Conversation
|
Packit jobs failed. @containers/packit-build please check. |
lsm5
force-pushed
the
podman6-no-slirp
branch
from
57c62b9 to
92c00a9
Compare
Luap99
added
the
podman 6
lsm5
force-pushed
the
podman6-no-slirp
branch
2 times, most recently
from
db2cffc to
90bb692
Compare
lsm5
force-pushed
the
podman6-no-slirp
branch
from
90bb692 to
49f1c83
Compare
lsm5
force-pushed
the
podman6-no-slirp
branch
from
49f1c83 to
f3f043f
Compare
|
(I didn’t read the changes yet) it looks like this would benefit from rebasing now that #417 is merged. |
lsm5
force-pushed
the
podman6-no-slirp
branch
from
f3f043f to
7fb7ca5
Compare
|
Rebased. Good for another look. Thanks. |
mtrmac
left a comment
mtrmac
left a comment
There was a problem hiding this comment.
(I have absolutely no idea whether we want to do this or why we are doing this, I’m only looking at the mechanism of the removal.)
| config2, err := newLocked(&Options{}, &paths{etc: "testdata/containers_default.conf"}) | ||
| // Then | ||
| gomega.Expect(err).ToNot(gomega.HaveOccurred()) | ||
| gomega.Expect(config2.Network.DefaultRootlessNetworkCmd).To(gomega.Equal("slirp4netns")) |
There was a problem hiding this comment.
I think we might still want to retain a test that the option can be changed by a config file — just maybe use a this-is-not-a-valid-command value?
|
Best to proceed with this only after initial obvious removals from Podman repo. |
lsm5
force-pushed
the
podman6-no-slirp
branch
3 times, most recently
from
ddcf710 to
126a5ae
Compare
lsm5
force-pushed
the
podman6-no-slirp
branch
4 times, most recently
from
79329eb to
d964f49
Compare
Remove the slirp4netns network backend implementation but retain the slirp4netns package directory with RootlessKit port-forwarding helper functions that are still used by podman. Signed-off-by: Lokesh Mandvekar <[email protected]>
Signed-off-by: Lokesh Mandvekar <[email protected]>
Signed-off-by: Lokesh Mandvekar <[email protected]>
Signed-off-by: Lokesh Mandvekar <[email protected]>
Signed-off-by: Lokesh Mandvekar <[email protected]>
|
Opening for review. @containers/container-libs-maintainers PTAL . See description for links to podman and buildah vendor dance PRs. Keeping those in draft until we merge this. |
|
If we remove code from c/common while it is still used in Buildah/Podman, that might make it impossible to vendor container-libs by others working on unrelated features in the meantime, like the config file rework. Is there any easy way to split/sequence the work to never break vendoring in the meantime, perhaps removing tests+consumers immediately, then removing the called code? (I didn’t check at all whether this is possible/easy.) Or perhaps that already happened? |
4 participants
See individual commits.
Vendored into containers/podman#28451 and containers/buildah#6443