test¶
Run tests.
Backend: pants.core
Config section: [test]
Basic options¶
debug¶
--[no-]test-debug
PANTS_TEST_DEBUG
default: False
Run tests sequentially in an interactive process. This is necessary, for example, when you add breakpoints to your code.
debug_adapter¶
--[no-]test-debug-adapter
PANTS_TEST_DEBUG_ADAPTER
default: False
Run tests sequentially in an interactive process, using a Debug Adapter (https://microsoft.github.io/debug-adapter-protocol/) for the language if supported.
The interactive process used will be immediately blocked waiting for a client before continuing.
This option implies --debug.
force¶
--[no-]test-force
PANTS_TEST_FORCE
default: False
Force the tests to run, even if they could be satisfied from cache.
output¶
--test-output=<ShowOutput>
PANTS_TEST_OUTPUT
one of: all, failed, none
default: failed
Show stdout/stderr for these tests.
use_coverage¶
--[no-]test-use-coverage
PANTS_TEST_USE_COVERAGE
default: False
Generate a coverage report if the test runner supports it.
open_coverage¶
--[no-]test-open-coverage
PANTS_TEST_OPEN_COVERAGE
default: False
If a coverage report file is generated, open it on the local system if the system supports this.
shard¶
--test-shard=<str>
PANTS_TEST_SHARD
default: ``
A shard specification of the form "k/N", where N is a positive integer and k is a non-negative integer less than N.
If set, the request input targets will be deterministically partitioned into N disjoint subsets of roughly equal size, and only the k'th subset will be used, with all others discarded.
Useful for splitting large numbers of test files across multiple machines in CI. For example, you can run three shards with --shard=0/3, --shard=1/3, --shard=2/3.
Note that the shards are roughly equal in size as measured by number of files. No attempt is made to consider the size of different files, the time they have taken to run in the past, or other such sophisticated measures.
timeouts¶
--[no-]test-timeouts
PANTS_TEST_TIMEOUTS
default: True
Enable test target timeouts. If timeouts are enabled then test targets with a timeout= parameter set on their target will time out after the given number of seconds if not completed. If no timeout is set, then either the default timeout is used or no timeout is configured.
extra_env_vars¶
--test-extra-env-vars="['<str>', '<str>', ...]"
PANTS_TEST_EXTRA_ENV_VARS
default: []
Additional environment variables to include in test processes. Entries are strings in the form ENV_VAR=value to use explicitly; or just ENV_VAR to copy the value of a variable in Pants's own environment.
Can be overridden by field test_extra_env_vars on local_environment, docker_environment, or remote_environment targets.
Advanced options¶
report¶
--[no-]test-report
PANTS_TEST_REPORT
default: False
Write test reports to --report-dir.
report_dir¶
--test-report-dir=<str>
PANTS_TEST_REPORT_DIR
default: {distdir}/test/reports
Path to write test reports to. Must be relative to the build root.
timeout_default¶
--test-timeout-default=<int>
PANTS_TEST_TIMEOUT_DEFAULT
default: None
The default timeout (in seconds) for a test target if the timeout field is not set on the target.
timeout_maximum¶
--test-timeout-maximum=<int>
PANTS_TEST_TIMEOUT_MAXIMUM
default: None
The maximum timeout (in seconds) that may be used on a test target.
batch_size¶
--test-batch-size=<int>
PANTS_TEST_BATCH_SIZE
default: 128
The target maximum number of files to be included in each run of batch-enabled test runners.
Some test runners can execute tests from multiple files in a single run. Test implementations will return all tests that can run together as a single group - and then this may be further divided into smaller batches, based on this option. This is done:
1. to avoid OS argument length limits (in processes which don't support argument files)
2. to support more stable cache keys than would be possible if all files were operated on in a single batch
3. to allow for parallelism in test runners which don't have internal parallelism, or -- if they do support internal parallelism -- to improve scheduling behavior when multiple processes are competing for cores and so internal parallelism cannot be used perfectly
In order to improve cache hit rates (see 2.), batches are created at stable boundaries, and so this value is only a "target" max batch size (rather than an exact value).
NOTE: This parameter has no effect on test runners/plugins that do not implement support for batched testing.
Deprecated options¶
None