Octri

lab run

Runs the whole build loop in one go: validates the spec, builds the languages you give, draws a lane per language while they build, then downloads and unpacks every package that shipped. It ends with a verdict per language, with the failing output inline for any that failed.

Runs are kept on this machine under ~/.octri/cache/lab/<project>/<buildId>, with a manifest, which is what the other lab commands read. It exits with an error when any language fails, so it fits a shell loop or a CI step.

bash
octri lab run [--lang <languages> | --all] [--keep-going] [--skip-validate]

Flags

--lang <languages>listdefault: the profile's defaultLanguages

Comma-separated language IDs.

--allboolean

Every language in the catalogue.

--keep-goingboolean

Builds even when the spec has validation errors. Without it, the run stops before building.

--skip-validateboolean

Skips validation altogether.

--version <version>string

The version the packages are built with.

--interval <ms>numberdefault: 2000

How often to check on the build.

Examples

Build and pull two languages

bash
octri lab run --lang typescript,python
  • octri lab diff: Compare two kept runs file by file, per language.
  • octri lab files: Show the file tree one language generated in a kept run.
  • octri sdk build: Build SDKs for the languages you choose and follow each one live.