Flush queued IDE writes and restart validation after host revision changes
This commit is contained in:
@@ -58,3 +58,7 @@ Evidence logs are `/tmp/generated-vm-bootstrap.log`, `/tmp/generated-vm-processo
|
||||
The first registered IntelliJ run passed the classpath workflow but stopped in module-path mode at the previously recorded immediate exact-save assertion, before generation checks. An unchanged module-path retry timed out at initial canvas rendering (`WorkflowProbe` line 438), also before generation checks. No assertion was weakened. Existing startup/persistence/preview-gate reliability work remains open; this slice does not claim to fix those intermittent failures.
|
||||
|
||||
The final unchanged module-path retry hit the same initial-render timeout. Consequently, **registered IntelliJ module-path whole-host qualification remains open** for this slice. Registered IntelliJ classpath passed generated-registry discovery, full shared editing/history/diagnostic checks, exclusive JavaFX 26 native verification and worker shutdown. Standalone and real target-process module-path checks passed, including generated classes and registry behavior; these do not substitute for the missing registered IDE pass. Follow up on initial rendering and save synchronization before declaring that host gate complete.
|
||||
|
||||
## Save/startup gate follow-up
|
||||
|
||||
The previously open module-path host gate is closed by the [IntelliJ save/startup reliability follow-up](ide-host-reliability.md): deterministic regressions reproduced and corrected queued VFS persistence and stale validation state, followed by three consecutive passing registered module-path workflows with the generated registry checks included. The failed attempts above remain historical evidence.
|
||||
|
||||
Reference in New Issue
Block a user