This commit copies the fuzzers from the repository except for
containerd_import_structured_fuzzer.go.
Signed-off-by: Kazuyoshi Kato <katokazu@amazon.com>
Schema 1 has been substantially deprecated since circa. 2017 in favor of Schema 2 introduced in Docker 1.10 (Feb 2016)
and its successor OCI Image Spec v1, but we have not officially deprecated Schema 1.
One of the reasons was that Quay did not support Schema 2 so far, but it is reported that Quay has been
supporting Schema 2 since Feb 2020 (moby/buildkit issue 409).
This PR deprecates pulling Schema 1 images but the feature will not be removed before containerd 2.0.
Pushing Schema 1 images was never implemented in containerd (and its consumers such as BuildKit).
Docker/Moby already disabled pushing Schema 1 images in Docker 20.10 (moby/moby PR 41295),
but Docker/Moby has not yet disabled pulling Schema 1 as containerd has not yet deprecated Schema 1.
(See the comments in moby/moby PR 42300.)
Docker/Moby is expected to disable pulling Schema 1 images in future after this deprecation.
Signed-off-by: Akihiro Suda <akihiro.suda.cz@hco.ntt.co.jp>
The directory created by `T.TempDir` is automatically removed when the
test and all its subtests complete.
Reference: https://pkg.go.dev/testing#T.TempDir
Signed-off-by: Eng Zer Jun <engzerjun@gmail.com>
This makes it easy for share functionality across tools to prevent
pushing layers that are not supposed to be re-distributed.
Signed-off-by: Brian Goff <cpuguy83@gmail.com>
`dockerPusher` provides `pushWriter` which implements `content.Writer`.
However, even if `pushWriter` become abort status (i.e. `Close()` is called
before `Commit()`), `dockerPusher` doesn't recognise that status and treats that
writer as on-going.
This behaviour doesn't allow the client to retry an aborted push.
This commit fixes this issue.
This commit also adds an test to ensure that the issue is fixed.
Signed-off-by: Kohei Tokunaga <ktokunaga.mail@gmail.com>
Test are being added for GenerateTokenOptions to cover multiple scope cases and the error cases handled.
Signed-off-by: Jacob MacElroy <jacob@okteto.com>
Currently scopes added to token options are added with all scopes included in space delimited string. This changes it so that each scope is added to the string slice as a separate string. This seems to be the desire behavior based on the fact that a string slice is used and the usage of this function in github.com/moby/buildkit.
Signed-off-by: Jacob MacElroy <jacob@okteto.com>
The io/ioutil package has been deprecated as of Go 1.16, see
https://golang.org/doc/go1.16#ioutil. This commit replaces the existing
io/ioutil functions with their new definitions in io and os packages.
Signed-off-by: Eng Zer Jun <engzerjun@gmail.com>
Currently, containerd doesn't restart pull when it encounters unexpected EOF of
blob strem withtout error codes.
There are cases where this lead to pull failure.
This commit tries to fix this issue.
Signed-off-by: Kohei Tokunaga <ktokunaga.mail@gmail.com>
Remove build tags which are already implied by the name of the file.
Ensures build tags are used consistently
Signed-off-by: Derek McGowan <derek@mcg.dev>
Discovered this while using HTTP tracing via OpenTelemetry inside of
buildkitd, where the trace spans were not being reported for the
registry PUT http requests. The spans are only reported on the Close
for the Body, after adding these Close calls, the spans are reported as
expected.
Signed-off-by: coryb <cbennett@netflix.com>
Adds support for mirrors which are non-compliant with the
OCI distribution specification but have previously mirrored
content with a namespace prefix after the API root `/v2`.
Signed-off-by: Derek McGowan <derek@mcg.dev>
Previously, containerd uses Go's default UA "Go-http-client/1.1" while authenticating with registry.
This commit changes it to the pattern like "containerd/v1.5.2" which is used for all other requests.
Signed-off-by: Ethan Chen <pov@mahou-shoujo.moe>
Allows redirects to be authorized if authorization is provided
for the redirected to host. The authorization will always go
to the redirect and never to the referrer.
Signed-off-by: Derek McGowan <derek@mcg.dev>
The `(dockerPusher).Push` method uses a `StatusTracker` to check if an
upload already happened, before repeating the upload. However, there is
no provision for failure handling. If a PUT request returns an error,
the `StatusTracker` will still see the upload as if it happened
successfully. Add a status boolean so that only successful uploads
short-circuit `Push`.
Signed-off-by: Aaron Lehmann <alehmann@netflix.com>
This allows a pusher to be used for more than one tag without creating a
new resolver/pusher. The current implementation checks the ref key
tracker status based on type and hash and will skip the push even if the
repository reference is unique.
Signed-off-by: Phil Estes <estesp@amazon.com>
