Compare commits
1
Commits
| Author | SHA1 | Date | |
|---|---|---|---|
|
|
c91fade6b9 |
No files matched your search
@@ -26,6 +26,14 @@ FrameLedger uses the **`poli0981/.github` ops repo** where its templates fit, an
|
||||
wrong by the end of the same evening; it is corrected rather than quietly deleted, because a frequency
|
||||
claim made from one observation is exactly the shape this file keeps catching.
|
||||
**Re-run before investigating**; if it reproduces, read the module list and the scan-set reason, not the test.
|
||||
- **Two more shapes had one cause, found and fixed 2026-09-16** after four failures in one evening's merge train:
|
||||
`AKilledHostLeavesAPartialThatRecoverTurnsIntoAnInterruptedSession` ("the process cannot access the file …
|
||||
`.partial`") and `WithNoConsentRecordTheHostRefusesAndNothingIsEverInjected` (a leftover consent record). The
|
||||
suite's `Kill` helper returned before the killed host had released its files, so the next line read a `.partial`
|
||||
the dead process still held, and the ledger delete in `Dispose` silently failed and left the record for the next
|
||||
case. `Kill` now waits for exit (bounded, asserted) and the delete retries. Not §S19(b) either — a test-harness
|
||||
race, and the third flake shape of the day (the registry-subtree race between `VkLayerRegistrationTests` and the
|
||||
GOG reader's test, fixed in #188) was the same kind.
|
||||
- **A re-run ERASES the evidence, which is why the count above had to be kept by hand.**
|
||||
`gh run rerun --failed` updates the original run's conclusion, so `gh run list` showed **18 success, 1
|
||||
cancelled, 0 failed** across the twenty runs that contained all three failures above. Anyone measuring
|
||||
|
||||
@@ -89,16 +89,26 @@ public sealed class CaptureHostEndToEndTests : IDisposable
|
||||
}
|
||||
|
||||
/// <summary>The ledger and WAL's two sidecars; a guarded delete of each.</summary>
|
||||
/// <summary>
|
||||
/// Deletes the host's ledger, retrying for a few seconds: a host this suite just killed may still hold the file
|
||||
/// while Windows tears the process down, and a delete that silently failed is how the refusal case found a
|
||||
/// leftover consent record on CI (2026-09-16, three runs in one evening).
|
||||
/// </summary>
|
||||
private static void DeleteHostLedger()
|
||||
{
|
||||
foreach (string suffix in new[] { "", "-wal", "-shm" })
|
||||
{
|
||||
try
|
||||
{
|
||||
File.Delete(HostLedger + suffix);
|
||||
}
|
||||
catch (Exception ex) when (ex is IOException or UnauthorizedAccessException)
|
||||
string path = HostLedger + suffix;
|
||||
for (int attempt = 0; attempt < 20 && File.Exists(path); attempt++)
|
||||
{
|
||||
try
|
||||
{
|
||||
File.Delete(path);
|
||||
}
|
||||
catch (Exception ex) when (ex is IOException or UnauthorizedAccessException)
|
||||
{
|
||||
Thread.Sleep(250);
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
@@ -158,6 +168,12 @@ public sealed class CaptureHostEndToEndTests : IDisposable
|
||||
return Process.Start(psi)!;
|
||||
}
|
||||
|
||||
/// <summary>
|
||||
/// Kills the process tree AND waits for it to be gone. `Kill` returns before the process has released its
|
||||
/// files; the kill-and-recover case then opened the `.partial` the dead host still held ("being used by another
|
||||
/// process"), and the refusal case's ledger delete failed the same way — the two CaptureHost flakes CI kept
|
||||
/// producing (13_CI_CD's tally). A wait bounded at 15 s: a process that will not die is a failure to report.
|
||||
/// </summary>
|
||||
private static void Kill(Process p)
|
||||
{
|
||||
try
|
||||
@@ -165,6 +181,7 @@ public sealed class CaptureHostEndToEndTests : IDisposable
|
||||
if (!p.HasExited)
|
||||
{
|
||||
p.Kill(entireProcessTree: true);
|
||||
p.WaitForExit(15_000).Should().BeTrue("a killed process must be gone before its files are read or deleted");
|
||||
}
|
||||
}
|
||||
catch (InvalidOperationException)
|
||||
|
||||
Reference in new issue
Block a user