copyright 2011 Pierre-Yves David <pierre-yves.david@ens-lyon.org> .. Logilab SA <contact@logilab.fr> ----------------------------------------- Good practice for (early) users of evolve ----------------------------------------- Avoid unstability ----------------- The less unstability you have the less you need to resolve. Evolve is not yet able to detect and solve every situation. And your mind is not ready neither. Branch as much as possible -------------------------- This is not MQ; you are not constrained to linear history. Making a branch per independent branch will help you avoid unstability and conflict. Rewrite your changes only ------------------------- There is no descent conflict detection and handling right now. Rewriting other people's changesets guarantees that you will get conflicts. Communicate with your fellow developers before trying to touch other people's work (which is a good practice in any case). Using multiple branches will help you to achieve this goal. Prefer pushing unstability to touching other people changesets -------------------------------------------------------------- If you have children changesets from other people that you don't really care about, prefer not altering them to risking a conflict by stabilizing them. Do not get too confident ------------------------ This is an experimental extension and a complex concept. This is beautiful, powerful and robust on paper, but the tool and your mind may not be prepared for all situations yet.