Make agent suspend tests release on the suspend signal

The single- and multi-level suspend/restore tests gated their slow
leaf tool on a manual release channel closed 50ms after cancel().
That sleep was a guess at how long the suspend signal takes to reach
the running sub-agent, so the post-tool turn boundary could observe
the release before cancellation and complete the run instead of
checkpointing, making the assertions timing-dependent.

Expose the per-run suspend signal through SuspendSignalFrom in an
export_test shim and have the leaf tools block on it directly. The
tool now returns only once the graceful-suspend signal has actually
propagated to its agent, so suspension is observed deterministically
without sleeps or release channels.

Signed-off-by: Bryan Frimin <bryan@probo.com>
This commit is contained in:
Bryan Frimin
2026-06-07 12:15:00 +02:00
parent a8d4da3916
commit 5a8654969e
3 changed files with 45 additions and 13 deletions

View File

@@ -924,16 +924,22 @@ func TestAgentTool_Execute_SuspendAndRestoreSingleLevel(t *testing.T) {
store := newMemoryCheckpointer()
toolReady := make(chan struct{})
toolRelease := make(chan struct{})
var readyOnce sync.Once
slowTool := agent.FunctionTool[struct{}](
"slow_inner_work",
"Slow inner work",
func(_ context.Context, _ struct{}) (agent.ToolResult, error) {
func(ctx context.Context, _ struct{}) (agent.ToolResult, error) {
readyOnce.Do(func() { close(toolReady) })
<-toolRelease
// Release only once the graceful-suspend signal has
// reached this inner agent, so the post-tool turn
// boundary deterministically observes cancellation and
// checkpoints instead of completing the run.
if sig := agent.SuspendSignalFrom(ctx); sig != nil {
<-sig.Done()
}
return agent.ToolResult{Content: "inner tool done"}, nil
},
@@ -996,8 +1002,6 @@ func TestAgentTool_Execute_SuspendAndRestoreSingleLevel(t *testing.T) {
}
cancel()
time.Sleep(50 * time.Millisecond)
close(toolRelease)
select {
case err := <-errCh:
@@ -1043,16 +1047,22 @@ func TestAgentTool_Execute_SuspendAndRestoreMultiLevel(t *testing.T) {
store := newMemoryCheckpointer()
toolReady := make(chan struct{})
toolRelease := make(chan struct{})
var readyOnce sync.Once
slowTool := agent.FunctionTool[struct{}](
"slow_grandchild_work",
"Slow grandchild work",
func(_ context.Context, _ struct{}) (agent.ToolResult, error) {
func(ctx context.Context, _ struct{}) (agent.ToolResult, error) {
readyOnce.Do(func() { close(toolReady) })
<-toolRelease
// Release only once the graceful-suspend signal has
// propagated down to this grandchild agent, so the
// post-tool turn boundary deterministically observes
// cancellation and checkpoints instead of completing.
if sig := agent.SuspendSignalFrom(ctx); sig != nil {
<-sig.Done()
}
return agent.ToolResult{Content: "grandchild tool done"}, nil
},
@@ -1134,8 +1144,6 @@ func TestAgentTool_Execute_SuspendAndRestoreMultiLevel(t *testing.T) {
}
cancel()
time.Sleep(50 * time.Millisecond)
close(toolRelease)
select {
case err := <-errCh: