1.8 KiB
Corrected navigation: Windows desktop test result
Tested source: 46b4f3c0d0f04050ad6d705051ee0199138e47a2 on codex/diagram-coverage-fixes-ytr83r-20260912.
cargo test -p oa-desktop --all-targets --locked --offline completed with 894 passed, 0 failed, 1 ignored, exit 0. Compilation reported 12m44s; execution reported 1137.48 seconds. Raw output, receipt, and all 562 verified compiler inputs are preserved. Each input matched its pre-check hash and byte length after the desktop run. The receipt conservatively records uncommitted work: the checkout also contained an unwired, untracked experiment directory, which was not included by this build. The compiled correction source matches the published commit.
All three failures of the older source now pass. The passing run includes the new exact qualified-preview link checks, independent Allocation grouping fixtures, port/state semantic-selection and visual-focus checks, stale chooser/deferred viewport cases, camera persistence, and existing desktop/source/history regressions.
The one ignored case is app::source_symbols::tests::source_symbols_100_modules_1000_definitions_10000_uses, explicitly a manual latency/process-memory trial intended to run alone. This result does not establish its performance threshold or large-model native responsiveness. The host also ran an independent core compilation and Capella startup during portions of execution; elapsed time here is not a comparative performance benchmark.
Core-specific outline tests and strict lint are separate gates. A matching GUI build and native acceptance of the corrected source remain pending. The process/scenario free-editing requirements are still unfinished, and Linux/WSL remains paused by the user.