CLI reference

Getting started
Requirements
Python 3.5 or newer
Java 8 or newer
Install
The Launchable CLI is a Python3 package that can be installed via pip:
1
$ pip3 install --user --upgrade launchable~=1.0
Copied!
This creates a ~/.local/bin/launchable executable that should be in your PATH. (See PEP-370 for further details.)
Authenticate
Set your API key:
1
$ export LAUNCHABLE_TOKEN=your_API_key
Copied!
Verify
Then run launchable verify in your CI environment to see if you've successfully configured the CLI. If it succeeds, you'll see a message like the one below. If you see an error message, see Troubleshooting.
1
$ launchable verify
2
​
3
Organization: <organization>
4
Workspace: <workspace>
5
Proxy: None
6
Platform: macOS-11.5.2-x86_64-i386-64bit
7
Python version: 3.9.6
8
Java command: java
9
launchable version: 1.28.0
10
Your CLI configuration is successfully verified 🎉
Copied!
Commands
inspect subset
Display the details of a subset request. See Subsetting your test runs for more info.
1
$ launchable inspect subset --subset-id 26876
Copied!
Option
Description
Required
--subset-id ID
The ID of the subset request. Can be obtained from the output of launchable subset.
Yes
inspect tests
Display the details of a record tests command. See Sending data to Launchable for more info.
1
$ launchable inspect tests --test-session-id 209575
Copied!
Option
Description
Required
--test-session-id ID
The ID of the test session. Can be obtained from the output of launchable record tests.
Yes
record commit
Sends commit details to Launchable.
1
$ launchable record commit --source src=.
Copied!
Option
Description
Required
--executable jar
Run commit collection using Java.
No (default)
--executable docker
Run commit collection using Docker.
No
--max-days DAYS
The maximum number of days to collect commits retroactively.
No. Defaults to 30
--source REPO_NAME=DIR
Name and path of a local Git repository.
No. Defaults to $PWD
Commit collection happens automatically as a part of record build, so normally this command need not be invoked separately.
--executable jar is faster as the Java version of the commit collector is bundled with the CLI, but it requires that your system has Java installed. --executable docker allows you to run the equivalent commit collector packaged as a Docker image. You may choose to do this if your system allows you to run Docker containers but not Java. Containers will be downloaded on demand. This option is more of an escape hatch.
record build
Creates a record of a build in Launchable.
1
$ launchable record build [OPTIONS]
Copied!
Option
Description
Required
--name BUILD_NAME
Unique identifier that you assign to your build. See [Naming builds](build-names.md) for more discussion of how to choose a build name.
Yes
--max-days DAYS
The maximum number of days to collect commits retroactively.
No. Defaults to 30
--no-submodules
Stop collecting build information from Git Submodules.
No. Defaults to False
--source main=path/to/ws (recommanded) or --source path/to/ws
Path to a local Git repository/workspace. Use this option multiple times when code from multiple Git repositories are contributing to the build. Note that Git submodules are discovered and recorded automatically, so there's no need to enumerate them separately.
To distinguish different Git repositories, every repository is labeled internally in Launchable. By default, the literal path given to this option is used as a label (for example, label would be dir/source for --source dir/source). We recommand naming labels explicitly (e.g. to keep them stable even when directory names move around), by prepending a label name followed by =, such as --source vendor=$VENDOR_PATH.
Yes
The act of recording a build teaches Launchable that the specified set of commits have turned into a build, and that this build is henceforth identified by the given name. This forms the basis of how Launchable calculates the changes.
For more details about what we do with commits and what we don't, see this block in the getting started guide.
record session
Creates a record of a test session in Launchable.
1
$ launchable record session [OPTIONS]
Copied!
Option
Description
Required
--build BUILD_NAME
Name of the build being tested. (See record build --name)
Yes
This command tells Launchable that you are about to begin testing a build that was been recorded earlier with the record build command. This is only needed in more complex scenarios.
The command writes out a session ID to ~/.config/launchable/sessions/{hash}.txt. Subsequent commands read the session ID from this file.
split-subset
Retrieves a specific portion of an existing subset from Launchable. See replacing static parallel suites with a dynamic parallel subset.
1
launchable split-subset [OPTIONS] TESTRUNNER ...
Copied!
Option
Description
Required
--subset-id SUBSETID
ID of the subset output from launchable subset --split ... (see --split under subset)
Yes
--bin BIN_NUMBER/BIN_COUNT
The portion of the subset to retrieve
Yes
--rest FILE
Output the remainder of the subset to a file. This is useful for running the "rest of the tests" after you've run a subset.
No
subset
Produces a subset of tests to pass to your test runner.
1
launchable subset [OPTIONS] TESTRUNNER ...
Copied!
Option
Description
Required
--build BUILD_NAME
Name of the build being tested. (See record build --name)
One of --build or --session is required
--session SESSIONID
ID of the test session (see record session)
One of --build or --session is required
--base DIR
Advanced option. A large number of test runners use file names to identify tests, and for those test runners, so does Launchable. By default Launchable record test file names as given to it; IOW we expect those to be relative paths, so that identities of tests remain stable no matter where in the file system a Git workspace gets checked out. But in the rare circumstances where this behavior is inadequate, the --base option lets you specify a separate directory to relativize the path of tests before recording them.
No
--target PERCENTAGE
Create a variable time-based subset of the given percentage. (0%-100%)
One of --target, --time or --confidence is required
--time TIME
Create a fixed time-based subset. Select the best set of tests that run within the given time bound. (e.g. 10m for 10 minutes, 2h30m for 2.5 hours, 1w3d for 7+3=10 days. )
One of --target, --time or --confidence is required
--confidence PERCENTAGE
Create a confidence-based subset of the given percentage. (0%-100%)
One of --target, --time or --confidence is required
--flavor KEY=VALUE
Advanced option. Restrict the subset of tests by flavor. Flavors must be submitted ahead of time with test reports (see launchable record tests --flavor below). Supports multiples, e.g. --flavor key1=value1 --flavor key2=value2.
No
--rest FILE
Output the remainder of the subset to a file. This is useful for running the "rest of the tests" after you've run a subset.
No
--split
Output a subset ID instead of the subset list itself. For use with launchable split-subset
No
Exactly how this command generates the subset and what's required to do this depends on test runners. For available supported TESTRUNNERs, see Integrations.
record tests
Send test results for the test session to Launchable.
1
launchable record tests [OPTIONS] TESTRUNNER ...
Copied!
Option
Description
Required
--build BUILD_NAME
Name of the build being tested. (See record build --name)
One of --build or --session is required
--session SESSIONID
ID of the test session (see record session)
One of --build or --session is required
--flavor KEY=VALUE
Advanced option. Submit additional non-code-related metadata that influenced the test results, such as environment. To be used in combination with launchable subset --flavor. Supports multiples, e.g. --flavor key1=value1 --flavor key2=value2.
No
--base DIR
See the explanation of launchable subset --base option above.
No
This command reads JUnit (or similar) XML report files produced by test runners and sends them to Launchable.
Exactly how this command generates the subset and what's required to do this depends on test runners. For available supported TESTRUNNER, see Integrations​
verify
Verify that the CLI can communicate with the Launchable service and that you're authenticated properly.
1
launchable verify
Copied!
In order to avoid disrupting your CI/test process, the Launchable CLI is designed to tolerate & recover from service disruptions and other recoverable error conditions by falling back to no-op. This is an intentional design, but the downside is that such transparent failures can make troubleshooting difficult.
Therefore, we recommend you keep launchable verify || true in a recognizable spot in your CI process. This way, when you suspect a problem in Launchable, you can check the output of this command as a starting point.
Global options
--log-level
You can use the --log-level option to output extra information from each command.
--log-level audit is particularly useful if you want to see exactly what data gets passed to Launchable when you run CLI commands. For example:
1
% launchable --log-level audit record build --name 1234 --source src=.
2
Processed 1 commits
3
AUDIT:launchable:send request method:post path:/intake/organizations/launchableinc/workspaces/awilkes/builds headers:{'Content-Type': 'application/json'} args:{'data': b'{"buildNumber": "1234", "commitHashes": [{"repositoryName": "src", "commitHash": "45b2e6d9df8e0013334354f30df1978c8b4196f8"}]}', 'timeout': (5, 60)}
Copied!