Under the Hood

The commands you use every day — plcc-scan, plcc-parse, plcc-rep, and plcc-diagram — each orchestrate a pipeline of lower-level commands. This page shows how the pieces fit together.

Dependency diagram

The diagram below shows which commands call which. An arrow from A to B means A calls B.

The core pipeline

plcc-make is the build orchestrator at the centre of the core package. Every author-facing command calls it before doing its own work.

Command	Input → Output
plcc-make	.plcc spec file → build artifacts in build/
plcc-spec	.plcc file → spec JSON
plcc-ll1	spec JSON → LL(1) analysis JSON
plcc-tokens	spec JSON + source files → token JSONL
plcc-trees	token JSONL + LL(1) JSON → parse tree JSON (dispatches to parser plugin)
plcc-model	spec JSON → language-neutral model JSON
plcc-lang-emit	model JSON → language source files (dispatches to language plugin)
plcc-lang-build	language source files → compiled output (dispatches to language plugin)
plcc-lang-run	compiled output + parse tree JSON → evaluation result (dispatches to language plugin)

The diagram pipeline

plcc-diagram (in the plcc-diagram package) calls plcc-make to build the model, then runs its own sub-pipeline:

Command	Input → Output
plcc-diagram-emit	model JSON → diagram source (dispatches to diagram plugin)
plcc-diagram-build	diagram source → PNG image (dispatches to diagram plugin)
plcc-diagram-run	PNG path → prints path to stdout (dispatches to diagram plugin)