feat: upgrade dependencies and migrate deprecated APIs to new ones#139
feat: upgrade dependencies and migrate deprecated APIs to new ones#139kezhenxu94 wants to merge 8 commits intomainfrom
Conversation
- Upgrade testcontainers-go from v0.11.1 to v0.40.0 - Upgrade docker/docker from v20.10.7 to v28.5.2 - Upgrade k8s.io/* packages from v0.22.2 to v0.35.1 - Upgrade sigs.k8s.io/kind from v0.27.0 to v0.31.0 - Migrate deprecated LocalDockerCompose API to new ComposeStack API - Replace deprecated types.EventsOptions with events.ListOptions - Replace deprecated types.Container* types with container.* types - Replace deprecated types.Network* types with network.* types - Update RESTMapper.NewShortcutExpander to use warning callback - Add parseComposeFile() to parse compose files directly with yaml.v3 - Add loadEnvFromFile() to load environment variables from files
There was a problem hiding this comment.
Pull request overview
This PR upgrades several core dependencies (testcontainers-go, docker client, Kubernetes libs, kind) and migrates the codebase off deprecated APIs, especially around docker-compose orchestration and Docker SDK type packages.
Changes:
- Migrates deprecated Testcontainers
LocalDockerComposeusage to themodules/composestack API. - Updates Docker SDK usages from deprecated
types.*to newercontainer.*,network.*,events.*, andimage.*packages. - Updates Kubernetes RESTMapper shortcut expander usage to the newer callback-enabled API.
Reviewed changes
Copilot reviewed 13 out of 14 changed files in this pull request and generated 8 comments.
Show a summary per file
| File | Description |
|---|---|
| internal/util/k8s.go | Updates RESTMapper shortcut expander to new signature with warning callback. |
| internal/components/setup/kind.go | Migrates Docker image list/pull to types/image option types. |
| internal/components/setup/compose_provider.go | Updates Docker container/network types and expands wait.StrategyTarget implementation surface. |
| internal/components/setup/compose_listener.go | Updates Docker events options/types and adjusts event message ID access. |
| internal/components/setup/compose.go | Reworks compose startup to use compose module stacks and adds compose/env parsing helpers. |
| internal/components/cleanup/compose.go | Migrates compose cleanup to compose module stack teardown. |
| go.mod | Upgrades module dependencies and adds replace directives for Docker archive compatibility. |
Comments suppressed due to low confidence (1)
internal/components/setup/compose_provider.go:185
inspectContainerduplicates the sameContainerInspectcall added inInspect. To reduce duplication and keep behavior consistent, consider implementinginspectContainerby callingInspect(or remove one of them if only one is needed).
func (c *DockerContainer) inspectContainer(ctx context.Context) (*container.InspectResponse, error) {
inspect, err := c.provider.client.ContainerInspect(ctx, c.ID)
if err != nil {
return nil, err
}
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
| func (c *ComposeContainerListener) Listen(consumer func(container *ComposeContainer)) error { | ||
| containerEvents, errors := c.client.Events(c.ctx, types.EventsOptions{ | ||
| containerEvents, errors := c.client.Events(c.ctx, events.ListOptions{ | ||
| Filters: filters.NewArgs( | ||
| filters.Arg("type", "container"), | ||
| filters.Arg("event", "start"), |
There was a problem hiding this comment.
client.Events returns channels; downstream handling should not rely on len(errCh) to detect failures and should also handle channel closure (ok checks) to avoid spinning on closed channels. Consider restructuring Listen to propagate the first error deterministically (e.g., via select on errCh) and exit the goroutine when channels close.
| @@ -68,27 +69,43 @@ func ComposeSetup(e2eConfig *config.E2EConfig) error { | |||
| return err | |||
| } | |||
There was a problem hiding this comment.
The Docker client created here is never closed. The Docker SDK client holds underlying HTTP connections; please defer cli.Close() after successful creation (similar to pullImages in kind.go) to avoid leaking resources in long-running processes/tests.
| } | |
| } | |
| defer cli.Close() |
| envVars, err := loadEnvFromFile(profilePath) | ||
| if err != nil { | ||
| return fmt.Errorf("load env file error: %v", err) | ||
| } |
There was a problem hiding this comment.
Same as above: this returns a propagated error but uses %v instead of %w, which prevents callers from unwrapping the underlying error.
| // Compatibility fix for docker/cli which imports github.com/docker/docker/pkg/archive | ||
| // but the package has been moved to github.com/moby/go-archive | ||
| replace github.com/moby/go-archive => github.com/moby/go-archive v0.1.0 | ||
|
|
||
| replace github.com/docker/docker/pkg/archive => github.com/moby/go-archive v0.1.0 |
There was a problem hiding this comment.
This self-replace (replace X => X v...) is effectively a version pin but is non-idiomatic and easy to misread as a no-op. Prefer adding an explicit require github.com/moby/go-archive v0.1.0 (or letting go mod tidy select) and keep replace only for the actual path substitution that’s needed.
| // Listen container create | ||
| listener := NewComposeContainerListener(context.Background(), cli, services) |
There was a problem hiding this comment.
The callback passed to listener.Listen (below) mutates the outer err variable from another goroutine, which is a data race and can overwrite the function’s control-flow error value. Use a local variable inside the callback to handle/log errors without assigning to the outer err.
| composeServices, err := parseComposeFile(composeConfigPath) | ||
| if err != nil { | ||
| return fmt.Errorf("parse compose file error: %v", err) | ||
| } |
There was a problem hiding this comment.
These errors are formatted with %v, which loses wrapping semantics and makes errors.Is/As harder upstream. Prefer %w when returning an error you’re propagating (as done elsewhere in this file for other failures).
| } | ||
|
|
||
| // No networks found | ||
| return "", nil |
There was a problem hiding this comment.
When no networks are found, this returns an empty IP with a nil error. That can lead to confusing downstream failures (treating "" as a valid IP). Prefer returning a non-nil error when the container has no network/IP information available.
| return "", nil | |
| return "", fmt.Errorf("container has no network/IP information available") |
| func (c *DockerContainer) Exec(ctx context.Context, cmd []string, options ...exec.ProcessOption) (int, io.Reader, error) { | ||
| cli := c.provider.client | ||
| response, err := cli.ContainerExecCreate(ctx, c.ID, types.ExecConfig{ | ||
| Cmd: cmd, | ||
| Detach: false, | ||
| response, err := cli.ContainerExecCreate(ctx, c.ID, container.ExecOptions{ | ||
| Cmd: cmd, | ||
| Detach: false, |
There was a problem hiding this comment.
Exec accepts options ...exec.ProcessOption and returns an io.Reader, but the implementation currently ignores all options and never attaches to capture stdout/stderr (it later returns a dummy reader). This can break callers/strategies that rely on ProcessOption behavior or output content. Consider applying options to the exec configuration and using ContainerExecAttach (or otherwise documenting/removing the output return).
…ndencies - Delete compose_provider.go (601 lines) — custom DockerContainer/DockerProvider re-implementation replaced by native testcontainers-go APIs - Delete compose_listener.go (85 lines) — Docker event listener no longer needed; containers are accessible after stack.Up() via stack.ServiceContainer() - Rewrite compose.go to use compose.NewDockerComposeWith + stack.Up/ServiceContainer - Rewrite cleanup/compose.go to use stack.Down - Upgrade testcontainers-go v0.11.1 → v0.42.0 - Upgrade docker/docker v20.10.7 → v28.5.2 - Upgrade k8s.io/* v0.22.2 → v0.35.3 - Upgrade sigs.k8s.io/kind v0.27.0 → v0.31.0 - Upgrade spf13/cobra v1.8.0 → v1.10.2 - Remove docker-compose v1 binary install from CI (v0.42.0 uses Docker Compose v2 plugin) - Update action.yaml Go version to 1.26 Closes #139 Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Split Build workflow into lint-and-test + build-platform (ubuntu/windows/macos) matrix jobs. Enable CGO for native builds. Update Dockerfile with build-essential for CGO. Change `make build` to build for current OS only (CI matrix handles cross-platform). Based on #139 Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
|
Closed in favor of #146, which completed the full migration to testcontainers-go v0.42.0 with all dependency upgrades, multi-platform CI build matrix, and CGO support originally proposed here. |
…146) * feat: migrate to testcontainers-go v0.42.0 and upgrade all major dependencies - Delete compose_provider.go (601 lines) — custom DockerContainer/DockerProvider re-implementation replaced by native testcontainers-go APIs - Delete compose_listener.go (85 lines) — Docker event listener no longer needed; containers are accessible after stack.Up() via stack.ServiceContainer() - Rewrite compose.go to use compose.NewDockerComposeWith + stack.Up/ServiceContainer - Rewrite cleanup/compose.go to use stack.Down - Upgrade testcontainers-go v0.11.1 → v0.42.0 - Upgrade docker/docker v20.10.7 → v28.5.2 - Upgrade k8s.io/* v0.22.2 → v0.35.3 - Upgrade sigs.k8s.io/kind v0.27.0 → v0.31.0 - Upgrade spf13/cobra v1.8.0 → v1.10.2 - Remove docker-compose v1 binary install from CI (v0.42.0 uses Docker Compose v2 plugin) - Update action.yaml Go version to 1.26 Closes #139 Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * chore: add multi-platform CI build matrix and CGO_ENABLED support Split Build workflow into lint-and-test + build-platform (ubuntu/windows/macos) matrix jobs. Enable CGO for native builds. Update Dockerfile with build-essential for CGO. Change `make build` to build for current OS only (CI matrix handles cross-platform). Based on #139 Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * fix: remove compose.Wait(true) to avoid requiring all containers healthy compose.Wait(true) maps to `docker compose up --wait` which requires ALL containers to remain running and healthy. This is too strict for e2e tests where some containers may not have healthchecks or may exit after init. Port readiness is already handled by WaitForService() with per-port wait strategies. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * chore: remove containerd image store disable workaround No longer needed — testcontainers-go v0.42.0 uses Docker Compose v2 plugin which fully supports the containerd image store. Reverts the workaround added in 8c21e43 (#140). Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * fix: add manual port waiting after compose up WaitForService strategies only execute when compose.Wait(true) is passed, but compose.Wait(true) is too strict (requires all containers healthy). Instead, manually wait for each port via TCP dial after Up() returns, matching the old code's behavior of waitPortUntilReady(). Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * fix: restore DOCKER_API_VERSION negotiation in action.yaml Docker SDK v28.5.2 defaults to API v1.51, but GitHub Actions runners may only support up to v1.48. Set DOCKER_API_VERSION to the daemon's supported version to ensure compatibility. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * fix: add logging to port waiting and try immediately before polling Add INFO-level logs for port wait start/completion and DEBUG-level for retry attempts. Also try connecting immediately before the first ticker interval to avoid unnecessary 500ms delay when ports are already ready. Log when a service has no ports to wait for. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * fix: wait for Docker healthcheck before port readiness check After stack.Up(), wait for each container's Docker healthcheck to pass (if defined) before doing TCP port checks. This matches the old behavior where docker-compose up -d would respect depends_on: condition: service_healthy. Also restored internal container port check (/proc/net/tcp + nc + /dev/tcp) from the old WaitPort implementation to align with original behavior. The wait sequence is now: 1. stack.Up(ctx) — start containers 2. waitForHealthy — poll Docker health status until healthy (if healthcheck defined) 3. waitForPort — external TCP dial + internal container check 4. Export env vars Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * fix: only wait for healthcheck on services with ports The old code only waited for services that had ports: in docker-compose.yml. Services without ports were completely skipped. The healthcheck wait was incorrectly running on ALL services, causing timeouts on services like init containers or senders whose healthchecks depend on other services. Now: services without ports get log streaming only. Services with ports get healthcheck wait + port wait + env export (matching old behavior). Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * fix: remove healthcheck waiting, use only TCP+internal port check The old code never checked Docker healthcheck status. It only did TCP dial + internal /proc/net/tcp check. Many compose files use healthchecks with tools (nc) not available in the container image, causing Docker to mark containers as unhealthy even when the application is running fine. Remove waitForHealthy entirely to match old behavior exactly: port-level waiting only, no Docker healthcheck dependency. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * fix: restore build target to build all platforms for release The build target must build all platforms (windows/linux/darwin) because release-bin depends on it. Each platform build uses CGO_FLAG to enable CGO only for native builds. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * fix: build only current platform per CI matrix runner Each matrix runner (ubuntu/windows/macos) now builds only its own platform instead of all three redundantly. The `make build` target is preserved for release which needs all platforms. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * ci: simplify build matrix by moving OS-to-target mapping into matrix include --------- Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com> Co-authored-by: kezhenxu94 <kezhenxu94@apache.org>
Summary
LocalDockerComposeAPI to newComposeStackAPItypes.*with newcontainer.*,network.*,events.*,image.*packagesRESTMapper.NewShortcutExpanderto use warning callbackTest plan
go build ./...passesgo test ./...passes