dockerrun

package
v0.2.36 Latest Latest
Warning

This package is not in the latest version of its module.

 Go to latest Published: Jul 25, 2026 License: MIT Imports: 31 Imported by: 0

Documentation

Index

Constants

View Source 
const (
	LabelCodeflyOwner     = "codefly.owner"     // always "true"
	LabelCodeflySession   = "codefly.session"   // PID of the spawning CLI
	LabelCodeflyName      = "codefly.name"      // container's logical name
	LabelCodeflyEphemeral = "codefly.ephemeral" // "true" for SDK/test (--cli-server) containers
	// EphemeralContainersEnvironment carries ephemeral lifecycle intent across
	// the CLI → agent process boundary: the CLI owns orchestration, but service
	// agents create the containers. The CLI plants this marker before spawning
	// agents, which inherit it via os.Environ() (see agents/manager loader).
	//
	// Its value is the PID of the process that enabled ephemeral mode, not a
	// bare flag. EphemeralContainers honors an inherited marker only when it
	// names the reader's live parent — scoping propagation to a genuine spawn
	// (one hop, CLI → agent) and ignoring a stale value left in an interactive
	// shell or CI environment. That matters because the marker decides whether
	// running containers are reaped mid-run: a stale ambient value must never
	// silently mark real containers for reaping.
	EphemeralContainersEnvironment = "CODEFLY_EPHEMERAL_CONTAINERS"
)

Container labels used by the startup sweep to identify codefly-owned containers and their spawning CLI. These are set on every container created via DockerEnvironment and consumed by ReapStaleContainers.

Variables

This section is empty.

Functions

func ContainerName 

func ContainerName(name string) string

func DockerEngineRunning 

func DockerEngineRunning(ctx context.Context) bool

func EphemeralContainers 

func EphemeralContainers() bool

EphemeralContainers reports whether this process spawns ephemeral containers. It is true when this process enabled the mode in-process, or when it inherited the marker from its live parent (the process that spawned it). An inherited marker that does not name the current parent is stale and deliberately ignored.

func Forward 

func Forward(ctx context.Context, reader io.Reader, writer io.Writer)

func GetImageID 

func GetImageID(im *resources.DockerImage) (string, error)

func GetImageIfNotPresent 

func GetImageIfNotPresent(ctx context.Context, c *client.Client, imag *resources.DockerImage, out io.Writer) error

func ImageExists 

func ImageExists(ctx context.Context, c *client.Client, imag *resources.DockerImage) (bool, error)

func NewClient added in v0.2.13 

func NewClient() (*client.Client, error)

NewClient builds a Docker client that honors the active docker context, mirroring the docker CLI's endpoint resolution (see resolveDockerHost). It exists so packages outside dockerrun — notably the image builder — reach the same engine as the runners instead of client.FromEnv's context-unaware default.

func PrintDownloadPercentage 

func PrintDownloadPercentage(reader io.ReadCloser, out io.Writer) error

func ReapStaleContainers 

func ReapStaleContainers(ctx context.Context) error

func SetEphemeralContainers 

func SetEphemeralContainers(v bool)

SetEphemeralContainers marks dependency containers spawned by this process as ephemeral so the startup sweep reaps them even while running. The CLI calls this before spawning agents, which inherit the environment marker.

This is owned by the process that establishes the run (the CLI). Enabling records this process's PID in the environment so freshly spawned children inherit the scoped marker; disabling clears both the in-process flag and the marker for this process and its future children. Agents query state via EphemeralContainers and must not call this.

Types

type DockerContainerInstance 

type DockerContainerInstance struct {
	ID string
}

type DockerEnvironment 

type DockerEnvironment struct {
	// contains filtered or unexported fields
}

func NewDockerEnvironment 

func NewDockerEnvironment(ctx context.Context, image *resources.DockerImage, dir, name string) (*DockerEnvironment, error)

NewDockerEnvironment creates a new docker environment.

If GetImageIfNotPresent fails, the docker client is closed before returning the error so the caller doesn't have to know about the half-constructed env (and isn't expected to call Shutdown on a nil return). Same defensive close in NewDockerHeadlessEnvironment.

func NewDockerHeadlessEnvironment 

func NewDockerHeadlessEnvironment(ctx context.Context, image *resources.DockerImage, name string) (*DockerEnvironment, error)

NewDockerHeadlessEnvironment creates a new docker runner

func (*DockerEnvironment) ContainerDeleted 

func (docker *DockerEnvironment) ContainerDeleted() (bool, error)

ContainerDeleted checks if the container with ID is gone

func (*DockerEnvironment) ContainerID 

func (docker *DockerEnvironment) ContainerID() (string, error)

func (*DockerEnvironment) GetContainer 

func (docker *DockerEnvironment) GetContainer(ctx context.Context) error

func (*DockerEnvironment) GetImageIfNotPresent 

func (docker *DockerEnvironment) GetImageIfNotPresent(ctx context.Context, imag *resources.DockerImage) error

func (*DockerEnvironment) GetLogs 

func (docker *DockerEnvironment) GetLogs(ctx context.Context) error

func (*DockerEnvironment) ImageExists 

func (docker *DockerEnvironment) ImageExists(ctx context.Context, imag *resources.DockerImage) (bool, error)

func (*DockerEnvironment) Init 

func (docker *DockerEnvironment) Init(ctx context.Context) error

func (*DockerEnvironment) IsContainerPresent 

func (docker *DockerEnvironment) IsContainerPresent(ctx context.Context) (bool, error)

func (*DockerEnvironment) NewProcess 

func (docker *DockerEnvironment) NewProcess(bin string, args ...string) (base.Proc, error)

func (*DockerEnvironment) PortMappings 

func (docker *DockerEnvironment) PortMappings() []*DockerPortMapping

func (*DockerEnvironment) Shutdown 

func (docker *DockerEnvironment) Shutdown(ctx context.Context) error

func (*DockerEnvironment) Stop 

func (docker *DockerEnvironment) Stop(ctx context.Context) error

func (*DockerEnvironment) TailLogs 

func (docker *DockerEnvironment) TailLogs(ctx context.Context, lines int) string

TailLogs returns the last `lines` lines of combined stdout+stderr for the container. Used by agent runtimes to enrich startup-failure errors with the actual container output (otherwise the user only sees a generic "not ready" timeout). Returns "" on any error so callers can safely append without conditional logic.

func (*DockerEnvironment) WithBinary 

func (docker *DockerEnvironment) WithBinary(bin string) error

func (*DockerEnvironment) WithCommand 

func (docker *DockerEnvironment) WithCommand(cmd ...string)

func (*DockerEnvironment) WithDir 

func (docker *DockerEnvironment) WithDir(dir string)

func (*DockerEnvironment) WithDockerSocket added in v0.2.15 

func (docker *DockerEnvironment) WithDockerSocket() *DockerEnvironment

WithDockerSocket explicitly grants the container control of the host Docker daemon. The socket is root-equivalent host access, so ordinary service and database containers never receive it by default.

func (*DockerEnvironment) WithEnvironmentVariables 

func (docker *DockerEnvironment) WithEnvironmentVariables(ctx context.Context, envs ...*resources.EnvironmentVariable)

func (*DockerEnvironment) WithEphemeral added in v0.2.11 

func (docker *DockerEnvironment) WithEphemeral() *DockerEnvironment

WithEphemeral marks this environment's container as ephemeral, so the startup sweep reaps it even when still running with a dead owner (no keep-alive reuse). For infra with no persistent state to preserve across runs.

func (*DockerEnvironment) WithMount 

func (docker *DockerEnvironment) WithMount(sourceDir string, targetDir string)

func (*DockerEnvironment) WithOutput 

func (docker *DockerEnvironment) WithOutput(w io.Writer)

func (*DockerEnvironment) WithPause 

func (docker *DockerEnvironment) WithPause()

func (*DockerEnvironment) WithPort 

func (docker *DockerEnvironment) WithPort(ctx context.Context, port uint16)

func (*DockerEnvironment) WithPortMapping 

func (docker *DockerEnvironment) WithPortMapping(ctx context.Context, local uint16, container uint16)

func (*DockerEnvironment) WithPublicPorts added in v0.2.15 

func (docker *DockerEnvironment) WithPublicPorts() *DockerEnvironment

WithPublicPorts explicitly publishes subsequent Docker port mappings on all host interfaces. The secure default is loopback-only because Codefly's service and companion ports are consumed by the local CLI and workspace.

func (*DockerEnvironment) WithUser added in v0.2.35 

func (docker *DockerEnvironment) WithUser(user string)

WithUser runs the container as the given user, in Docker's "UID[:GID]" form. Bind-mounted output is then owned by that user instead of root, so a non-root host process can read (and clean up) what the container wrote. Empty string keeps the image's default user.

func (*DockerEnvironment) WithWorkDir 

func (docker *DockerEnvironment) WithWorkDir(dir string)

type DockerPortMapping 

type DockerPortMapping struct {
	Host      uint16
	Container uint16
}

type DockerProc 

type DockerProc struct {
	ID string
	// contains filtered or unexported fields
}

func (*DockerProc) FindPid 

func (proc *DockerProc) FindPid(ctx context.Context) (int, error)

FindPid locates the PID of the matching child process inside the container by scanning /proc. /proc is guaranteed on every Linux container; `ps` is not (debian-slim images ship without procps, Alpine busybox has a different flag set). Output format is one line per PID: 

<pid>\t<cmdline-space-separated>

where cmdline is the full argv reconstructed from /proc/<pid>/cmdline (NUL-separated in the kernel — we convert NUL→space for splitting).

func (*DockerProc) IsRunning 

func (proc *DockerProc) IsRunning(ctx context.Context) (bool, error)

func (*DockerProc) Match 

func (proc *DockerProc) Match(cmd []string) bool

func (*DockerProc) Run 

func (proc *DockerProc) Run(ctx context.Context) error

func (*DockerProc) Start 

func (proc *DockerProc) Start(ctx context.Context) error

func (*DockerProc) StdinPipe 

func (proc *DockerProc) StdinPipe() (io.WriteCloser, error)

func (*DockerProc) StdoutPipe 

func (proc *DockerProc) StdoutPipe() (io.ReadCloser, error)

func (*DockerProc) Stop 

func (proc *DockerProc) Stop(ctx context.Context) error

func (*DockerProc) Wait 

func (proc *DockerProc) Wait(ctx context.Context) error

Wait blocks until the container exec exits or ctx is cancelled. Returns the exec's exit error: nil for a clean (exit code 0) exit, non-nil carrying the code otherwise — matching Run and the Proc.Wait contract. The exit status comes from ContainerExecInspect (the authoritative source Run polls too), which stays queryable after the exec finishes, so repeated or late Wait calls all observe the same result. Polling IsRunning (the previous approach) scanned /proc and could only ever return nil or ctx.Err(), so a supervisor never saw a non-zero exit.

func (*DockerProc) WaitOn 

func (proc *DockerProc) WaitOn(bin string)

func (*DockerProc) WithDir 

func (proc *DockerProc) WithDir(dir string)

func (*DockerProc) WithEnvironmentVariables 

func (proc *DockerProc) WithEnvironmentVariables(ctx context.Context, envs ...*resources.EnvironmentVariable)

func (*DockerProc) WithEnvironmentVariablesAppend 

func (proc *DockerProc) WithEnvironmentVariablesAppend(ctx context.Context, added *resources.EnvironmentVariable, sep string)

func (*DockerProc) WithOutput 

func (proc *DockerProc) WithOutput(output io.Writer)

type DockerPullResponse 

type DockerPullResponse struct {
	ID             string         `json:"id"`
	Status         string         `json:"status"`
	ProgressDetail ProgressDetail `json:"progressDetail"`
	Error          string         `json:"error"`
	ErrorDetail    struct {
		Message string `json:"message"`
	} `json:"errorDetail"`
}

type ProgressDetail 

type ProgressDetail struct {
	Current int `json:"current"`
	Total   int `json:"total"`
}

Source Files

View all Source files

Jump to

Keyboard shortcuts

? : This menu
/ : Search site
f or F : Jump to
y or Y : Canonical URL
go.dev uses cookies from Google to deliver and enhance the quality of its services and to analyze traffic. Learn more.