Merge pull request 'Phase 5 (3/4): every test on a bus of its own, at the release the mesh runs (hq ADR 0237)' (#97) from feat/a-suite-that-cannot-flake into main

internal/link/advisories_test.go

@@ -8,6 +8,8 @@ import (
 	"time"
 
 	"github.com/nats-io/nats.go"
+
+	"github.com/novox/mesh-controller/internal/testbus"
 )
 
 // **A reader's own consumer gone is not a consumer lost**: every bucket watch and stream read-back
@@ -64,10 +66,7 @@ func TestNatsARefusedQuestionIsSaidAtOnce(t *testing.T) {
 
 // The refusal reaches whoever waits on that subject, and only them.
 func TestNatsARefusalReachesItsWaiter(t *testing.T) {
-	url := os.Getenv("MESH_TEST_NATS")
-	if url == "" {
-		t.Skip("MESH_TEST_NATS unset")
-	}
+	url := testbus.URL(t)
 	conn, err := nats.Connect(url)
 	if err != nil {
 		t.Fatal(err)

internal/link/builds_nats_test.go

@@ -5,7 +5,6 @@ import (
 	"encoding/json"
 	"io"
 	"log"
-	"os"
 	"sync"
 	"testing"
 	"time"
@@ -13,6 +12,8 @@ import (
 	"github.com/nats-io/nats.go"
 
 	"github.com/novox/mesh-controller/internal/broker"
+
+	"github.com/novox/mesh-controller/internal/testbus"
 )
 
 // A build, end to end, against a real server.
@@ -22,15 +23,11 @@ import (
 // catalogue places it. On the old bus that fan-out came from a shared exchange, and it would be easy
 // to write a version where only the asker hears it and nobody notices for weeks.
 //
-//	docker run -d --rm --name t -p 14230:4222 nats:2.10-alpine -js
-//	MESH_TEST_NATS=nats://127.0.0.1:14230 go test ./internal/link/ -run TestNatsABuild
+//	go test ./internal/link/ -run TestNatsABuild   (each test on a bus of its own: internal/testbus)
 
 func aBusWithTheBuildRole(t *testing.T) *broker.JetStream {
 	t.Helper()
-	url := os.Getenv("MESH_TEST_NATS")
-	if url == "" {
-		t.Skip("MESH_TEST_NATS unset")
-	}
+	url := testbus.URL(t)
 	js, err := broker.Dial(url)
 	if err != nil {
 		t.Fatal(err)
@@ -349,7 +346,7 @@ func TestNatsAMachineOnTheRetiredBuildRoleTakesThatRolesAsks(t *testing.T) {
 		})
 	}()
 
-	asker, err := BuildsOverNATSOn(os.Getenv("MESH_TEST_NATS"), TheBuildMachineBefore)
+	asker, err := BuildsOverNATSOn(testbus.URL(t), TheBuildMachineBefore)
 	if err != nil {
 		t.Fatal(err)
 	}

internal/link/bus_test.go

@@ -3,12 +3,13 @@ package link
 import (
 	"context"
 	"encoding/json"
-	"os"
 	"strings"
 	"testing"
 	"time"
 
 	"github.com/nats-io/nats.go"
+
+	"github.com/novox/mesh-controller/internal/testbus"
 )
 
 // **Both implementations, one fixture.** The point of the seam is not that the transport can be
@@ -17,13 +18,9 @@ import (
 //
 // Against a real server, because what the fixture pins is what reaches the wire:
 //
-//	docker run -d --rm --name t -p 14222:4222 nats:2.10-alpine -js
-//	MESH_TEST_NATS=nats://127.0.0.1:14222 go test ./internal/link/ -run TestTheNatsBus
+//	go test ./internal/link/ -run TestTheNatsBus   (each test on a bus of its own: internal/testbus)
 func TestTheNatsBusEmitsTheEnvelopeTheFixturePins(t *testing.T) {
-	url := os.Getenv("MESH_TEST_NATS")
-	if url == "" {
-		t.Skip("MESH_TEST_NATS unset")
-	}
+	url := testbus.URL(t)
 	f := loadFixture(t, "events/module-event.json")
 
 	conn, err := nats.Connect(url)
@@ -99,10 +96,7 @@ func TestAnEventsSubjectIsDerivedFromItsSource(t *testing.T) {
 // tool is slow" is the first thing a person asking wants, and waiting out the timeout to say it
 // is how a fast answer becomes a slow non-answer.
 func TestAskingAToolNobodyServesFailsAtOnce(t *testing.T) {
-	url := os.Getenv("MESH_TEST_NATS")
-	if url == "" {
-		t.Skip("MESH_TEST_NATS unset")
-	}
+	url := testbus.URL(t)
 	conn, err := nats.Connect(url)
 	if err != nil {
 		t.Fatal(err)
@@ -125,10 +119,7 @@ func TestAskingAToolNobodyServesFailsAtOnce(t *testing.T) {
 
 // And a served tool answers, with no reply queue to declare and no correlation to check.
 func TestAskingAServedToolAnswers(t *testing.T) {
-	url := os.Getenv("MESH_TEST_NATS")
-	if url == "" {
-		t.Skip("MESH_TEST_NATS unset")
-	}
+	url := testbus.URL(t)
 	conn, err := nats.Connect(url)
 	if err != nil {
 		t.Fatal(err)

internal/link/enrol_reply_probe_test.go

@@ -1,11 +1,12 @@
 package link
 
 import (
-	"os"
 	"testing"
 	"time"
 
 	"github.com/nats-io/nats.go"
+
+	"github.com/novox/mesh-controller/internal/testbus"
 )
 
 // **Verified 2026-09-27**: the caller asked for a reply to `_INBOX.LCr3M83q…` and the consumer
@@ -18,13 +19,9 @@ import (
 // enrolment design rests on it — the reply subject travels in the payload instead — so it is
 // checked rather than believed.
 //
-//	docker run -d --rm --name t -p 14222:4222 nats:2.10-alpine -js
-//	MESH_TEST_NATS=nats://127.0.0.1:14222 go test ./internal/link/ -run TestAReplyAddress
+//	go test ./internal/link/ -run TestAReplyAddress   (each test on a bus of its own: internal/testbus)
 func TestAReplyAddressDoesNotSurviveAStream(t *testing.T) {
-	url := os.Getenv("MESH_TEST_NATS")
-	if url == "" {
-		t.Skip("MESH_TEST_NATS unset")
-	}
+	url := testbus.URL(t)
 	conn, err := nats.Connect(url)
 	if err != nil {
 		t.Fatal(err)

internal/link/receive_nats_test.go

@@ -6,7 +6,6 @@ import (
 	"errors"
 	"io"
 	"log"
-	"os"
 	"sync"
 	"testing"
 	"time"
@@ -14,6 +13,8 @@ import (
 	"github.com/nats-io/nats.go"
 
 	"github.com/novox/mesh-controller/internal/broker"
+
+	"github.com/novox/mesh-controller/internal/testbus"
 )
 
 // The consume side against a real server, because what is being checked is what the server does.
@@ -23,15 +24,11 @@ import (
 // to an address carried in the payload reaches a caller waiting on its own inbox. Each is a claim
 // about a server, so each is asked of one:
 //
-//	docker run -d --rm --name t -p 14222:4222 nats:2.10-alpine -js
-//	MESH_TEST_NATS=nats://127.0.0.1:14222 go test ./internal/link/ -run TestNats
+//	go test ./internal/link/ -run TestNats   (each test on a bus of its own: internal/testbus)
 
 func aBus(t *testing.T) *broker.JetStream {
 	t.Helper()
-	url := os.Getenv("MESH_TEST_NATS")
-	if url == "" {
-		t.Skip("MESH_TEST_NATS unset")
-	}
+	url := testbus.URL(t)
 	js, err := broker.Dial(url)
 	if err != nil {
 		t.Fatal(err)

internal/link/standby_nats_test.go

@@ -3,11 +3,12 @@ package link
 import (
 	"context"
 	"encoding/json"
-	"os"
 	"testing"
 	"time"
 
 	"github.com/novox/mesh-controller/internal/broker"
+
+	"github.com/novox/mesh-controller/internal/testbus"
 )
 
 // novox/hq issue 213: while a machine hands its controller over, the new controller (the process)
@@ -29,7 +30,7 @@ func TestNatsASecondControllerStandsByAndTakesOverWhenTheFirstLetsGo(t *testing.
 	})
 
 	// The new one, on a connection of its own as the process would have.
-	second, err := broker.Dial(os.Getenv("MESH_TEST_NATS"))
+	second, err := broker.Dial(testbus.URL(t))
 	if err != nil {
 		t.Fatal(err)
 	}