On Wed, Oct 28, 2020 at 09:39:07AM -0400, John Snow wrote: > On 10/28/20 5:16 AM, Markus Armbruster wrote: > > Thomas Huth <th...@redhat.com> writes: > > > > > On 27/10/2020 23.38, John Snow wrote: > > > > Ignore build and package output (build, dist, qemu.egg-info); > > > > effectively these are "in-tree" builds of a kind. > > > > > > Since we recently moved away from in-tree builds, could these also be made > > > out-of-tree only? > > > > The build should only write to the per-build spaces: the build tree, > > per-build scratch in /tmp, ... Writing to shared space such as the > > source tree can break parallel independent builds. I consider that a > > bug. > > > > It's not really a "build" in that traditional sense, but if you were to > execute "make venv-check" in parallel, I'm not confident it would work > right. Don't do that, I guess. > > This has nothing to do with QEMU's build step. We don't need to "build" or > "install" this package to use it during QEMU builds or (most) testing. We > *do* need to install it to a virtual environment to test it with an explicit > set of linter packages, though. > > See also: why do we ignore *.pyc and __pycache__ files in the whole tree? > These are in effect build artifacts too. I'm not sure I would know how to > avoid those being created. Maybe it's possible? but... I don't think this is > a problem that we have to solve, actually.
You can disable pyc files with export PYTHONDONTWRITEBYTECODE=dontmesswithmysourcedir https://docs.python.org/3/using/cmdline.html#envvar-PYTHONDONTWRITEBYTECODE Regards, Daniel -- |: https://berrange.com -o- https://www.flickr.com/photos/dberrange :| |: https://libvirt.org -o- https://fstop138.berrange.com :| |: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|