Add end-to-end CLI workflows and user documentation #18

New issue

Open

opened 2026-09-03 14:22:18 +00:00 by JayKobe6k · 0 comments

JayKobe6k commented

2026-09-03 14:22:18 +00:00

Owner

Goal

Prove and document the complete CLI and core workflow through supported frontend boundaries.

End-to-end scenarios

Index a local multi-TU C project into the standard cache.
Index a local Git-origin fixture at a fixed revision.
Resolve the same index by project name, local directory, Git origin, and explicit path where supported.
Search values, functions, structs, and enums with text and JSON output.
Feed a returned value handle into upstream/downstream lineage traversal.
Modify a source file and update incrementally.
Modify a shared header and verify all affected TUs update.
Demonstrate a no-op update and a forced --reindex.
Simulate parse/persistence failure and prove the prior active revision remains queryable.

Documentation

Document installation/runtime dependencies, cache locations on Linux and Windows, project/build/compilation-database inputs, Git acquisition, index naming/selectors, update semantics, value handles, regex behavior, grouping/ordering JSON shape, lineage direction/depth, diagnostics, exit statuses, schema compatibility, and current analysis limitations. Include copy-pastable text and JSON examples.

Acceptance criteria

Tests are deterministic, isolated, network-free, and use unique temporary databases/cache roots.
Tests exercise the real CLI process and LadybugDB adapter, not only mocks.
JSON examples validate against the actual serialized response types.
Documentation contains no raw Cypher and does not imply references/calls alone prove value flow.
Linux CI runs the complete suite and the pinned Nix environment retains the Windows cross-compilation check.
README implementation-status claims are updated to match delivered behavior.

Dependencies

Depends on #8, #9, #12, #14, #15, #16, and #17.

## Goal Prove and document the complete CLI and core workflow through supported frontend boundaries. ## End-to-end scenarios - Index a local multi-TU C project into the standard cache. - Index a local Git-origin fixture at a fixed revision. - Resolve the same index by project name, local directory, Git origin, and explicit path where supported. - Search values, functions, structs, and enums with text and JSON output. - Feed a returned value handle into upstream/downstream lineage traversal. - Modify a source file and update incrementally. - Modify a shared header and verify all affected TUs update. - Demonstrate a no-op update and a forced `--reindex`. - Simulate parse/persistence failure and prove the prior active revision remains queryable. ## Documentation Document installation/runtime dependencies, cache locations on Linux and Windows, project/build/compilation-database inputs, Git acquisition, index naming/selectors, update semantics, value handles, regex behavior, grouping/ordering JSON shape, lineage direction/depth, diagnostics, exit statuses, schema compatibility, and current analysis limitations. Include copy-pastable text and JSON examples. ## Acceptance criteria - Tests are deterministic, isolated, network-free, and use unique temporary databases/cache roots. - Tests exercise the real CLI process and LadybugDB adapter, not only mocks. - JSON examples validate against the actual serialized response types. - Documentation contains no raw Cypher and does not imply references/calls alone prove value flow. - Linux CI runs the complete suite and the pinned Nix environment retains the Windows cross-compilation check. - README implementation-status claims are updated to match delivered behavior. ## Dependencies Depends on #8, #9, #12, #14, #15, #16, and #17.

JayKobe6k added this to the CLI + Core MVP milestone

2026-09-03 14:22:18 +00:00

JayKobe6k added the

labels

2026-09-03 14:22:18 +00:00

JayKobe6k referenced this issue

2026-09-06 19:21:31 +00:00