Repository navigation
TESTS_FILE edges are never created for JVM (Kotlin/JUnit) tests #1556
Description
Environment: v0.10.2, Windows 11, --mode full. Kotlin (JUnit5/Testcontainers) + TypeScript (vitest) monorepo.
Symptom
TESTS_FILE edges are created only for the TypeScript side. Measured on a full index:
- Files matching
.*(Test|IT)\.ktor.*\.test\.tsx?: 283 TESTS_FILEedges: 92 — all of them.test.tsx(66) /.test.ts(26)- JVM tests linked: 0, including trivially-conventional pairs like
AgendarUseCaseTest.kt→AgendarUseCase.kt(same package, siblingsrc/test/kotlinsource set) andJooqCompromissoRepositoryIT.kt→JooqCompromissoRepository.kt.
Impact
Any "which tests cover this class?" query over the graph silently answers "none" for the whole JVM half of a polyglot repo. We now instruct agents to fall back to naming convention + OVERRIDE edges of test fakes for Kotlin, but the edge type advertises coverage it doesn't have.
Suggestion
Apply the same filename/convention heuristic used for *.test.ts(x) to JVM conventions (XTest/XIT, src/test/{kotlin,java} mirroring src/main/...), or document TESTS_FILE as TS-only.
Activity
DeusData commentedon Aug 12, 2026
Thank you @PsyTech-Rob. Confirmed in
pass_tests.c:XTest.ktis recognized as a test path, buttest_to_prod_path()only derives production paths for Go, Python, and JavaScript/TypeScript naming conventions. It has no Kotlin/Java source-set mirror and noXIT.ktmapping, so JVMTESTS_FILEedges cannot be created by the current pass.Routed priority/high under the parsing and extraction epic #592. The measured TS/JVM split is a useful regression fixture.
DeusData commentedon Oct 19, 2026
Thank you again for the measured TS/JVM difference and the concrete Kotlin naming pairs. I have no new fix to report for JVM TESTS_FILE mapping; no additional reproduction is requested at this stage.
DeusData commentedon Oct 25, 2026
Thank you for the concrete JVM test/source pairs. The TESTS_FILE mapping gap remains for maintainer follow-through with no additional reproduction needed. Sorry for the delay; this is waiting on maintainer follow-up, and I’ll follow up soon.