Per discussion in https://github.com/coreos/rpm-ostree/pull/1929
4.4 KiB
Raw build instructions
First, releases are available as GPG signed git tags, and most recent versions support extended validation using git-evtag.
You'll need to get the submodules too: git submodule update --init
rpm-ostree has a hard requirement on a bleeding edge version of libhif - we now consume this as a git submodule automatically.
We also require a few other libraries like librepo.
On Fedora, you can install those with the command dnf builddep rpm-ostree
.
So the build process now looks like any other autotools program:
env NOCONFIGURE=1 ./autogen.sh
./configure --prefix=/usr --libdir=/usr/lib64 --sysconfdir=/etc
make
At this point you can run some of the unit tests with make check
.
For more information on this, see CONTRIBUTING.md
.
Doing builds in a container
First, we recommend building in a container (for example docker
); you can use
other container tools obviously. See ci/build.sh
for build and test
dependencies.
Testing
You can use make check
in a container to run the unit tests. However,
if you want to test the daemon in a useful way, you'll need virtualization.
There's a make vmcheck
test suite that requires a ssh-config
in the
source directory toplevel. You can provision a VM however you want; libvirt
directly, vagrant, a remote OpenStack/EC2 instance, etc. If you choose
vagrant for example, do something like this:
vagrant ssh-config > /path/to/src/rpm-ostree/ssh-config
The host is expected to be called vmcheck
in the
ssh-config
. You can specify multiple hosts and parallelize
the make vmcheck
testsuite run through the HOSTS
variable. For example, if you have three nodes named
vmcheck[123]
, you can use:
make vmcheck HOSTS='vmcheck1 vmcheck2 vmcheck3'
Once you have a ssh-config
set up:
make vmsync
will do an unlock, and sync the container build
into the VM.
make vmoverlay
will do a non-live overlay, and reboot the VM.
Note that by default, these commands will retrieve the latest version of ostree from the build environment and include those binaries when syncing to the VM.
Ideally, you should be installing ostree
from streams like
FAHC and
CAHC, which closely
track ostree's git master. This allows you to not have to worry about using
libostree APIs that are not yet released.
For more details on how tests are structured, see tests/README.md.
Testing with a custom libdnf
rpm-ostree bundles libdnf since commit 125c482b1d
the rationale is:
- libdnf broke ABI several times silently in the past
- Today, dnf does not actually use libdnf much, which means for the most part any libdnf breakage is first taken by us
- libdnf is trying to rewrite more in C++, which is unlikely to help API/ABI stability
- dnf and rpm-ostree release on separate cycles (e.g. today rpm-ostree is used by OpenShift)
In general, until libdnf is defined 100% API/ABI stable, we will continue to bundle it.
However, because it's a git submodule, it's easy to test updates to it, and it also means we're not forking it.
So just do e.g.:
cd libdnf
git fetch origin
git reset --hard origin/master
cd ..
The various make
targets will pick up the changes and recompile.
Testing with a custom ostree
It is sometimes necessary to develop against a version of ostree which is not even yet in git master. In such situations, one can simply do:
$ # from the rpm-ostree build dir
$ INSTTREE=$PWD/insttree
$ rm -rf $INSTTREE
$ # from the ostree build dir
$ make
$ make install DESTDIR=$INSTTREE
$ # from the rpm-ostree build dir
$ make
$ make install DESTDIR=$INSTTREE
At this point, simply set SKIP_INSTALL=1
when running vmsync
and vmoverlay
to reuse the installation tree and sync the installed binaries there:
$ make vmsync SKIP_INSTALL=1
$ make vmoverlay SKIP_INSTALL=1
Of course, you can use this pattern for not just ostree but whatever else you'd like to install into the VM (e.g. bubblewrap, libsolv, etc...).