mesh-sdk/go/delivery/buildrun/standing_test.go

package buildrun

import (
	"bytes"
	"context"
	"errors"
	"fmt"
	"log"
	"strings"
	"testing"
	"time"

	"github.com/nats-io/nats.go/jetstream"

	"git.novox.be/novox/mesh-sdk/go/delivery"
)

func gateReport(t *testing.T, node, version string, at time.Time) (string, []byte) {
	t.Helper()
	key, err := delivery.ToolchainKey(node, "typescript")
	if err != nil {
		t.Fatal(err)
	}
	value, err := delivery.EncodeToolchain(delivery.Toolchain{Language: "typescript", Version: version, Recipe: "f59571cf4be75adf", At: at})
	if err != nil {
		t.Fatal(err)
	}
	return key, value
}

// placing is a value of the state `placement` placing modules on node.
func placing(t *testing.T, node string, modules ...string) (string, []byte) {
	t.Helper()
	p := delivery.Placement{Node: node}
	for _, m := range modules {
		p.Modules = append(p.Modules, delivery.PlacedModule{Module: m})
	}
	key, err := delivery.PlacementKey(node)
	if err != nil {
		t.Fatal(err)
	}
	value, err := delivery.EncodePlacement(p)
	if err != nil {
		t.Fatal(err)
	}
	return key, value
}

func agreedOf(set *ToolchainSet) (string, string) {
	tc, none := delivery.AgreedToolchain(set.Entries("typescript"))
	return tc.Version, none
}

// Issue 1940, the gate over the catalogue's placement (issue 774, ADR 0549 §3): a build node's report counts only while
// the placement places its build module on its node, and only once written since; before the placement is settled, and
// while no node's placement is held, every report counts. The live cases: mesh-build-node taken off ace left
// ace.typescript at v26.12.0 against v26.12.1 on the three others; put on ace again, its report of 2026-10-23 06:29 came
// back before the module ran there.
func TestToolchainGate(t *testing.T) {
	old := time.Date(2026, 10, 23, 6, 29, 18, 0, time.UTC)     // ace's report from its earlier run
	placed := time.Date(2026, 10, 25, 19, 56, 45, 0, time.UTC) // mesh-build-node placed on ace again
	three := []string{"g14", "novox", "shanks"}
	start := func(t *testing.T, seeded bool) (*ToolchainGate, *ToolchainSet, *[]string) {
		set := &ToolchainSet{}
		said := &[]string{}
		gate := NewToolchainGate([]string{GoBuildModule, NodeBuildModule}, set.Put, func(f string, a ...any) { *said = append(*said, fmt.Sprintf(f, a...)) })
		for _, node := range three {
			if seeded {
				key, value := placing(t, node, NodeBuildModule)
				gate.Placement(key, value, false, old.Add(-time.Hour))
			}
			key, value := gateReport(t, node, "v26.12.1", placed.Add(time.Minute))
			gate.Toolchain(NodeBuildModule, key, value, false)
		}
		return gate, set, said
	}

	t.Run("not placed on ace: its report is no builder's", func(t *testing.T) {
		gate, set, said := start(t, true)
		key, value := gateReport(t, "ace", "v26.12.0", old)
		gate.Toolchain(NodeBuildModule, key, value, false)
		if _, none := agreedOf(set); !strings.HasPrefix(none, "builders disagree") {
			t.Fatalf("before the placement is settled every report counts; got %q", none)
		}
		k, v := placing(t, "ace", GoBuildModule) // another build module on ace: no matter
		gate.Placement(k, v, false, old)
		if !gate.Settled() {
			t.Fatal("settling left ace's report out, so what was passed on changed")
		}
		if v, none := agreedOf(set); v != "v26.12.1" || none != "" {
			t.Fatalf("the placed nodes agree on v26.12.1; got %q, %q", v, none)
		}
		if len(*said) != 1 || !strings.Contains((*said)[0], "does not place mesh-build-node on ace") {
			t.Fatalf("ace's report is said left out once; said %q", *said)
		}
		if gate.Toolchain(NodeBuildModule, key, value, false) {
			t.Fatal("a report of a node not placed changed what was passed on")
		}
		// Placed on ace with a report written since: it counts, and the builders really disagree.
		k, v = placing(t, "ace", GoBuildModule, NodeBuildModule)
		gate.Placement(k, v, false, old.Add(-time.Minute))
		if _, none := agreedOf(set); !strings.HasPrefix(none, "builders disagree: v26.12.0 recipe f59571cf4be75adf on ace") {
			t.Fatalf("builders that run different versions still disagree; got %q", none)
		}
		// Taken off ace again (its key deleted): its report goes again.
		if !gate.Placement(k, nil, true, placed) {
			t.Fatal("taking ace's placement away passed nothing on")
		}
		if v, none := agreedOf(set); v != "v26.12.1" || none != "" {
			t.Fatalf("got %q, %q", v, none)
		}
		// A report deleted from the bucket goes whether its node is placed or not.
		key, _ = gateReport(t, "g14", "v26.12.1", placed)
		if !gate.Toolchain(NodeBuildModule, key, nil, true) || len(set.Entries("typescript")) != 2 {
			t.Fatalf("deleting g14's report: %d entries left, want novox and shanks", len(set.Entries("typescript")))
		}
	})

	for _, c := range []struct {
		name   string
		at     time.Time // when ace wrote its report
		counts bool
	}{
		{"a report left before the placement", old, false},
		{"a report written since the placement", placed.Add(30 * time.Second), true},
	} {
		t.Run(c.name, func(t *testing.T) {
			gate, set, said := start(t, true)
			k, v := placing(t, "ace") // ace's key, first heard placing nothing
			gate.Placement(k, v, false, old.Add(-time.Hour))
			gate.Settled()
			k, v = placing(t, "ace", NodeBuildModule)
			if gate.Placement(k, v, false, placed) {
				t.Fatal("a placement with no report passed something on")
			}
			key, value := gateReport(t, "ace", "v26.12.0", c.at)
			if got := gate.Toolchain(NodeBuildModule, key, value, false); got != c.counts {
				t.Fatalf("ace's report passed on: %v, want %v", got, c.counts)
			}
			// ace's placement written again with another module beside: its standing does not move, nothing is said again.
			k, v = placing(t, "ace", GoBuildModule, NodeBuildModule)
			if gate.Placement(k, v, false, placed.Add(time.Minute)) {
				t.Fatal("the same placement of mesh-build-node changed what was passed on")
			}
			if c.counts {
				return
			}
			if v, none := agreedOf(set); v != "v26.12.1" || none != "" {
				t.Fatalf("ace's left report is no builder's; got %q, %q", v, none)
			}
			if len(*said) != 1 || !strings.Contains((*said)[0], "before mesh-build-node was placed on ace at 2026-10-25T19:56:45Z") {
				t.Fatalf("said %q, want once why ace's report is left out", *said)
			}
			key, value = gateReport(t, "ace", "v26.12.1", placed.Add(time.Minute))
			if !gate.Toolchain(NodeBuildModule, key, value, false) || len(set.Entries("typescript")) != 4 {
				t.Fatal("ace's new report was not passed on")
			}
		})
	}

	t.Run("no placement held: every report counts", func(t *testing.T) {
		gate, set, _ := start(t, false)
		key, value := gateReport(t, "ace", "v26.12.0", old)
		gate.Toolchain(NodeBuildModule, key, value, false)
		gate.Settled()
		if n := len(set.Entries("typescript")); n != 4 {
			t.Fatalf("before the first placement file every report counts; got %d", n)
		}
		k, v := placing(t, "g14", NodeBuildModule)
		if !gate.Placement(k, v, false, old.Add(-time.Hour)) || len(set.Entries("typescript")) != 1 {
			t.Fatalf("the first placement leaves every node it does not place out; got %d", len(set.Entries("typescript")))
		}
		if !gate.Placement(k, nil, true, placed) || len(set.Entries("typescript")) != 4 {
			t.Fatal("the last placement gone: every report counts again")
		}
	})

	t.Run("the history: a module in the first revision heard of a key stands since before it", func(t *testing.T) {
		p := delivery.NewPlaced()
		k, v := placing(t, "ace", NodeBuildModule)
		p.Put(k, v, false, old)
		k2, v2 := placing(t, "g14")
		p.Put(k2, v2, false, old)
		k2, v2 = placing(t, "g14", NodeBuildModule)
		if _, changed, _ := p.Put(k2, v2, false, placed); strings.Join(changed, ",") != NodeBuildModule {
			t.Errorf("a module placed says changed %v", changed)
		}
		if _, _, err := p.Put("placement.novox", []byte("not json"), false, old); err == nil {
			t.Error("a value that does not read was not said")
		}
		if s, ok := p.Since("ace", NodeBuildModule); !ok || !s.IsZero() {
			t.Errorf("ace's module stands since %v, %v; want the zero time", s, ok)
		}
		if s, ok := p.Since("g14", NodeBuildModule); !ok || !s.Equal(placed) {
			t.Errorf("g14's module stands since %v, %v; want the revision that placed it", s, ok)
		}
		if got := p.PlacedNodes(NodeBuildModule); strings.Join(got, ",") != "ace,g14" || len(p.ByNode()) != 2 || !p.Seeded() {
			t.Errorf("PlacedNodes = %v, ByNode = %v", got, p.ByNode())
		}
	})

	t.Run("a placement the bus refuses is said, and the part goes on", func(t *testing.T) {
		var logs bytes.Buffer
		logger := log.New(&logs, "", 0)
		refused := errors.New("nats: permissions violation for publish to $JS.API.CONSUMER.CREATE.KV_mesh-delivery-assigner_placement")
		gate, _, _ := start(t, false)
		for _, bus := range []refusingBus{{err: refused}, {refuseOpen: true, err: refused}} {
			if _, err := FollowPlacement(context.Background(), bus, gate, func(context.Context) {}, logger.Printf); err != nil {
				t.Errorf("a refused placement stopped the start: %v", err)
			}
		}
		if !strings.Contains(logs.String(), "cannot be watched") || !strings.Contains(logs.String(), "cannot be read") {
			t.Errorf("the refusals were not said: %q", logs.String())
		}
	})
}

// refusingBus is a bus that refuses: every open (refuseOpen), or every watch of a bucket it opens, with err.
type refusingBus struct {
	refuseOpen bool
	err        error
}

func (r refusingBus) KeyValue(context.Context, string) (jetstream.KeyValue, error) {
	if r.refuseOpen {
		return nil, r.err
	}
	return refusedWatch{err: r.err}, nil
}

type refusedWatch struct {
	jetstream.KeyValue
	err error
}

func (r refusedWatch) WatchAll(context.Context, ...jetstream.WatchOpt) (jetstream.KeyWatcher, error) {
	return nil, r.err
}