Feat/prune unreferenced models#50
Merged
Merged
Conversation
Definition provenance was emitted inline during the schema build under the type's short leaf name, but reduceDefinitionNames renames collision groups afterwards (cross-package same-named models: Item -> XItem/YItem) without touching provenance. Every /definitions/<leaf>... pointer for a renamed definition was left dangling, so the OnProvenance consumer (the genspec TUI) could not resolve it to a spec node. This affected any scan with a cross-package name collision, independent of model pruning. OnProvenance is a push stream (RecordOrigin fires immediately and accumulates nothing), so inline anchors cannot be retracted after the rename. Buffer definition-scoped anchors instead: BeginDefOrigins/EndDefOrigins open a window around each definition build keyed by its fully-qualified DefKey; RecordOrigin buffers inside the window and fires inline outside it (path, response, info and parameter anchors are never renamed by reduction). reduceDefinitionNames now returns its old->new rename map, and FlushDefOrigins re-points each buffered anchor to the definition's final name and emits it once names are settled. Witness: name-identity-mixed and name-identity-3way join the geometry test, plus a focused TestCoverage_ProvenanceCollisionRename. Before the fix the mixed fixture emits 8 dangling anchors; after, every emitted anchor resolves against the rendered spec. Zero golden drift: provenance is a side channel and never alters the spec. Signed-off-by: Frederic BIDON <fredbi@yahoo.com> Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
ScanModels (`-m`) deliberately emits every swagger:model type, reachable or not. On a large shared library that buries the wanted models under noise; users worked around it with go-swagger's `flatten`, partly to dodge cross-package name collisions that an unused model should never have caused (go-swagger/go-swagger#2639). Add Options.PruneUnusedModels: with ScanModels, run discovery as before, then drop every discovered definition not transitively referenced from a reachability root. Roots are the paths (operation body parameters + response schemas), the shared responses and parameters, and every definition supplied via InputSpec — overlay definitions are pinned (never pruned, seeded as roots). A model referenced only by another unreferenced model is pruned too; recursive/cyclic models terminate on a visited set. Without ScanModels the flag is a no-op (the set is already reachable-only) and says so with one Hint. The prune runs BEFORE reduceDefinitionNames, in the fully-qualified definition-key namespace. That ordering is the point of the feature, not a detail: name reduction deconflicts colliding leaves (a.Thing / b.Thing -> AThing / BThing). Pruning the unused twin first means the collision never materialises, so the surviving model keeps its bare leaf name with no concat churn. Each prune raises a located scan.pruned-unused Hint; the collision renames the reduce stage still performs are surfaced as located scan.renamed-definition Hints. Buffered provenance for a pruned node is dropped, so no anchor dangles. Fixture enhancements/prune-unused exercises the reachable closure (root -> chain -> recursive node), the dead chain, and the collision pair; tests lock both regression witnesses (collision evaporation to a bare name; InputSpec pinning) plus the no-op, diagnostics and no-orphan-provenance contracts. Default false: flag-off output is byte-identical. Signed-off-by: Frederic BIDON <fredbi@yahoo.com> Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Document Options.PruneUnusedModels as a new shaping-the-output page (weight 16, between type-discovery and resolving-name-conflicts). Prose-only: presents the option, what it is useful for (scanning a shared model library with -m and keeping only the reachable subset), and contrasts the three emission modes — reachable-only (default), every model (ScanModels), and models-then-pruned (ScanModels + PruneUnusedModels). Covers what counts as reachable (path/response/parameter roots, InputSpec defs pinned), the before-name-reduction ordering that makes pruning avoid spurious cross-package collision renames, and the scan.pruned-unused / scan.renamed- definition diagnostics. Cross-links the type-discovery page (forward, from its "Standalone" note) and the resolving-name-conflicts "What's next". Markdown-lint and link-check clean. Signed-off-by: Frederic BIDON <fredbi@yahoo.com> Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Codecov Report❌ Patch coverage is Additional details and impacted files@@ Coverage Diff @@
## master #50 +/- ##
==========================================
- Coverage 81.98% 81.85% -0.14%
==========================================
Files 79 80 +1
Lines 7440 7594 +154
==========================================
+ Hits 6100 6216 +116
- Misses 994 1015 +21
- Partials 346 363 +17 ☔ View full report in Codecov by Harness. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Change type
Please select: 🆕 New feature or enhancement|🔧 Bug fix'|📃 Documentation update
Short description
Fixes
Full description
Checklist