App::Y...::start(3pm)

📦
version 1.000172 · 8 versions
, in ubuntu stonking

🌸 NAME 🌸

App::Yath::Command::start — Start the persistent test runner.

This manual page is App::Y...::start(3pm) from User Contributed Perl Documentation, bestie. ✨

📝 DESCRIPTION 📝

This command starts a persistant instance of yath. A persistant instance is useful because it lets you preload modules in advance, which reduces start time for any tests you decide to run as you work. 💖

A running instance will watch for changes to any preloaded files, and restart itself if anything changes. Changed files are blacklisted for subsequent reloads, so reloading is not a frequent occurence when you edit the same file over and over again. 🦋

💻 USAGE 💻

$ yath [YATH OPTIONS] start [COMMAND OPTIONS]

✨ YATH OPTIONS ✨

🛠️ Developer 🛠️

  • --dev-lib, --dev-lib=lib, -D, -D=lib, -Dlib, --no-dev-lib — Add paths to @INC before loading ANYTHING. This is what you use if you are developing yath or yath plugins to make sure the yath script finds the local code instead of the installed versions of the same code. You can provide an argument (-Dfoo) to provide a custom path, or you can just use -D without an arg to add lib, blib/lib and blib/arch. Can be specified multiple times.

🌱 Environment 🌱

  • --persist-dir ARG, --persist-dir=ARG, --no-persist-dir — Where to find persistence files.
  • --persist-file ARG, --persist-file=ARG, --pfile ARG, --pfile=ARG, --no-persist-file — Where to find the persistence file. The default is /{system-tempdir}/project-yath-persist.json. If no project is specified then it will fall back to the current directory. If the current directory is not writable it will default to /tmp/yath-persist.json, which limits you to one persistent runner on your system.
  • --project ARG, --project=ARG, --project-name ARG, --project-name=ARG, --no-project — This lets you provide a label for your current project/codebase. This is best used in a .yath.rc file. This is necessary for a persistent runner.

🆘 Help and Debugging 🆘

  • --show-opts, --no-show-opts — Exit after showing what yath thinks your options mean.
  • --version, -V, --no-version — Exit after showing a helpful usage message.

🧩 Plugins 🧩

  • --no-scan-plugins, --no-no-scan-plugins — Normally yath scans for and loads all App::Yath::Plugin::* modules in order to bring in command-line options they may provide. This flag will disable that. This is useful if you have a naughty plugin that is loading other modules when it should not.
  • --plugins PLUGIN, --plugins +App::Yath::Plugin::PLUGIN, --plugins PLUGIN=arg1,arg2,..., --plugin PLUGIN, --plugin +App::Yath::Plugin::PLUGIN, --plugin PLUGIN=arg1,arg2,..., -pPLUGIN, --no-plugins — Load a yath plugin. Can be specified multiple times.

👑 COMMAND OPTIONS 👑

📊 Collector Options 📊

  • --max-open-jobs 18, --no-max-open-jobs — Maximum number of jobs a collector can process at a time. If more jobs are pending, their output will be delayed until the earlier jobs have been processed. (Default: double the -j value.)
  • --max-poll-events 1000, --no-max-poll-events — Maximum number of events to poll from a job before jumping to the next job. (Default: 1000.)

🎀 Cover Options 🎀

  • --cover-aggregator ByTest, --cover-aggregator ByRun, --cover-aggregator +Custom::Aggregator, --cover-agg ByTest, --cover-agg ByRun, --cover-agg +Custom::Aggregator, --no-cover-aggregator — Choose a custom aggregator subclass.
  • --cover-class ARG, --cover-class=ARG, --no-cover-class — Choose a Test2::Plugin::Cover subclass.
  • --cover-dirs ARG, --cover-dirs=ARG, --cover-dir ARG, --cover-dir=ARG, --no-cover-dirs — NO DESCRIPTION - FIX ME 😭 the original is literally asking for help. Can be specified multiple times.
  • --cover-exclude-private, --no-cover-exclude-private, --cover-files, --no-cover-files — Use Test2::Plugin::Cover to collect coverage data for what files are touched by what tests. Unlike Devel::Cover this has very little performance impact (About 4% difference).
  • --cover-from path/to/log.jsonl, --cover-from http://example.com/coverage, --cover-from path/to/coverage.jsonl, --no-cover-from — This can be a test log, a coverage dump (old style json or new jsonl format), or a url to any of the previous. Tests will not be run if the file/url is invalid.
  • --cover-from-type json, --cover-from-type jsonl, --cover-from-type log, --no-cover-from-type — File type for coverage source. Usually it can be detected, but when it cannot you should specify. "json" is old style single-blob coverage data, "jsonl" is the new by-test style, "log" is a logfile from a previous run.
  • --cover-manager My::Coverage::Manager, --no-cover-manager — Coverage 'from' manager to use when coverage data does not provide one.
  • --cover-maybe-from path/to/log.jsonl, --cover-maybe-from http://example.com/coverage, --cover-maybe-from path/to/coverage.jsonl, --no-cover-maybe-from — This can be a test log, a coverage dump (old style json or new jsonl format), or a url to any of the previous. Tests will continue even if the coverage file/url is invalid.
  • --cover-maybe-from-type json, --cover-maybe-from-type jsonl, --cover-maybe-from-type log, --no-cover-maybe-from-type — Same as "from_type" but for "maybe_from". Defaults to "from_type" if that is specified, otherwise auto-detect.
  • --cover-metrics, --no-cover-metrics — No description in the original. 👻
  • --cover-types ARG, --cover-types=ARG, --cover-type ARG, --cover-type=ARG, --no-cover-types — NO DESCRIPTION - FIX ME again, the source is leaving us on read. Can be specified multiple times.
  • --cover-write, --cover-write=coverage.jsonl, --cover-write=coverage.json, --no-cover-write — Create a json or jsonl file of all coverage data seen during the run (This implies --cover-files).

🐙 Git Options 🐙

  • --git-change-base master, --git-change-base HEAD^, --git-change-base df22abe4, --no-git-change-base — Find files changed by all commits in the current branch from most recent stopping when a commit is found that is also present in the history of the branch/commit specified as the change base.

🆘 Help and Debugging 🆘

  • --dummy, -d, --no-dummy — Dummy run, do not actually execute anything. Can also be set with the following environment variables: "T2_HARNESS_DUMMY".
  • --help, -h, --no-help — exit after showing help information.
  • --interactive, -i, --no-interactive — Use interactive mode, 1 test at a time, stdin forwarded to it.
  • --keep-dirs, --keep_dir, -k, --no-keep-dirs — Do not delete directories when done. This is useful if you want to inspect the directories used for various commands.
  • --procname-prefix ARG, --procname-prefix=ARG, --no-procname-prefix — Add a prefix to all proc names (as seen by ps).

💾 Persistent Runner Options 💾

  • --quiet, -q, --no-quiet — Be very quiet. Can be specified multiple times.
  • --reload, -r, --no-reload — Attempt to reload modified modules in-place, restarting entire stages only when necessary.
  • --restrict-reload, --restrict-reload=path, --no-restrict-reload — Only reload modules under the specified path, if no path is specified look at anything under the .yath.rc path, or the current working directory. Can be specified multiple times.

🏃 Runner Options 🏃

  • --abort-on-bail, --no-abort-on-bail — Abort all testing if a bail-out is encountered (default: on).
  • --blib, -b, --no-blib — (Default: include if it exists) Include 'blib/lib' and 'blib/arch' in your module path.
  • --cover, --cover=-silent,1,+ignore,^t/,+ignore,^t2/,+ignore,^xt,+ignore,^test.pl, --no-cover — Use Devel::Cover to calculate test coverage. This disables forking. If no args are specified the following are used: -silent,1,+ignore,^t/,+ignore,^t2/,+ignore,^xt,+ignore,^test.pl.
  • --daemon, --no-daemon — Start the runner as a daemon (Default: True).
  • --dump-depmap, --no-dump-depmap — When using staged preload, dump the depmap for each stage as json files.
  • --event-timeout SECONDS, --et SECONDS, --no-event-timeout — Kill test if no output is received within timeout period. (Default: 60 seconds). Add the # HARNESS-NO-TIMEOUT comment to the top of a test file to disable timeouts on a per-test basis. This prevents a hung test from running forever.
  • --fail-on-resource-skip, --no-fail-on-resource-skip — Treat resource-skipped tests as failures instead of skips. When enabled, tests that would be skipped due to unavailable resources will be marked as failing.
  • --include ARG, --include=ARG, -I ARG, -I=ARG, --no-include — Add a directory to your include paths. Can be specified multiple times.
  • --job-count 4, --job-count 8:2, --jobs 4, --jobs 8:2, -j4, -j8:2, --no-job-count — Set the number of concurrent jobs to run. Add a :# if you also wish to designate multiple slots per test. 8:2 means 8 slots, but each test gets 2 slots, so 4 tests run concurrently. Tests can find their concurrency assignment in the "T2_HARNESS_MY_JOB_CONCURRENCY" environment variable. Can also be set with the following environment variables: "YATH_JOB_COUNT", "T2_HARNESS_JOB_COUNT", "HARNESS_JOB_COUNT".
  • --lib, -l, --no-lib — (Default: include if it exists) Include 'lib' in your module path.
  • --nytprof, --no-nytprof — Use Devel::NYTProf on tests. This will set addpid=1 for you. This works with or without fork.
  • --post-exit-timeout SECONDS, --pet SECONDS, --no-post-exit-timeout — Stop waiting post-exit after the timeout period. (Default: 15 seconds) Some tests fork and allow the parent to exit before writing all their output. If Test2::Harness detects an incomplete plan after the test exits it will monitor for more events until the timeout period. Add the # HARNESS-NO-TIMEOUT comment to the top of a test file to disable timeouts on a per-test basis.
  • --preload-threshold ARG, --preload-threshold=ARG, --Pt ARG, --Pt=ARG, -W ARG, -W=ARG, --no-preload-threshold — okay so this one only preloads if at least N tests are actually gonna run 💅. Sometimes a full preload takes longer than just running the tests, so this lets you set a minimum number of test jobs that need to be happening before preload will even bother. Literally no effect for a persistent runner, bestie. The default is 0, which means always preload. Periodt.
  • --preloads ARG, --preloads=ARG, --preload ARG, --preload=ARG, -P ARG, -P=ARG, --no-preloads — preload a module before running tests ✨. Can be specified multiple times, so stack 'em up like your fave scented candles 🕯️.
  • --resource-timeout SECONDS, --rt SECONDS, --no-resource-timeout — abort the whole test run if no tests have been able to start for SECONDS seconds while there are pending tests and none running 😭. This is the moment when a resource class is broken and keeps claiming a resource will become available, which means yath would never ever finish. Iconic safety net. (Default: 0, meaning no timeout.)
  • --resource Port, --resource +Test2::Harness::Runner::Resource::Port, -R Port, --no-resource — use a resource module to assign resource assignments to individual tests 👑. Can be specified multiple times.
  • --runner-id ARG, --runner-id=ARG, --no-runner-id — the runner ID (usually a generated uuid) 🦋.
  • --shared-jobs-config .sharedjobslots.yml, --shared-jobs-config relative/path/.sharedjobslots.yml, --shared-jobs-config /absolute/path/.sharedjobslots.yml, --no-shared-jobs-config — where to look for a shared slot config file 🔍. If you hand it a filename with no path, yath will search the current directory and every parent directory for that name.
  • --slots-per-job 2, -x2, --no-slots-per-job — sets the number of slots each job will use (default 1) 🌸. This is normally set by the :# in -j#:#.

Can also be set with the following environment variables: T2_HARNESS_JOB_CONCURRENCY

  • --switch ARG, --switch=ARG, -S ARG, -S=ARG, --no-switch — pass the specified switch to perl for each test 🎀. Not compatible with preload, so pick your fighter. Can be specified multiple times.
  • --tlib, --no-tlib — (Default: off) include t/lib in your module path 🫶.
  • --unsafe-inc, --no-unsafe-inc — perl is removing . from @INC as a security concern 🔥. This option keeps things from breaking for now.

Can also be set with the following environment variables: PERL_USE_UNSAFE_INC

  • --use-fork, --fork, --no-use-fork — (default: on, except on windows) normally tests are run by forking, which is what makes features like preloading possible 💖. This turns that behavior off globally (and that is not compatible with preloading). It's slower, so honestly it's better to tag misbehaving tests with the # HARNESS-NO-PRELOAD comment in their header to disable forking only for those tests. Main character solution, fr fr.

Can also be set with the following environment variables: !T2_NO_FORK, T2_HARNESS_FORK, !T2_HARNESS_NO_FORK, YATH_FORK, !YATH_NO_FORK

  • --use-timeout, --timeout, --no-use-timeout — (default: on) enable/disable timeouts ⏱️.

🧹 Workspace Options 🧹

  • --clear, -C, --no-clear — clear the work directory if it is not already empty 🧼.
  • --tmp-dir ARG, --tmp-dir=ARG, --tmpdir ARG, --tmpdir=ARG, -t ARG, -t=ARG, --no-tmp-dir — use a specific temp directory (Default: use system temp dir) 🍓.

Can also be set with the following environment variables: T2_HARNESS_TEMP_DIR, YATH_TEMP_DIR, TMPDIR, TEMPDIR, TMP_DIR, TEMP_DIR

  • --workdir ARG, --workdir=ARG, -w ARG, -w=ARG, --no-workdir — set the work directory (Default: new temp directory) 🏠.

Can also be set with the following environment variables: T2_WORKDIR, YATH_WORKDIR

🌈 YathUI Options 🌈

  • --yathui-api-key ARG, --yathui-api-key=ARG, --no-yathui-api-key — your Yath-UI API key 🔑. Not necessary if your Yath-UI instance is set to single-user.
  • --yathui-grace, --no-yathui-grace — if yath cannot connect to yath-ui it normally throws an error 😱, so use this to make it fail gracefully. You get a warning, but things keep going. Soft launch energy.
  • --yathui-long-duration 10, --no-yathui-long-duration — minimum duration length (seconds) before a test goes from MEDIUM to LONG 📏.
  • --yathui-medium-duration 5, --no-yathui-medium-duration — minimum duration length (seconds) before a test goes from SHORT to MEDIUM 📐.
  • --yathui-mode summary, --yathui-mode qvf, --yathui-mode qvfd, --yathui-mode complete, --no-yathui-mode — set the upload mode (default 'qvfd') 💅.
  • --yathui-project ARG, --yathui-project=ARG, --no-yathui-project — the Yath-UI project for your test results 🎀.
  • --yathui-retry, --no-yathui-retry — how many times to try an operation before giving up 🙈. Can be specified multiple times.
  • --yathui-url http://my-yath-ui.com/..., --uri http://my-yath-ui.com/..., --no-yathui-url — Yath-UI url 🌐.

📦 SOURCE 📦

The source code repository for Test2-Harness can be found at http://github.com/Test-More/Test2-Harness/ ⭐.

🫡 MAINTAINERS 🫡

Chad Granum <exodist@cpan.org>

👑 AUTHORS 👑

Chad Granum <exodist@cpan.org>

Copyright 2026 Chad Granum <exodist7@gmail.com>.

This program is free software; you can redistribute it and/or modify it under the same terms as Perl itself 🌸.

See http://dev.perl.org/licenses/

And that's the whole page, bestie, we ate 🍽️✨ no crumbs left. This lil masterpiece is from perl v5.40.1, dated 2026-05-10, and it's the App::Y...::start(3pm) manpage — go forth and run those tests, you gorgeous genius 🦋💜

📦 where this page comes from