A Taskcluster client library for the command line.
This directory builds an executable, taskcluster
, suitable for interacting
with Taskcluster from the comfort of you command-line and in simple shell
scripts. It provides utilities ranging from direct calls to the specific API
endpoints to more complex and practical tasks like listing and cancelling
scheduled runs.
To install, Linux users should download the taskcluster
binary for the latest release your
platform, run chmod +x
and run it!
MacOS users run the following command:
curl -L https://github.com/taskcluster/taskcluster/releases/download/v39.1.1/taskcluster-darwin-amd64 --output taskcluster
This is to ensure the binary is not quarantined by MacOS.
You will need to chmod +x
of
course.
For a list of all commands run taskcluster help
, detailed information about
each command is available with
taskcluster help <command> [<sub-command> [...]]
. You can also use the -h
or --help
parameter to get a command's help information.
For a general guide to using Taskcluster clients, see Calling Taskcluster APIs.
Provide Taskcluster credentials to this tool with environment variables.
At least TASKCLUSTER_ROOT_URL
must be given.
For API calls that require authentication, additionally TASKCLUSTER_CLIENT_ID
, TASKCLUSTER_ACCESS_TOKEN
, and perhaps TASKCLUSTER_CERTIFICATE
are also required.
The taskcluster signin
command provides an easy method to get credentials for use with this tool
See below.
To call an API method, use the taskcluster api <service> <apiMethod>
subcommand.
Help for these commands is extensive and based on the reference documentation; try
taskcluster api --help
taskcluster api auth --help
taskcluster api auth createClient --help
Positional URL arguments are given on the command line, with query arguments
given with options (e.g., --limit
). Methods that expect a payload body will
read that body in JSON format from stdin. Response bodies are written to
stdout in JSON, or to the destination file given by -o
.
jq
is a useful tool for dealing with JSON
inputs and outputs.
The taskcluster signin
subcommand provides an easy way to get credentials encoded into environment variables via a browser session.
$ eval `taskcluster signin`
This will open a web browser to get credentials, then set the corresponding environment variables in your shell session.
You might make this easy to use with a shell function in ~/.bashrc:
tc-signin() { eval `taskcluster signin "${@}"`; }
It's common to pass a --name
(to help you differentiate clients from one another) and one or more --scope
arguments:
tc-signin --name smoketest --scope assume:project:taskcluster:smoketests
See the taskcluster signin --help
output or Calling Taskcluster APIs for more information.
The taskcluster from-now
subcommand can be used to generate timestamps relative to the current time. For example:
echo '{"expires": "'`taskcluster from-now 1 hour`'", ...}' | taskcluster api ..
The taskcluster slugid
subcommand can generate (and encode and decode) slugids.
To generate a nice slugid:
taskcluster slugid generate -n
The following higher-level commands can be useful in day-to-day operations.
This list may be incomplete; consult taskcluster --help
for the full list.
taskcluster group cancel
- cancel a whole task group by taskGroupId.taskcluster group list
- list tasks (taskId and label) in a task grouptaskcluster group status
- show the status of a task grouptaskcluster task artifacts
- get the name of the artifacts of a task.taskcluster task cancel
- cancel a task.taskcluster task complete
- completes a task.taskcluster task def
- get the full definition of a task.taskcluster task group
- get the taskGroupID of a task.taskcluster task log
- streams the log until completion.taskcluster task name
- get the name of a task.taskcluster task rerun
- rerun a task.taskcluster task retrigger
- re-trigger a task (new taskId, updated timestamps).taskcluster task run
- create and schedule a task through a 'docker run'-like interface.taskcluster task status
- get the status of a task.
This library is co-versioned with Taskcluster itself. That is, a client with version x.y.z contains API methods corresponding to Taskcluster version x.y.z. Taskcluster is careful to maintain API compatibility, and guarantees it within a major version. That means that any client with version x.* will work against any Taskcluster services at version x.*, and is very likely to work for many other major versions of the Taskcluster services. Any incompatibilities are noted in the Changelog.
This package requires Go version 1.12 and uses Go Modules.
To build the client, clone the Taskcluster repository, switch to the clients/client-shell
directory, and run go build -o taskcluster .
.
The API specifications are generated automatically as part of running yarn generate
in the root directory of this repository.
We are using cobra to manage the various commands and sub-commands that are implemented in taskcluster-cli.
Each command is a instance of the cobra.Command
struct, and is dynamically
registered at run-time in the command tree (in func init() {...}
). Thus,
commands are registered as an import side-effect. Commands are implemented in
sub-packages.
To add a new command, create a new sub-package under cmds
and add an import
for that sub-package to subtree_import.go
, keeping the imports in order.