mirror of
https://github.com/go-delve/delve.git
synced 2026-05-07 05:06:32 +02:00
1994 lines
58 KiB
Go
1994 lines
58 KiB
Go
package terminal
|
|
|
|
import (
|
|
"bytes"
|
|
"cmp"
|
|
"flag"
|
|
"fmt"
|
|
"net"
|
|
"net/http"
|
|
"os"
|
|
"path/filepath"
|
|
"reflect"
|
|
"regexp"
|
|
"runtime"
|
|
"slices"
|
|
"strconv"
|
|
"strings"
|
|
"testing"
|
|
"time"
|
|
|
|
"github.com/go-delve/delve/pkg/config"
|
|
"github.com/go-delve/delve/pkg/goversion"
|
|
"github.com/go-delve/delve/pkg/logflags"
|
|
"github.com/go-delve/delve/pkg/proc/test"
|
|
"github.com/go-delve/delve/service"
|
|
"github.com/go-delve/delve/service/api"
|
|
"github.com/go-delve/delve/service/debugger"
|
|
"github.com/go-delve/delve/service/rpc2"
|
|
"github.com/go-delve/delve/service/rpccommon"
|
|
"github.com/go-delve/liner"
|
|
)
|
|
|
|
var testBackend, buildMode string
|
|
|
|
func TestMain(m *testing.M) {
|
|
flag.StringVar(&testBackend, "backend", "", "selects backend")
|
|
flag.StringVar(&buildMode, "test-buildmode", "", "selects build mode")
|
|
var logConf string
|
|
flag.StringVar(&logConf, "log", "", "configures logging")
|
|
flag.Parse()
|
|
test.DefaultTestBackend(&testBackend)
|
|
if buildMode != "" && buildMode != "pie" {
|
|
fmt.Fprintf(os.Stderr, "unknown build mode %q", buildMode)
|
|
os.Exit(1)
|
|
}
|
|
logflags.Setup(logConf != "", logConf, "")
|
|
test.RunTestsWithFixtures(m)
|
|
}
|
|
|
|
type FakeTerminal struct {
|
|
*Term
|
|
t testing.TB
|
|
}
|
|
|
|
const logCommandOutput = false
|
|
|
|
func (ft *FakeTerminal) Exec(cmdstr string) (outstr string, err error) {
|
|
var buf bytes.Buffer
|
|
ft.Term.stdout.pw.w = &buf
|
|
if ft.Term.starlarkEnv != nil {
|
|
ft.Term.starlarkEnv.Redirect(ft.Term.stdout)
|
|
}
|
|
err = ft.cmds.Call(cmdstr, ft.Term)
|
|
outstr = buf.String()
|
|
if logCommandOutput {
|
|
ft.t.Logf("command %q -> %q", cmdstr, outstr)
|
|
}
|
|
ft.Term.stdout.Flush()
|
|
return
|
|
}
|
|
|
|
func (ft *FakeTerminal) ExecStarlark(starlarkProgram string) (outstr string, err error) {
|
|
var buf bytes.Buffer
|
|
ft.Term.stdout.pw.w = &buf
|
|
ft.Term.starlarkEnv.Redirect(ft.Term.stdout)
|
|
_, err = ft.Term.starlarkEnv.Execute("<stdin>", starlarkProgram, "main", nil)
|
|
outstr = buf.String()
|
|
if logCommandOutput {
|
|
ft.t.Logf("command %q -> %q", starlarkProgram, outstr)
|
|
}
|
|
ft.Term.stdout.Flush()
|
|
return
|
|
}
|
|
|
|
func (ft *FakeTerminal) MustExec(cmdstr string) string {
|
|
ft.t.Helper()
|
|
outstr, err := ft.Exec(cmdstr)
|
|
if err != nil {
|
|
ft.t.Errorf("output of %q: %q", cmdstr, outstr)
|
|
ft.t.Fatalf("Error executing <%s>: %v", cmdstr, err)
|
|
}
|
|
return outstr
|
|
}
|
|
|
|
func (ft *FakeTerminal) MustExecStarlark(starlarkProgram string) string {
|
|
outstr, err := ft.ExecStarlark(starlarkProgram)
|
|
if err != nil {
|
|
ft.t.Errorf("output of %q: %q", starlarkProgram, outstr)
|
|
ft.t.Fatalf("Error executing <%s>: %v", starlarkProgram, err)
|
|
}
|
|
return outstr
|
|
}
|
|
|
|
func (ft *FakeTerminal) AssertExec(cmdstr, tgt string) {
|
|
out := ft.MustExec(cmdstr)
|
|
if out != tgt {
|
|
ft.t.Fatalf("Error executing %q, expected %q got %q", cmdstr, tgt, out)
|
|
}
|
|
}
|
|
|
|
func (ft *FakeTerminal) AssertExecError(cmdstr, tgterr string) {
|
|
_, err := ft.Exec(cmdstr)
|
|
if err == nil {
|
|
ft.t.Fatalf("Expected error executing %q", cmdstr)
|
|
}
|
|
if err.Error() != tgterr {
|
|
ft.t.Fatalf("Expected error %q executing %q, got error %q", tgterr, cmdstr, err.Error())
|
|
}
|
|
}
|
|
|
|
func withTestTerminal(name string, t testing.TB, fn func(*FakeTerminal)) {
|
|
withTestTerminalBuildFlags(name, t, 0, fn)
|
|
}
|
|
|
|
func withTestTerminalBuildFlags(name string, t testing.TB, buildFlags test.BuildFlags, fn func(*FakeTerminal)) {
|
|
if testBackend == "rr" {
|
|
test.MustHaveRecordingAllowed(t)
|
|
}
|
|
t.Setenv("TERM", "dumb")
|
|
listener, err := net.Listen("tcp", "127.0.0.1:0")
|
|
if err != nil {
|
|
t.Fatalf("couldn't start listener: %s\n", err)
|
|
}
|
|
defer listener.Close()
|
|
if buildMode == "pie" {
|
|
buildFlags |= test.BuildModePIE
|
|
}
|
|
server := rpccommon.NewServer(&service.Config{
|
|
Listener: listener,
|
|
ProcessArgs: []string{test.BuildFixture(t, name, buildFlags).Path},
|
|
Debugger: debugger.Config{
|
|
Backend: testBackend,
|
|
},
|
|
})
|
|
if err := server.Run(); err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
client := rpc2.NewClient(listener.Addr().String())
|
|
defer func() {
|
|
client.Detach(true)
|
|
}()
|
|
|
|
ft := &FakeTerminal{
|
|
t: t,
|
|
Term: New(client, &config.Config{}),
|
|
}
|
|
fn(ft)
|
|
}
|
|
|
|
func TestCommandDefault(t *testing.T) {
|
|
var (
|
|
cmds = Commands{}
|
|
cmd = cmds.Find("non-existent-command", noPrefix).cmdFn
|
|
)
|
|
|
|
err := cmd(nil, callContext{}, "")
|
|
if err == nil {
|
|
t.Fatal("cmd() did not default")
|
|
}
|
|
|
|
if err.Error() != "command not available" {
|
|
t.Fatal("wrong command output")
|
|
}
|
|
}
|
|
|
|
func TestCommandReplayWithoutPreviousCommand(t *testing.T) {
|
|
var (
|
|
cmds = DebugCommands(nil)
|
|
cmd = cmds.Find("", noPrefix).cmdFn
|
|
err = cmd(nil, callContext{}, "")
|
|
)
|
|
|
|
if err != nil {
|
|
t.Error("Null command not returned", err)
|
|
}
|
|
}
|
|
|
|
func TestCommandThread(t *testing.T) {
|
|
var (
|
|
cmds = DebugCommands(nil)
|
|
cmd = cmds.Find("thread", noPrefix).cmdFn
|
|
)
|
|
|
|
err := cmd(nil, callContext{}, "")
|
|
if err == nil {
|
|
t.Fatal("thread terminal command did not default")
|
|
}
|
|
|
|
if err.Error() != "you must specify a thread" {
|
|
t.Fatal("wrong command output: ", err.Error())
|
|
}
|
|
}
|
|
|
|
func TestExecuteFile(t *testing.T) {
|
|
breakCount := 0
|
|
traceCount := 0
|
|
c := &Commands{
|
|
client: nil,
|
|
cmds: []command{
|
|
{aliases: []string{"trace"}, cmdFn: func(t *Term, ctx callContext, args string) error {
|
|
traceCount++
|
|
return nil
|
|
}},
|
|
{aliases: []string{"break"}, cmdFn: func(t *Term, ctx callContext, args string) error {
|
|
breakCount++
|
|
return nil
|
|
}},
|
|
},
|
|
}
|
|
|
|
fixturesDir := test.FindFixturesDir()
|
|
err := c.executeFile(nil, filepath.Join(fixturesDir, "bpfile"))
|
|
if err != nil {
|
|
t.Fatalf("executeFile: %v", err)
|
|
}
|
|
|
|
if breakCount != 1 || traceCount != 1 {
|
|
t.Fatalf("Wrong counts break: %d trace: %d\n", breakCount, traceCount)
|
|
}
|
|
}
|
|
|
|
func TestIssue354(t *testing.T) {
|
|
printStack(&Term{}, os.Stdout, []api.Stackframe{}, "", false)
|
|
printStack(&Term{}, os.Stdout, []api.Stackframe{
|
|
{Location: api.Location{PC: 0, File: "irrelevant.go", Line: 10, Function: nil},
|
|
Bottom: true}}, "", false)
|
|
}
|
|
|
|
func TestIssue411(t *testing.T) {
|
|
test.AllowRecording(t)
|
|
withTestTerminal("math", t, func(term *FakeTerminal) {
|
|
term.MustExec("break _fixtures/math.go:8")
|
|
term.MustExec("trace _fixtures/math.go:9")
|
|
term.MustExec("continue")
|
|
out := term.MustExec("next")
|
|
if !strings.HasPrefix(out, "> goroutine(1): main.main()") {
|
|
t.Fatalf("Wrong output for next: <%s>", out)
|
|
}
|
|
})
|
|
}
|
|
|
|
func TestCustomCommandNestedBreakpoint(t *testing.T) {
|
|
withTestTerminal("nestedbp", t, func(term *FakeTerminal) {
|
|
term.MustExec("source " + findStarFile("test_custom_cmd_nested"))
|
|
|
|
term.MustExec("break nestedbp.go:13") // BP 1
|
|
term.MustExec("break nestedbp.go:18") // BP 2
|
|
term.MustExec("break nestedbp.go:19") // BP 3
|
|
|
|
// Add custom commands to BP1:
|
|
// 1. cmd before continue (should execute)
|
|
// 2. continue cmd (should execute and hit BP2)
|
|
// 3. cmd after continue (should NOT execute)
|
|
term.MustExec("on 1 test_bp1_before_continue")
|
|
term.MustExec("on 1 test_bp1_continue_cmd")
|
|
term.MustExec("on 1 test_bp1_after_continue")
|
|
|
|
// Add custom commands to BP2 which should execute due to BP 1 custom command continue:
|
|
term.MustExec(fmt.Sprintf("on %d test_bp2_cmd", 2))
|
|
term.MustExec("on 2 test_bp1_continue_cmd")
|
|
term.MustExec("on 2 test_bp1_after_continue")
|
|
|
|
// Add command on BP3 which should execute due to BP 2 custom command continue:
|
|
term.MustExec("on 3 test_bp3_cmd")
|
|
|
|
out := term.MustExec("continue")
|
|
|
|
if !strings.Contains(out, "BP1_BEFORE_CONTINUE") {
|
|
t.Errorf("expected BP1_BEFORE_CONTINUE to be printed, got: %q", out)
|
|
}
|
|
|
|
if !strings.Contains(out, "BP1_CONTINUE_CMD") {
|
|
t.Errorf("expected BP1_CONTINUE_CMD to be printed, got: %q", out)
|
|
}
|
|
|
|
// Verify BP1 command after continue did NOT execute
|
|
if strings.Contains(out, "BP1_AFTER_CONTINUE") {
|
|
t.Errorf("BP1_AFTER_CONTINUE was printed, which means the command ran after continue, got: %q", out)
|
|
}
|
|
|
|
if !strings.Contains(out, "BP2_CMD_EXECUTED") {
|
|
t.Errorf("BP2's custom commands not executed")
|
|
}
|
|
if !strings.Contains(out, "BP3_CMD_EXECUTED") {
|
|
t.Errorf("BP3's custom commands not executed")
|
|
}
|
|
})
|
|
}
|
|
|
|
func TestTrace(t *testing.T) {
|
|
test.AllowRecording(t)
|
|
withTestTerminal("issue573", t, func(term *FakeTerminal) {
|
|
term.MustExec("trace foo")
|
|
out, _ := term.Exec("continue")
|
|
// The output here is a little strange, but we don't filter stdout vs stderr so it gets jumbled.
|
|
// Therefore we assert about the call and return values separately.
|
|
if !strings.Contains(out, "> goroutine(1): main.foo(99, 9801)") {
|
|
t.Fatalf("Wrong output for tracepoint: %s", out)
|
|
}
|
|
if !strings.Contains(out, "=> (9900)") {
|
|
t.Fatalf("Wrong output for tracepoint return value: %s", out)
|
|
}
|
|
})
|
|
}
|
|
|
|
func TestTraceWithName(t *testing.T) {
|
|
test.AllowRecording(t)
|
|
withTestTerminal("issue573", t, func(term *FakeTerminal) {
|
|
term.MustExec("trace foobar foo")
|
|
out, _ := term.Exec("continue")
|
|
// The output here is a little strange, but we don't filter stdout vs stderr so it gets jumbled.
|
|
// Therefore we assert about the call and return values separately.
|
|
if !strings.Contains(out, "> goroutine(1): [foobar] main.foo(99, 9801)") {
|
|
t.Fatalf("Wrong output for tracepoint: %s", out)
|
|
}
|
|
if !strings.Contains(out, "=> (9900)") {
|
|
t.Fatalf("Wrong output for tracepoint return value: %s", out)
|
|
}
|
|
})
|
|
}
|
|
|
|
func TestTraceOnNonFunctionEntry(t *testing.T) {
|
|
test.AllowRecording(t)
|
|
withTestTerminal("issue573", t, func(term *FakeTerminal) {
|
|
term.MustExec("trace foobar issue573.go:19")
|
|
out, _ := term.Exec("continue")
|
|
if !strings.Contains(out, "> goroutine(1): [foobar] main.foo(99, 9801)") {
|
|
t.Fatalf("Wrong output for tracepoint: %s", out)
|
|
}
|
|
if strings.Contains(out, "=> (9900)") {
|
|
t.Fatalf("Tracepoint on non-function locspec should not have return value:\n%s", out)
|
|
}
|
|
})
|
|
}
|
|
|
|
func TestExitStatus(t *testing.T) {
|
|
withTestTerminal("continuetestprog", t, func(term *FakeTerminal) {
|
|
term.Exec("continue")
|
|
status, err := term.handleExit()
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
if status != 0 {
|
|
t.Fatalf("incorrect exit status, expected 0, got %d", status)
|
|
}
|
|
})
|
|
}
|
|
|
|
func TestScopePrefix(t *testing.T) {
|
|
if runtime.GOARCH == "ppc64le" && buildMode == "pie" {
|
|
t.Skip("pie mode broken on ppc64le")
|
|
}
|
|
const goroutinesLinePrefix = " Goroutine "
|
|
const goroutinesCurLinePrefix = "* Goroutine "
|
|
test.AllowRecording(t)
|
|
|
|
lenient := 0
|
|
if runtime.GOOS == "windows" {
|
|
lenient = 1
|
|
}
|
|
|
|
withTestTerminal("goroutinestackprog", t, func(term *FakeTerminal) {
|
|
term.MustExec("b stacktraceme")
|
|
term.MustExec("continue")
|
|
|
|
goroutinesOut := strings.Split(term.MustExec("goroutines"), "\n")
|
|
agoroutines := []int{}
|
|
nonagoroutines := []int{}
|
|
curgid := -1
|
|
|
|
for _, line := range goroutinesOut {
|
|
iscur := strings.HasPrefix(line, goroutinesCurLinePrefix)
|
|
if !iscur && !strings.HasPrefix(line, goroutinesLinePrefix) {
|
|
continue
|
|
}
|
|
|
|
dash := strings.Index(line, " - ")
|
|
if dash < 0 {
|
|
continue
|
|
}
|
|
|
|
gid, err := strconv.Atoi(line[len(goroutinesLinePrefix):dash])
|
|
if err != nil {
|
|
continue
|
|
}
|
|
|
|
if iscur {
|
|
curgid = gid
|
|
}
|
|
|
|
if found := strings.Contains(line, " main.agoroutine "); !found {
|
|
nonagoroutines = append(nonagoroutines, gid)
|
|
continue
|
|
}
|
|
|
|
agoroutines = append(agoroutines, gid)
|
|
}
|
|
|
|
if len(agoroutines) > 10 {
|
|
t.Fatalf("Output of goroutines did not have 10 goroutines stopped on main.agoroutine (%d found): %q", len(agoroutines), goroutinesOut)
|
|
}
|
|
|
|
if len(agoroutines) < 10 {
|
|
extraAgoroutines := 0
|
|
for _, gid := range nonagoroutines {
|
|
stackOut := strings.SplitSeq(term.MustExec(fmt.Sprintf("goroutine %d stack", gid)), "\n")
|
|
for line := range stackOut {
|
|
if strings.HasSuffix(line, " main.agoroutine") {
|
|
extraAgoroutines++
|
|
break
|
|
}
|
|
}
|
|
}
|
|
if len(agoroutines)+extraAgoroutines < 10-lenient {
|
|
t.Fatalf("Output of goroutines did not have 10 goroutines stopped on main.agoroutine (%d+%d found): %q", len(agoroutines), extraAgoroutines, goroutinesOut)
|
|
}
|
|
}
|
|
|
|
if curgid < 0 {
|
|
t.Fatalf("Could not find current goroutine in output of goroutines: %q", goroutinesOut)
|
|
}
|
|
|
|
seen := make([]bool, 10)
|
|
for _, gid := range agoroutines {
|
|
stackOut := strings.Split(term.MustExec(fmt.Sprintf("goroutine %d stack", gid)), "\n")
|
|
fid := -1
|
|
for _, line := range stackOut {
|
|
line = strings.TrimLeft(line, " ")
|
|
before, _, ok := strings.Cut(line, " ")
|
|
if !ok {
|
|
continue
|
|
}
|
|
curfid, err := strconv.Atoi(before)
|
|
if err != nil {
|
|
continue
|
|
}
|
|
|
|
if found := strings.Contains(line, " main.agoroutine"); found {
|
|
fid = curfid
|
|
break
|
|
}
|
|
}
|
|
if fid < 0 {
|
|
t.Fatalf("Could not find frame for goroutine %d: %q", gid, stackOut)
|
|
}
|
|
term.AssertExec(fmt.Sprintf("goroutine %d frame %d locals", gid, fid), "(no locals)\n")
|
|
argsOut := strings.Split(term.MustExec(fmt.Sprintf("goroutine %d frame %d args", gid, fid)), "\n")
|
|
if len(argsOut) != 4 || argsOut[3] != "" {
|
|
t.Fatalf("Wrong number of arguments in goroutine %d frame %d: %v", gid, fid, argsOut)
|
|
}
|
|
out := term.MustExec(fmt.Sprintf("goroutine %d frame %d p i", gid, fid))
|
|
ival, err := strconv.Atoi(out[:len(out)-1])
|
|
if err != nil {
|
|
t.Fatalf("could not parse value %q of i for goroutine %d frame %d: %v", out, gid, fid, err)
|
|
}
|
|
seen[ival] = true
|
|
}
|
|
|
|
for i := range seen {
|
|
if !seen[i] {
|
|
if lenient > 0 {
|
|
lenient--
|
|
} else {
|
|
t.Fatalf("goroutine %d not found", i)
|
|
}
|
|
}
|
|
}
|
|
|
|
term.MustExec("c")
|
|
|
|
term.AssertExecError("frame", "not enough arguments")
|
|
term.AssertExecError(fmt.Sprintf("goroutine %d frame 10 locals", curgid), fmt.Sprintf("Frame 10 does not exist in goroutine %d", curgid))
|
|
term.AssertExecError("goroutine 9000 locals", "unknown goroutine 9000")
|
|
|
|
term.AssertExecError("print n", "could not find symbol value for n")
|
|
term.AssertExec("frame 1 print n", "3\n")
|
|
term.AssertExec("frame 2 print n", "2\n")
|
|
term.AssertExec("frame 3 print n", "1\n")
|
|
term.AssertExec("frame 4 print n", "0\n")
|
|
term.AssertExecError("frame 5 print n", "could not find symbol value for n")
|
|
|
|
term.MustExec("frame 2")
|
|
term.AssertExec("print n", "2\n")
|
|
term.MustExec("frame 4")
|
|
term.AssertExec("print n", "0\n")
|
|
term.MustExec("down")
|
|
term.AssertExec("print n", "1\n")
|
|
term.MustExec("down 2")
|
|
term.AssertExec("print n", "3\n")
|
|
term.AssertExecError("down 2", "Invalid frame -1")
|
|
term.AssertExec("print n", "3\n")
|
|
term.MustExec("up 2")
|
|
term.AssertExec("print n", "1\n")
|
|
term.AssertExecError("up 100", "Invalid frame 103")
|
|
term.AssertExec("print n", "1\n")
|
|
|
|
term.MustExec("step")
|
|
term.AssertExecError("print n", "could not find symbol value for n")
|
|
term.MustExec("frame 2")
|
|
term.AssertExec("print n", "2\n")
|
|
})
|
|
}
|
|
|
|
func TestOnPrefix(t *testing.T) {
|
|
const prefix = "\ti: "
|
|
test.AllowRecording(t)
|
|
lenient := false
|
|
if runtime.GOOS == "windows" {
|
|
lenient = true
|
|
}
|
|
withTestTerminal("goroutinestackprog", t, func(term *FakeTerminal) {
|
|
term.MustExec("b agobp main.agoroutine")
|
|
term.MustExec("on agobp print i")
|
|
|
|
seen := make([]bool, 10)
|
|
|
|
for {
|
|
outstr, err := term.Exec("continue")
|
|
if err != nil {
|
|
if !strings.Contains(err.Error(), "exited") {
|
|
t.Fatalf("Unexpected error executing 'continue': %v", err)
|
|
}
|
|
break
|
|
}
|
|
out := strings.Split(outstr, "\n")
|
|
|
|
for i := range out {
|
|
if !strings.HasPrefix(out[i], prefix) {
|
|
continue
|
|
}
|
|
id, err := strconv.Atoi(out[i][len(prefix):])
|
|
if err != nil {
|
|
continue
|
|
}
|
|
if seen[id] {
|
|
t.Fatalf("Goroutine %d seen twice\n", id)
|
|
}
|
|
seen[id] = true
|
|
}
|
|
}
|
|
|
|
for i := range seen {
|
|
if !seen[i] {
|
|
if lenient {
|
|
lenient = false
|
|
} else {
|
|
t.Fatalf("Goroutine %d not seen\n", i)
|
|
}
|
|
}
|
|
}
|
|
})
|
|
}
|
|
|
|
func TestNoVars(t *testing.T) {
|
|
test.AllowRecording(t)
|
|
withTestTerminal("locationsUpperCase", t, func(term *FakeTerminal) {
|
|
term.MustExec("b main.main")
|
|
term.MustExec("continue")
|
|
term.AssertExec("args", "(no args)\n")
|
|
term.AssertExec("locals", "(no locals)\n")
|
|
term.AssertExec("vars filterThatMatchesNothing", "(no vars)\n")
|
|
})
|
|
}
|
|
|
|
func TestOnPrefixLocals(t *testing.T) {
|
|
const prefix = "\ti: "
|
|
test.AllowRecording(t)
|
|
withTestTerminal("goroutinestackprog", t, func(term *FakeTerminal) {
|
|
term.MustExec("b agobp main.agoroutine")
|
|
term.MustExec("on agobp args -v")
|
|
|
|
seen := make([]bool, 10)
|
|
|
|
for {
|
|
outstr, err := term.Exec("continue")
|
|
if err != nil {
|
|
if !strings.Contains(err.Error(), "exited") {
|
|
t.Fatalf("Unexpected error executing 'continue': %v", err)
|
|
}
|
|
break
|
|
}
|
|
out := strings.Split(outstr, "\n")
|
|
|
|
for i := range out {
|
|
if !strings.HasPrefix(out[i], prefix) {
|
|
continue
|
|
}
|
|
id, err := strconv.Atoi(out[i][len(prefix):])
|
|
if err != nil {
|
|
continue
|
|
}
|
|
if seen[id] {
|
|
t.Fatalf("Goroutine %d seen twice\n", id)
|
|
}
|
|
seen[id] = true
|
|
}
|
|
}
|
|
|
|
for i := range seen {
|
|
if !seen[i] {
|
|
t.Fatalf("Goroutine %d not seen\n", i)
|
|
}
|
|
}
|
|
})
|
|
}
|
|
|
|
func listIsAt(t *testing.T, term *FakeTerminal, listcmd string, cur, start, end int) {
|
|
t.Helper()
|
|
outstr := term.MustExec(listcmd)
|
|
lines := strings.Split(outstr, "\n")
|
|
|
|
t.Logf("%q: %q", listcmd, outstr)
|
|
|
|
if cur >= 0 && !strings.Contains(lines[0], fmt.Sprintf(":%d", cur)) {
|
|
t.Fatalf("Could not find current line number in first output line: %q", lines[0])
|
|
}
|
|
|
|
re := regexp.MustCompile(`(=>)?\s+(\d+):`)
|
|
|
|
outStart, outEnd := 0, 0
|
|
|
|
for _, line := range lines[1:] {
|
|
if line == "" {
|
|
continue
|
|
}
|
|
v := re.FindStringSubmatch(line)
|
|
if len(v) != 3 {
|
|
continue
|
|
}
|
|
curline, _ := strconv.Atoi(v[2])
|
|
if v[1] == "=>" {
|
|
if cur != curline {
|
|
t.Fatalf("Wrong current line, got %d expected %d", curline, cur)
|
|
}
|
|
}
|
|
if outStart == 0 {
|
|
outStart = curline
|
|
}
|
|
outEnd = curline
|
|
}
|
|
|
|
if start != -1 || end != -1 {
|
|
if outStart != start || outEnd != end {
|
|
t.Fatalf("Wrong output range, got %d:%d expected %d:%d", outStart, outEnd, start, end)
|
|
}
|
|
}
|
|
}
|
|
|
|
func TestListCmd(t *testing.T) {
|
|
withTestTerminal("testvariables", t, func(term *FakeTerminal) {
|
|
term.MustExec("continue")
|
|
term.MustExec("continue")
|
|
listIsAt(t, term, "list", 27, 22, 32)
|
|
listIsAt(t, term, "list 69", 69, 64, 74)
|
|
listIsAt(t, term, "frame 1 list", 66, 61, 71)
|
|
listIsAt(t, term, "frame 1 list 69", 69, 64, 74)
|
|
_, err := term.Exec("frame 50 list")
|
|
if err == nil {
|
|
t.Fatalf("Expected error requesting 50th frame")
|
|
}
|
|
listIsAt(t, term, "list testvariables.go:1", -1, 1, 6)
|
|
listIsAt(t, term, "list testvariables.go:10000", -1, 0, 0)
|
|
})
|
|
}
|
|
|
|
func TestReverseContinue(t *testing.T) {
|
|
test.AllowRecording(t)
|
|
if testBackend != "rr" {
|
|
return
|
|
}
|
|
withTestTerminal("continuetestprog", t, func(term *FakeTerminal) {
|
|
term.MustExec("break main.main")
|
|
term.MustExec("break main.sayhi")
|
|
listIsAt(t, term, "continue", 16, -1, -1)
|
|
listIsAt(t, term, "continue", 12, -1, -1)
|
|
listIsAt(t, term, "rewind", 16, -1, -1)
|
|
})
|
|
}
|
|
|
|
func TestCheckpoints(t *testing.T) {
|
|
test.AllowRecording(t)
|
|
if testBackend != "rr" {
|
|
return
|
|
}
|
|
withTestTerminal("continuetestprog", t, func(term *FakeTerminal) {
|
|
term.MustExec("break main.main")
|
|
listIsAt(t, term, "continue", 16, -1, -1)
|
|
term.MustExec("checkpoint")
|
|
term.MustExec("checkpoints")
|
|
listIsAt(t, term, "next", 17, -1, -1)
|
|
listIsAt(t, term, "next", 18, -1, -1)
|
|
term.MustExec("restart c1")
|
|
term.MustExec("goroutine 1")
|
|
listIsAt(t, term, "list", 16, -1, -1)
|
|
})
|
|
}
|
|
|
|
func TestNextWithCount(t *testing.T) {
|
|
test.AllowRecording(t)
|
|
withTestTerminal("nextcond", t, func(term *FakeTerminal) {
|
|
term.MustExec("break main.main")
|
|
listIsAt(t, term, "continue", 8, -1, -1)
|
|
listIsAt(t, term, "next 2", 10, -1, -1)
|
|
})
|
|
}
|
|
|
|
func TestRestart(t *testing.T) {
|
|
withTestTerminal("restartargs", t, func(term *FakeTerminal) {
|
|
term.MustExec("break main.printArgs")
|
|
term.MustExec("continue")
|
|
if out := term.MustExec("print main.args"); !strings.Contains(out, ", []") {
|
|
t.Fatalf("wrong args: %q", out)
|
|
}
|
|
// Reset the arg list
|
|
term.MustExec("restart hello")
|
|
term.MustExec("continue")
|
|
if out := term.MustExec("print main.args"); !strings.Contains(out, ", [\"hello\"]") {
|
|
t.Fatalf("wrong args: %q ", out)
|
|
}
|
|
// Restart w/o arg should retain the current args.
|
|
term.MustExec("restart")
|
|
term.MustExec("continue")
|
|
if out := term.MustExec("print main.args"); !strings.Contains(out, ", [\"hello\"]") {
|
|
t.Fatalf("wrong args: %q ", out)
|
|
}
|
|
// Empty arg list
|
|
term.MustExec("restart -noargs")
|
|
term.MustExec("continue")
|
|
if out := term.MustExec("print main.args"); !strings.Contains(out, ", []") {
|
|
t.Fatalf("wrong args: %q ", out)
|
|
}
|
|
})
|
|
}
|
|
|
|
func TestIssue827(t *testing.T) {
|
|
// switching goroutines when the current thread isn't running any goroutine
|
|
// causes nil pointer dereference.
|
|
withTestTerminal("notify-v2", t, func(term *FakeTerminal) {
|
|
go func() {
|
|
time.Sleep(1 * time.Second)
|
|
resp, err := http.Get("http://127.0.0.1:8888/test")
|
|
if err == nil {
|
|
resp.Body.Close()
|
|
}
|
|
time.Sleep(1 * time.Second)
|
|
term.client.Halt()
|
|
}()
|
|
term.MustExec("continue")
|
|
term.MustExec("goroutine 1")
|
|
})
|
|
}
|
|
|
|
func findCmdName(c *Commands, cmdstr string, prefix cmdPrefix) string {
|
|
for _, v := range c.cmds {
|
|
if v.match(cmdstr) {
|
|
if prefix != noPrefix && v.allowedPrefixes&prefix == 0 {
|
|
continue
|
|
}
|
|
return v.aliases[0]
|
|
}
|
|
}
|
|
return ""
|
|
}
|
|
|
|
func assertNoError(t *testing.T, err error, str string) {
|
|
t.Helper()
|
|
if err != nil {
|
|
t.Fatalf("%s: %v", str, err)
|
|
}
|
|
}
|
|
|
|
func assertNoErrorConfigureCmd(t *testing.T, term *Term, cmdstr string) {
|
|
t.Helper()
|
|
err := configureCmd(term, callContext{}, cmdstr)
|
|
assertNoError(t, err, fmt.Sprintf("error executing configureCmd(%s)", cmdstr))
|
|
}
|
|
|
|
func assertSubstitutePath(t *testing.T, sp config.SubstitutePathRules, v ...string) {
|
|
t.Helper()
|
|
if len(sp) != len(v)/2 {
|
|
t.Fatalf("wrong number of substitute path rules (expected: %d): %#v", len(v)/2, sp)
|
|
}
|
|
for i := range sp {
|
|
if sp[i].From != v[i*2] || sp[i].To != v[i*2+1] {
|
|
t.Fatalf("wrong substitute path rule %#v expected (from: %q to %q)", sp[i], v[i*2], v[i*2+1])
|
|
}
|
|
}
|
|
}
|
|
|
|
func assertDebugInfoDirs(t *testing.T, got []string, tgt ...string) {
|
|
if len(got) != len(tgt) {
|
|
t.Fatalf("wrong number of debug info directories (got %d expected %d)", len(got), len(tgt))
|
|
}
|
|
for i := range got {
|
|
if got[i] != tgt[i] {
|
|
t.Fatalf("debug info directories mismatch got: %v expected: %v", got, tgt)
|
|
}
|
|
}
|
|
}
|
|
|
|
func TestConfig(t *testing.T) {
|
|
var buf bytes.Buffer
|
|
var term Term
|
|
term.conf = &config.Config{}
|
|
term.cmds = DebugCommands(nil)
|
|
term.stdout = &transcriptWriter{pw: &pagingWriter{w: &buf}}
|
|
|
|
err := configureCmd(&term, callContext{}, "nonexistent-parameter 10")
|
|
if err == nil {
|
|
t.Fatalf("expected error executing configureCmd(nonexistent-parameter)")
|
|
}
|
|
|
|
assertNoErrorConfigureCmd(t, &term, "max-string-len 10")
|
|
if term.conf.MaxStringLen == nil {
|
|
t.Fatalf("expected MaxStringLen 10, got nil")
|
|
}
|
|
if *term.conf.MaxStringLen != 10 {
|
|
t.Fatalf("expected MaxStringLen 10, got: %d", *term.conf.MaxStringLen)
|
|
}
|
|
assertNoErrorConfigureCmd(t, &term, "show-location-expr true")
|
|
if term.conf.ShowLocationExpr != true {
|
|
t.Fatalf("expected ShowLocationExpr true, got false")
|
|
}
|
|
|
|
assertNoErrorConfigureCmd(t, &term, "max-variable-recurse 4")
|
|
if term.conf.MaxVariableRecurse == nil {
|
|
t.Fatalf("expected MaxVariableRecurse 4, got nil")
|
|
}
|
|
if *term.conf.MaxVariableRecurse != 4 {
|
|
t.Fatalf("expected MaxVariableRecurse 4, got: %d", *term.conf.MaxVariableRecurse)
|
|
}
|
|
|
|
assertNoErrorConfigureCmd(t, &term, "substitute-path a b")
|
|
assertSubstitutePath(t, term.conf.SubstitutePath, "a", "b")
|
|
|
|
assertNoErrorConfigureCmd(t, &term, "substitute-path a")
|
|
assertSubstitutePath(t, term.conf.SubstitutePath)
|
|
|
|
assertNoErrorConfigureCmd(t, &term, "alias print blah")
|
|
if len(term.conf.Aliases["print"]) != 1 {
|
|
t.Fatalf("aliases not changed after configure command %v", term.conf.Aliases)
|
|
}
|
|
if findCmdName(term.cmds, "blah", noPrefix) != "print" {
|
|
t.Fatalf("new alias not found")
|
|
}
|
|
|
|
assertNoErrorConfigureCmd(t, &term, "alias blah")
|
|
if len(term.conf.Aliases["print"]) != 0 {
|
|
t.Fatalf("alias not removed after configure command %v", term.conf.Aliases)
|
|
}
|
|
if findCmdName(term.cmds, "blah", noPrefix) != "" {
|
|
t.Fatalf("new alias found after delete")
|
|
}
|
|
|
|
err = configureCmd(&term, callContext{}, "show-location-expr")
|
|
if err == nil {
|
|
t.Fatalf("no error form configureCmd(show-location-expr)")
|
|
}
|
|
if !term.conf.ShowLocationExpr {
|
|
t.Fatalf("ShowLocationExpr not set to true")
|
|
}
|
|
|
|
assertNoErrorConfigureCmd(t, &term, "show-location-expr false")
|
|
if term.conf.ShowLocationExpr {
|
|
t.Fatalf("ShowLocationExpr set to true")
|
|
}
|
|
|
|
assertNoErrorConfigureCmd(t, &term, "substitute-path a b")
|
|
assertNoErrorConfigureCmd(t, &term, "substitute-path c d")
|
|
assertSubstitutePath(t, term.conf.SubstitutePath, "a", "b", "c", "d")
|
|
|
|
buf.Reset()
|
|
assertNoErrorConfigureCmd(t, &term, "substitute-path")
|
|
t.Logf("current substitute-path: %q", buf.String())
|
|
if buf.String() != "\"a\" → \"b\"\n\"c\" → \"d\"\n" {
|
|
t.Fatalf("wrong substitute-path value")
|
|
}
|
|
|
|
assertNoErrorConfigureCmd(t, &term, "substitute-path -clear c")
|
|
assertSubstitutePath(t, term.conf.SubstitutePath, "a", "b")
|
|
|
|
assertNoErrorConfigureCmd(t, &term, "substitute-path -clear")
|
|
assertSubstitutePath(t, term.conf.SubstitutePath)
|
|
|
|
assertNoErrorConfigureCmd(t, &term, "substitute-path \"\" something")
|
|
assertSubstitutePath(t, term.conf.SubstitutePath, "", "something")
|
|
|
|
assertNoErrorConfigureCmd(t, &term, "substitute-path somethingelse \"\"")
|
|
assertSubstitutePath(t, term.conf.SubstitutePath, "", "something", "somethingelse", "")
|
|
|
|
assertDebugInfoDirs(t, term.conf.DebugInfoDirectories)
|
|
|
|
assertNoErrorConfigureCmd(t, &term, "debug-info-directories -add a")
|
|
assertDebugInfoDirs(t, term.conf.DebugInfoDirectories, "a")
|
|
assertNoErrorConfigureCmd(t, &term, "debug-info-directories -add b")
|
|
assertDebugInfoDirs(t, term.conf.DebugInfoDirectories, "a", "b")
|
|
assertNoErrorConfigureCmd(t, &term, "debug-info-directories -add c")
|
|
assertDebugInfoDirs(t, term.conf.DebugInfoDirectories, "a", "b", "c")
|
|
assertNoErrorConfigureCmd(t, &term, "debug-info-directories -rm b")
|
|
assertDebugInfoDirs(t, term.conf.DebugInfoDirectories, "a", "c")
|
|
assertNoErrorConfigureCmd(t, &term, "debug-info-directories -clear")
|
|
assertDebugInfoDirs(t, term.conf.DebugInfoDirectories)
|
|
|
|
ft := &FakeTerminal{&term, t}
|
|
_, err = ft.Exec("help config blah")
|
|
if err == nil {
|
|
t.Errorf("expected error executing help config for an unknown parameter")
|
|
}
|
|
ft.MustExec("help config substitute-path")
|
|
}
|
|
|
|
func TestIssue1090(t *testing.T) {
|
|
// Exit while executing 'next' should report the "Process exited" error
|
|
// message instead of crashing.
|
|
withTestTerminal("math", t, func(term *FakeTerminal) {
|
|
term.MustExec("break main.main")
|
|
term.MustExec("continue")
|
|
for {
|
|
_, err := term.Exec("next")
|
|
if err != nil && strings.Contains(err.Error(), " has exited with status ") {
|
|
break
|
|
}
|
|
}
|
|
})
|
|
}
|
|
|
|
func TestPrintContextParkedGoroutine(t *testing.T) {
|
|
if runtime.GOARCH == "ppc64le" && buildMode == "pie" {
|
|
t.Skip("pie mode broken on ppc64le")
|
|
}
|
|
withTestTerminal("goroutinestackprog", t, func(term *FakeTerminal) {
|
|
term.MustExec("break stacktraceme")
|
|
term.MustExec("continue")
|
|
|
|
// pick a goroutine that isn't running on a thread
|
|
gid := ""
|
|
gout := strings.Split(term.MustExec("goroutines"), "\n")
|
|
t.Logf("goroutines -> %q", gout)
|
|
for _, gline := range gout {
|
|
if !strings.Contains(gline, "thread ") && strings.Contains(gline, "agoroutine") {
|
|
if dash := strings.Index(gline, " - "); dash > 0 {
|
|
gid = gline[len(" Goroutine "):dash]
|
|
break
|
|
}
|
|
}
|
|
}
|
|
|
|
t.Logf("picked %q", gid)
|
|
term.MustExec(fmt.Sprintf("goroutine %s", gid))
|
|
|
|
frameout := strings.Split(term.MustExec("frame 0"), "\n")
|
|
t.Logf("frame 0 -> %q", frameout)
|
|
if strings.Contains(frameout[0], "stacktraceme") {
|
|
t.Fatal("bad output for `frame 0` command on a parked goroutine")
|
|
}
|
|
|
|
listout := strings.Split(term.MustExec("list"), "\n")
|
|
t.Logf("list -> %q", listout)
|
|
if strings.Contains(listout[0], "stacktraceme") {
|
|
t.Fatal("bad output for list command on a parked goroutine")
|
|
}
|
|
})
|
|
}
|
|
|
|
func TestStepOutReturn(t *testing.T) {
|
|
ver, _ := goversion.Parse(runtime.Version())
|
|
if ver.Major >= 0 && !ver.AfterOrEqual(goversion.GoVersion{Major: 1, Minor: 10, Rev: -1}) {
|
|
t.Skip("return variables aren't marked on 1.9 or earlier")
|
|
}
|
|
withTestTerminal("stepoutret", t, func(term *FakeTerminal) {
|
|
term.MustExec("break main.stepout")
|
|
term.MustExec("continue")
|
|
out := term.MustExec("stepout")
|
|
t.Logf("output: %q", out)
|
|
if !strings.Contains(out, "num: ") || !strings.Contains(out, "str: ") {
|
|
t.Fatal("could not find parameter")
|
|
}
|
|
})
|
|
}
|
|
|
|
func TestOptimizationCheck(t *testing.T) {
|
|
withTestTerminal("continuetestprog", t, func(term *FakeTerminal) {
|
|
term.MustExec("break main.main")
|
|
out := term.MustExec("continue")
|
|
t.Logf("output %q", out)
|
|
if strings.Contains(out, optimizedFunctionWarning) {
|
|
t.Fatal("optimized function warning")
|
|
}
|
|
})
|
|
|
|
if goversion.VersionAfterOrEqual(runtime.Version(), 1, 10) {
|
|
withTestTerminalBuildFlags("continuetestprog", t, test.EnableOptimization|test.EnableInlining, func(term *FakeTerminal) {
|
|
term.MustExec("break main.main")
|
|
out := term.MustExec("continue")
|
|
t.Logf("output %q", out)
|
|
if !strings.Contains(out, optimizedFunctionWarning) {
|
|
t.Fatal("optimized function warning missing")
|
|
}
|
|
})
|
|
}
|
|
}
|
|
|
|
func TestTruncateStacktrace(t *testing.T) {
|
|
if runtime.GOARCH == "ppc64le" && buildMode == "pie" {
|
|
t.Skip("pie mode broken on ppc64le")
|
|
}
|
|
const stacktraceTruncatedMessage = "(truncated)"
|
|
withTestTerminal("stacktraceprog", t, func(term *FakeTerminal) {
|
|
term.MustExec("break main.stacktraceme")
|
|
term.MustExec("continue")
|
|
out1 := term.MustExec("stack")
|
|
t.Logf("untruncated output %q", out1)
|
|
if strings.Contains(out1, stacktraceTruncatedMessage) {
|
|
t.Fatalf("stacktrace was truncated")
|
|
}
|
|
out2 := term.MustExec("stack 1")
|
|
t.Logf("truncated output %q", out2)
|
|
if !strings.Contains(out2, stacktraceTruncatedMessage) {
|
|
t.Fatalf("stacktrace was not truncated")
|
|
}
|
|
})
|
|
}
|
|
|
|
func TestIssue1493(t *testing.T) {
|
|
// The 'regs' command without the '-a' option should only return
|
|
// general purpose registers.
|
|
if runtime.GOARCH == "ppc64le" {
|
|
t.Skip("skipping, some registers such as vector registers are currently not loaded")
|
|
}
|
|
withTestTerminal("continuetestprog", t, func(term *FakeTerminal) {
|
|
r := term.MustExec("regs")
|
|
nr := len(strings.Split(r, "\n"))
|
|
t.Logf("regs: %s", r)
|
|
ra := term.MustExec("regs -a")
|
|
nra := len(strings.Split(ra, "\n"))
|
|
t.Logf("regs -a: %s", ra)
|
|
if nr > nra/2+1 {
|
|
t.Fatalf("'regs' returned too many registers (%d) compared to 'regs -a' (%d)", nr, nra)
|
|
}
|
|
})
|
|
}
|
|
|
|
func findStarFile(name string) string {
|
|
return filepath.Join(test.FindFixturesDir(), name+".star")
|
|
}
|
|
|
|
func TestIssue1598(t *testing.T) {
|
|
if buildMode == "pie" && runtime.GOARCH == "ppc64le" {
|
|
t.Skip("Debug function call Test broken in PIE mode")
|
|
}
|
|
test.MustSupportFunctionCalls(t, testBackend)
|
|
withTestTerminal("issue1598", t, func(term *FakeTerminal) {
|
|
term.MustExec("break issue1598.go:5")
|
|
term.MustExec("continue")
|
|
term.MustExec("config max-string-len 500")
|
|
r := term.MustExec("call x()")
|
|
t.Logf("result %q", r)
|
|
if !strings.Contains(r, "Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut \\nlabore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut") {
|
|
t.Fatalf("wrong value returned")
|
|
}
|
|
})
|
|
}
|
|
|
|
func TestExamineMemoryCmd(t *testing.T) {
|
|
withTestTerminal("examinememory", t, func(term *FakeTerminal) {
|
|
term.MustExec("break examinememory.go:19")
|
|
term.MustExec("break examinememory.go:24")
|
|
term.MustExec("continue")
|
|
|
|
addressStr := strings.TrimSpace(term.MustExec("p bspUintptr"))
|
|
address, err := strconv.ParseInt(addressStr, 0, 64)
|
|
if err != nil {
|
|
t.Fatalf("could convert %s into int64, err %s", addressStr, err)
|
|
}
|
|
|
|
res := term.MustExec("examinemem -count 51 -fmt hex " + addressStr)
|
|
t.Logf("the result of examining memory \n%s", res)
|
|
// check first line
|
|
firstLine := fmt.Sprintf("%#x: 0x0a 0x0b 0x0c 0x0d 0x0e 0x0f 0x10 0x11", address)
|
|
if !strings.Contains(res, firstLine) {
|
|
t.Fatalf("expected first line: %s", firstLine)
|
|
}
|
|
|
|
// check last line
|
|
lastLine := fmt.Sprintf("%#x: 0x3a 0x3b 0x3c", address+6*8)
|
|
if !strings.Contains(res, lastLine) {
|
|
t.Fatalf("expected last line: %s", lastLine)
|
|
}
|
|
|
|
// second examining memory
|
|
term.MustExec("continue")
|
|
res = term.MustExec("x -count 52 -fmt bin " + addressStr)
|
|
t.Logf("the second result of examining memory result \n%s", res)
|
|
|
|
// check first line
|
|
firstLine = fmt.Sprintf("%#x: 11111111 00001011 00001100 00001101", address)
|
|
if !strings.Contains(res, firstLine) {
|
|
t.Fatalf("expected first line: %s", firstLine)
|
|
}
|
|
|
|
// third examining memory: -x addr
|
|
res = term.MustExec("examinemem -x " + addressStr)
|
|
t.Logf("the third result of examining memory result \n%s", res)
|
|
firstLine = fmt.Sprintf("%#x: 0xff", address)
|
|
if !strings.Contains(res, firstLine) {
|
|
t.Fatalf("expected first line: %s", firstLine)
|
|
}
|
|
|
|
// fourth examining memory: -x addr + offset
|
|
res = term.MustExec("examinemem -x " + addressStr + " + 8")
|
|
t.Logf("the fourth result of examining memory result \n%s", res)
|
|
firstLine = fmt.Sprintf("%#x: 0x12", address+8)
|
|
if !strings.Contains(res, firstLine) {
|
|
t.Fatalf("expected first line: %s", firstLine)
|
|
}
|
|
// fifth examining memory: -x &var
|
|
res = term.MustExec("examinemem -x &bs[0]")
|
|
t.Logf("the fifth result of examining memory result \n%s", res)
|
|
firstLine = fmt.Sprintf("%#x: 0xff", address)
|
|
if !strings.Contains(res, firstLine) {
|
|
t.Fatalf("expected first line: %s", firstLine)
|
|
}
|
|
|
|
// sixth examining memory: -fmt and double spaces
|
|
res = term.MustExec("examinemem -fmt hex -x &bs[0]")
|
|
t.Logf("the sixth result of examining memory result \n%s", res)
|
|
firstLine = fmt.Sprintf("%#x: 0xff", address)
|
|
if !strings.Contains(res, firstLine) {
|
|
t.Fatalf("expected first line: %s", firstLine)
|
|
}
|
|
})
|
|
|
|
withTestTerminal("testvariables2", t, func(term *FakeTerminal) {
|
|
tests := []struct {
|
|
Expr string
|
|
Want int
|
|
}{
|
|
{Expr: "&i1", Want: 1},
|
|
{Expr: "&i2", Want: 2},
|
|
{Expr: "p1", Want: 1},
|
|
{Expr: "*pp1", Want: 1},
|
|
{Expr: "&str1[1]", Want: '1'},
|
|
{Expr: "c1.pb", Want: 1},
|
|
{Expr: "&c1.pb.a", Want: 1},
|
|
{Expr: "&c1.pb.a.A", Want: 1},
|
|
{Expr: "&c1.pb.a.B", Want: 2},
|
|
}
|
|
term.MustExec("continue")
|
|
for _, test := range tests {
|
|
res := term.MustExec("examinemem -fmt dec -x " + test.Expr)
|
|
// strip addr from output, e.g. "0xc0000160b8: 023" -> "023"
|
|
res = strings.TrimSpace(strings.Split(res, ":")[1])
|
|
got, err := strconv.Atoi(res)
|
|
if err != nil {
|
|
t.Fatalf("expr=%q err=%s", test.Expr, err)
|
|
} else if got != test.Want {
|
|
t.Errorf("expr=%q got=%d want=%d", test.Expr, got, test.Want)
|
|
}
|
|
}
|
|
})
|
|
}
|
|
|
|
func TestPrintOnTracepoint(t *testing.T) {
|
|
withTestTerminal("increment", t, func(term *FakeTerminal) {
|
|
term.MustExec("trace main.Increment")
|
|
term.MustExec("on 1 print y+1")
|
|
out, _ := term.Exec("continue")
|
|
if !strings.Contains(out, "y+1: 4") || !strings.Contains(out, "y+1: 2") || !strings.Contains(out, "y+1: 1") {
|
|
t.Errorf("output did not contain breakpoint information: %q", out)
|
|
}
|
|
})
|
|
}
|
|
|
|
func TestPrintCastToInterface(t *testing.T) {
|
|
withTestTerminal("testvariables2", t, func(term *FakeTerminal) {
|
|
term.MustExec("continue")
|
|
out := term.MustExec(`p (*"interface {}")(uintptr(&iface2))`)
|
|
t.Logf("%q", out)
|
|
})
|
|
}
|
|
|
|
func TestParseNewArgv(t *testing.T) {
|
|
testCases := []struct {
|
|
in string
|
|
tgtargs string
|
|
tgtredir string
|
|
tgterr string
|
|
}{
|
|
{"-noargs", "", " | | ", ""},
|
|
{"-noargs arg1", "", "", "too many arguments to restart"},
|
|
{"arg1 arg2", "arg1 | arg2", " | | ", ""},
|
|
{"arg1 arg2 <input.txt", "arg1 | arg2", "input.txt | | ", ""},
|
|
{"arg1 arg2 < input.txt", "arg1 | arg2", "input.txt | | ", ""},
|
|
{"<input.txt", "", "input.txt | | ", ""},
|
|
{"< input.txt", "", "input.txt | | ", ""},
|
|
{"arg1 < input.txt > output.txt 2> error.txt", "arg1", "input.txt | output.txt | error.txt", ""},
|
|
{"< input.txt > output.txt 2> error.txt", "", "input.txt | output.txt | error.txt", ""},
|
|
{"arg1 <input.txt >output.txt 2>error.txt", "arg1", "input.txt | output.txt | error.txt", ""},
|
|
{"<input.txt >output.txt 2>error.txt", "", "input.txt | output.txt | error.txt", ""},
|
|
{"<input.txt <input2.txt", "", "", "redirect error: stdin redirected twice"},
|
|
}
|
|
|
|
for _, tc := range testCases {
|
|
resetArgs, newArgv, newRedirects, err := parseNewArgv(tc.in)
|
|
t.Logf("%q -> %q %q %v\n", tc.in, newArgv, newRedirects, err)
|
|
if tc.tgterr != "" {
|
|
if err == nil {
|
|
t.Errorf("Expected error %q, got no error", tc.tgterr)
|
|
} else if errstr := err.Error(); errstr != tc.tgterr {
|
|
t.Errorf("Expected error %q, got error %q", tc.tgterr, errstr)
|
|
}
|
|
} else {
|
|
if !resetArgs {
|
|
t.Errorf("parse error, resetArgs is false")
|
|
continue
|
|
}
|
|
argvstr := strings.Join(newArgv, " | ")
|
|
if argvstr != tc.tgtargs {
|
|
t.Errorf("Expected new arguments %q, got %q", tc.tgtargs, argvstr)
|
|
}
|
|
redirstr := strings.Join(newRedirects[:], " | ")
|
|
if redirstr != tc.tgtredir {
|
|
t.Errorf("Expected new redirects %q, got %q", tc.tgtredir, redirstr)
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
func TestContinueUntil(t *testing.T) {
|
|
withTestTerminal("continuetestprog", t, func(term *FakeTerminal) {
|
|
if runtime.GOARCH != "386" {
|
|
listIsAt(t, term, "continue main.main", 16, -1, -1)
|
|
} else {
|
|
listIsAt(t, term, "continue main.main", 17, -1, -1)
|
|
}
|
|
listIsAt(t, term, "continue main.sayhi", 12, -1, -1)
|
|
})
|
|
}
|
|
|
|
func TestContinueUntilExistingBreakpoint(t *testing.T) {
|
|
withTestTerminal("continuetestprog", t, func(term *FakeTerminal) {
|
|
term.MustExec("break main.main")
|
|
if runtime.GOARCH != "386" {
|
|
listIsAt(t, term, "continue main.main", 16, -1, -1)
|
|
} else {
|
|
listIsAt(t, term, "continue main.main", 17, -1, -1)
|
|
}
|
|
listIsAt(t, term, "continue main.sayhi", 12, -1, -1)
|
|
})
|
|
}
|
|
|
|
func TestPrintFormat(t *testing.T) {
|
|
withTestTerminal("testvariables2", t, func(term *FakeTerminal) {
|
|
term.MustExec("continue")
|
|
out := term.MustExec("print %#x m2[1].B")
|
|
if !strings.Contains(out, "0xb\n") {
|
|
t.Fatalf("output did not contain '0xb': %q", out)
|
|
}
|
|
})
|
|
}
|
|
|
|
func TestHitCondBreakpoint(t *testing.T) {
|
|
withTestTerminal("break", t, func(term *FakeTerminal) {
|
|
term.MustExec("break bp1 main.main:4")
|
|
term.MustExec("condition -hitcount bp1 > 2")
|
|
listIsAt(t, term, "continue", 7, -1, -1)
|
|
out := term.MustExec("print i")
|
|
t.Logf("%q", out)
|
|
if !strings.Contains(out, "3\n") {
|
|
t.Fatalf("wrong value of i")
|
|
}
|
|
})
|
|
|
|
withTestTerminal("condperghitcount", t, func(term *FakeTerminal) {
|
|
term.MustExec("break bp1 main.main:8")
|
|
term.MustExec("condition -per-g-hitcount bp1 == 2")
|
|
listIsAt(t, term, "continue", 16, -1, -1)
|
|
// first g hit
|
|
out := term.MustExec("print j")
|
|
t.Logf("%q", out)
|
|
if !strings.Contains(out, "2\n") {
|
|
t.Fatalf("wrong value of j")
|
|
}
|
|
listIsAt(t, term, "continue", 16, -1, -1)
|
|
// second g hit
|
|
out = term.MustExec("print j")
|
|
t.Logf("%q", out)
|
|
if !strings.Contains(out, "2\n") {
|
|
t.Fatalf("wrong value of j")
|
|
}
|
|
})
|
|
}
|
|
|
|
func TestCondBreakpointWithFrame(t *testing.T) {
|
|
withTestTerminal("condframe", t, func(term *FakeTerminal) {
|
|
term.MustExec("break bp1 callme2")
|
|
term.MustExec("condition bp1 runtime.frame(1).i == 3")
|
|
term.MustExec("continue")
|
|
out := term.MustExec("frame 1 print i")
|
|
t.Logf("%q", out)
|
|
if !strings.Contains(out, "3\n") {
|
|
t.Fatalf("wrong value of i")
|
|
}
|
|
})
|
|
}
|
|
|
|
func TestClearCondBreakpoint(t *testing.T) {
|
|
withTestTerminal("break", t, func(term *FakeTerminal) {
|
|
term.MustExec("break main.main:4")
|
|
term.MustExec("condition 1 i%3==2")
|
|
listIsAt(t, term, "continue", 7, -1, -1)
|
|
out := term.MustExec("print i")
|
|
t.Logf("%q", out)
|
|
if !strings.Contains(out, "2\n") {
|
|
t.Fatalf("wrong value of i")
|
|
}
|
|
term.MustExec("condition -clear 1")
|
|
listIsAt(t, term, "continue", 7, -1, -1)
|
|
out = term.MustExec("print i")
|
|
t.Logf("%q", out)
|
|
if !strings.Contains(out, "3\n") {
|
|
t.Fatalf("wrong value of i")
|
|
}
|
|
})
|
|
}
|
|
|
|
func TestBreakpointEditing(t *testing.T) {
|
|
term := &FakeTerminal{
|
|
t: t,
|
|
Term: New(nil, &config.Config{}),
|
|
}
|
|
_ = term
|
|
|
|
var testCases = []struct {
|
|
inBp *api.Breakpoint
|
|
inBpStr string
|
|
edit string
|
|
outBp *api.Breakpoint
|
|
}{
|
|
{ // tracepoint -> breakpoint
|
|
&api.Breakpoint{Tracepoint: true},
|
|
"trace",
|
|
"",
|
|
&api.Breakpoint{}},
|
|
{ // breakpoint -> tracepoint
|
|
&api.Breakpoint{Variables: []string{"a"}},
|
|
"print a",
|
|
"print a\ntrace",
|
|
&api.Breakpoint{Tracepoint: true, Variables: []string{"a"}}},
|
|
{ // add print var
|
|
&api.Breakpoint{Variables: []string{"a"}},
|
|
"print a",
|
|
"print b\nprint a\n",
|
|
&api.Breakpoint{Variables: []string{"b", "a"}}},
|
|
{ // add goroutine flag
|
|
&api.Breakpoint{},
|
|
"",
|
|
"goroutine",
|
|
&api.Breakpoint{Goroutine: true}},
|
|
{ // remove goroutine flag
|
|
&api.Breakpoint{Goroutine: true},
|
|
"goroutine",
|
|
"",
|
|
&api.Breakpoint{}},
|
|
{ // add stack directive
|
|
&api.Breakpoint{},
|
|
"",
|
|
"stack 10",
|
|
&api.Breakpoint{Stacktrace: 10}},
|
|
{ // remove stack directive
|
|
&api.Breakpoint{Stacktrace: 20},
|
|
"stack 20",
|
|
"print a",
|
|
&api.Breakpoint{Variables: []string{"a"}}},
|
|
{ // add condition
|
|
&api.Breakpoint{Variables: []string{"a"}},
|
|
"print a",
|
|
"print a\ncond a < b",
|
|
&api.Breakpoint{Variables: []string{"a"}, Cond: "a < b"}},
|
|
{ // remove condition
|
|
&api.Breakpoint{Cond: "a < b"},
|
|
"cond a < b",
|
|
"",
|
|
&api.Breakpoint{}},
|
|
{ // change condition
|
|
&api.Breakpoint{Cond: "a < b"},
|
|
"cond a < b",
|
|
"cond a < 5",
|
|
&api.Breakpoint{Cond: "a < 5"}},
|
|
{ // change hitcount condition
|
|
&api.Breakpoint{HitCond: "% 2"},
|
|
"cond -hitcount % 2",
|
|
"cond -hitcount = 2",
|
|
&api.Breakpoint{HitCond: "= 2"}},
|
|
}
|
|
|
|
for _, tc := range testCases {
|
|
bp := *tc.inBp
|
|
bpStr := strings.Join(formatBreakpointAttrs("", &bp, true), "\n")
|
|
if bpStr != tc.inBpStr {
|
|
t.Errorf("Expected %q got %q for:\n%#v", tc.inBpStr, bpStr, tc.inBp)
|
|
}
|
|
ctx := callContext{Prefix: onPrefix, Scope: api.EvalScope{GoroutineID: -1, Frame: 0, DeferredCall: 0}, Breakpoint: &bp}
|
|
err := term.cmds.parseBreakpointAttrs(nil, ctx, strings.NewReader(tc.edit))
|
|
if err != nil {
|
|
t.Errorf("Unexpected error during edit %q", tc.edit)
|
|
}
|
|
if !reflect.DeepEqual(bp, *tc.outBp) {
|
|
t.Errorf("mismatch after edit\nexpected: %#v\ngot: %#v", tc.outBp, bp)
|
|
}
|
|
}
|
|
}
|
|
|
|
func TestTranscript(t *testing.T) {
|
|
withTestTerminal("math", t, func(term *FakeTerminal) {
|
|
term.MustExec("break main.main")
|
|
out := term.MustExec("continue")
|
|
if !strings.HasPrefix(out, "> [Breakpoint 1] main.main()") {
|
|
t.Fatalf("Wrong output for next: <%s>", out)
|
|
}
|
|
fh, err := os.CreateTemp("", "test-transcript-*")
|
|
if err != nil {
|
|
t.Fatalf("CreateTemp: %v", err)
|
|
}
|
|
name := fh.Name()
|
|
fh.Close()
|
|
t.Logf("output to %q", name)
|
|
|
|
slurp := func() string {
|
|
b, err := os.ReadFile(name)
|
|
if err != nil {
|
|
t.Fatalf("could not read transcript file: %v", err)
|
|
}
|
|
return string(b)
|
|
}
|
|
|
|
term.MustExec(fmt.Sprintf("transcript %s", name))
|
|
out = term.MustExec("list")
|
|
//term.MustExec("transcript -off")
|
|
if out != slurp() {
|
|
t.Logf("output of list %s", out)
|
|
t.Logf("contents of transcript: %s", slurp())
|
|
t.Errorf("transcript and command out differ")
|
|
}
|
|
|
|
term.MustExec(fmt.Sprintf("transcript -t -x %s", name))
|
|
out = term.MustExec(`print "hello"`)
|
|
if out != "" {
|
|
t.Errorf("output of print is %q but should have been suppressed by transcript", out)
|
|
}
|
|
if slurp() != "\"hello\"\n" {
|
|
t.Errorf("wrong contents of transcript: %q", slurp())
|
|
}
|
|
|
|
os.Remove(name)
|
|
})
|
|
}
|
|
|
|
func TestDisassPosCmd(t *testing.T) {
|
|
if runtime.GOARCH == "ppc64le" && buildMode == "pie" {
|
|
t.Skip("pie mode broken on ppc64le")
|
|
}
|
|
withTestTerminal("testvariables2", t, func(term *FakeTerminal) {
|
|
term.MustExec("continue")
|
|
out := term.MustExec("step-instruction")
|
|
t.Logf("%q\n", out)
|
|
if !strings.Contains(out, "call $runtime.Breakpoint") && !strings.Contains(out, "CALL runtime.Breakpoint(SB)") {
|
|
t.Errorf("output doesn't look like disassembly")
|
|
}
|
|
})
|
|
}
|
|
|
|
func TestCreateBreakpointByLocExpr(t *testing.T) {
|
|
withTestTerminal("math", t, func(term *FakeTerminal) {
|
|
out := term.MustExec("break main.main")
|
|
position1 := strings.Split(out, " set at ")[1]
|
|
term.MustExec("continue")
|
|
term.MustExec("clear 1")
|
|
out = term.MustExec("break +0")
|
|
position2 := strings.Split(out, " set at ")[1]
|
|
if position1 != position2 {
|
|
t.Fatalf("mismatched positions %q and %q\n", position1, position2)
|
|
}
|
|
})
|
|
}
|
|
|
|
func TestCreateBreakpointWithCondition(t *testing.T) {
|
|
withTestTerminal("break", t, func(term *FakeTerminal) {
|
|
term.MustExec("break bp1 main.main:4 if i == 3")
|
|
out := term.MustExec("breakpoints")
|
|
if !strings.Contains(out, "Breakpoint bp1") {
|
|
t.Fatal("incorrect breakpoint name")
|
|
}
|
|
listIsAt(t, term, "continue", 7, -1, -1)
|
|
out = term.MustExec("print i")
|
|
t.Logf("%q", out)
|
|
if !strings.Contains(out, "3\n") {
|
|
t.Fatalf("wrong value of i")
|
|
}
|
|
})
|
|
}
|
|
|
|
func TestCreateBreakpointWithCondition2(t *testing.T) {
|
|
withTestTerminal("break", t, func(term *FakeTerminal) {
|
|
term.MustExec("continue main.main:4")
|
|
term.MustExec("break if i == 3")
|
|
out := term.MustExec("breakpoints")
|
|
if strings.Contains(out, "Breakpoint if") {
|
|
t.Fatal("incorrect breakpoint name, should be ID")
|
|
}
|
|
listIsAt(t, term, "continue", 7, -1, -1)
|
|
out = term.MustExec("print i")
|
|
t.Logf("%q", out)
|
|
if !strings.Contains(out, "3\n") {
|
|
t.Fatalf("wrong value of i")
|
|
}
|
|
})
|
|
}
|
|
|
|
func TestCreateBreakpointWithCondition3(t *testing.T) {
|
|
withTestTerminal("test if path/main", t, func(term *FakeTerminal) {
|
|
// We should not attempt to parse this as a condition.
|
|
term.MustExec(`break _fixtures/test if path/main.go:4`)
|
|
listIsAt(t, term, "continue", 4, -1, -1)
|
|
})
|
|
}
|
|
|
|
func TestRestartBreakpoints(t *testing.T) {
|
|
// Tests that breakpoints set using just a line number and with a line
|
|
// offset are preserved after restart. See issue #3423.
|
|
withTestTerminal("continuetestprog", t, func(term *FakeTerminal) {
|
|
term.MustExec("break main.main")
|
|
term.MustExec("continue")
|
|
term.MustExec("break 9")
|
|
term.MustExec("break +1")
|
|
out := term.MustExec("breakpoints")
|
|
t.Log("breakpoints before:\n", out)
|
|
term.MustExec("restart")
|
|
out = term.MustExec("breakpoints")
|
|
t.Log("breakpoints after:\n", out)
|
|
bps, err := term.client.ListBreakpoints(false)
|
|
assertNoError(t, err, "ListBreakpoints")
|
|
for _, bp := range bps {
|
|
if bp.ID < 0 {
|
|
continue
|
|
}
|
|
if bp.Addr == 0 {
|
|
t.Fatalf("breakpoint %d has address 0", bp.ID)
|
|
}
|
|
}
|
|
})
|
|
}
|
|
|
|
func TestListPackages(t *testing.T) {
|
|
test.AllowRecording(t)
|
|
withTestTerminal("goroutinestackprog", t, func(term *FakeTerminal) {
|
|
out := term.MustExec("packages")
|
|
t.Logf("> packages\n%s", out)
|
|
seen := map[string]bool{}
|
|
for p := range strings.SplitSeq(strings.TrimSpace(out), "\n") {
|
|
seen[p] = true
|
|
}
|
|
if !seen["main"] || !seen["runtime"] {
|
|
t.Error("output omits 'main' and 'runtime'")
|
|
}
|
|
|
|
out = term.MustExec("packages runtime")
|
|
t.Logf("> packages runtime\n%s", out)
|
|
|
|
for p := range strings.SplitSeq(strings.TrimSpace(out), "\n") {
|
|
if !strings.Contains(p, "runtime") {
|
|
t.Errorf("output includes unexpected %q", p)
|
|
}
|
|
seen[p] = true
|
|
}
|
|
if !seen["runtime"] {
|
|
t.Error("output omits 'runtime'")
|
|
}
|
|
})
|
|
}
|
|
|
|
func TestSubstitutePathAndList(t *testing.T) {
|
|
// checks that substitute path rules do not remain cached after a -clear.
|
|
// See issue #3565.
|
|
if runtime.GOOS == "windows" {
|
|
t.Skip("test is not valid on windows due to path separators")
|
|
}
|
|
withTestTerminal("math", t, func(term *FakeTerminal) {
|
|
term.MustExec("break main.main")
|
|
term.MustExec("continue")
|
|
fixturesDir, _ := filepath.Abs(test.FindFixturesDir())
|
|
term.MustExec("config substitute-path " + fixturesDir + " /blah")
|
|
out, _ := term.Exec("list")
|
|
t.Logf("list output %s", out)
|
|
if !strings.Contains(out, "/blah/math.go") {
|
|
t.Fatalf("bad output")
|
|
}
|
|
term.MustExec("config substitute-path -clear")
|
|
term.MustExec("config substitute-path " + fixturesDir + " /blah2")
|
|
out, _ = term.Exec("list")
|
|
t.Logf("list output %s", out)
|
|
if !strings.Contains(out, "/blah2/math.go") {
|
|
t.Fatalf("bad output")
|
|
}
|
|
})
|
|
}
|
|
|
|
func TestDisplay(t *testing.T) {
|
|
// Tests that display command works. See issue #3595.
|
|
type testCase struct{ in, tgt string }
|
|
withTestTerminal("testvariables2", t, func(term *FakeTerminal) {
|
|
term.MustExec("continue")
|
|
|
|
for _, tc := range []testCase{
|
|
{"string(byteslice)", `0: string(byteslice) = "tèst"`},
|
|
{"string(byteslice[1:])", `0: string(byteslice[1:]) = "èst"`},
|
|
{"%s string(byteslice)", `0: string(byteslice) = tèst`},
|
|
} {
|
|
out := term.MustExec("display -a " + tc.in)
|
|
t.Logf("%q -> %q", tc.in, out)
|
|
if !strings.Contains(out, tc.tgt) {
|
|
t.Errorf("wrong output for 'display -a %s':\n\tgot: %q\n\texpected: %q", tc.in, out, tc.tgt)
|
|
}
|
|
term.MustExec("display -d 0")
|
|
}
|
|
})
|
|
}
|
|
|
|
// compareBreakpoints compares two lists of breakpoints, considering only
|
|
// the fields that are preserved during save/restore. It ignores runtime state like
|
|
// IDs, names, addresses, and hit counts.
|
|
func compareBreakpoints(t *testing.T, before, after []*api.Breakpoint) {
|
|
t.Helper()
|
|
|
|
// Sort both lists by File:Line for stable comparison
|
|
// We mainly need this function for Tracepoints.
|
|
sortBps := func(bps []*api.Breakpoint) []*api.Breakpoint {
|
|
sorted := make([]*api.Breakpoint, len(bps))
|
|
copy(sorted, bps)
|
|
slices.SortFunc(sorted, func(a, b *api.Breakpoint) int {
|
|
if a.File != b.File {
|
|
return strings.Compare(a.File, b.File)
|
|
}
|
|
return cmp.Compare(a.Line, b.Line)
|
|
})
|
|
return sorted
|
|
}
|
|
|
|
beforeSorted := sortBps(before)
|
|
afterSorted := sortBps(after)
|
|
|
|
if len(beforeSorted) != len(afterSorted) {
|
|
t.Errorf("Breakpoint count mismatch: before=%d, after=%d", len(beforeSorted), len(afterSorted))
|
|
return
|
|
}
|
|
|
|
for i := range beforeSorted {
|
|
b := beforeSorted[i]
|
|
a := afterSorted[i]
|
|
|
|
if b.File != a.File {
|
|
t.Errorf("Breakpoint %d: File mismatch: before=%q, after=%q", i, b.File, a.File)
|
|
}
|
|
if b.Line != a.Line {
|
|
t.Errorf("Breakpoint %d: Line mismatch: before=%d, after=%d", i, b.Line, a.Line)
|
|
}
|
|
|
|
if b.Tracepoint != a.Tracepoint {
|
|
t.Errorf("Breakpoint %d: Tracepoint mismatch: before=%v, after=%v", i, b.Tracepoint, a.Tracepoint)
|
|
}
|
|
if b.Cond != a.Cond {
|
|
t.Errorf("Breakpoint %d: Cond mismatch: before=%q, after=%q", i, b.Cond, a.Cond)
|
|
}
|
|
if b.HitCond != a.HitCond {
|
|
t.Errorf("Breakpoint %d: HitCond mismatch: before=%q, after=%q", i, b.HitCond, a.HitCond)
|
|
}
|
|
if b.HitCondPerG != a.HitCondPerG {
|
|
t.Errorf("Breakpoint %d: HitCondPerG mismatch: before=%v, after=%v", i, b.HitCondPerG, a.HitCondPerG)
|
|
}
|
|
if b.Disabled != a.Disabled {
|
|
t.Errorf("Breakpoint %d: Disabled mismatch: before=%v, after=%v", i, b.Disabled, a.Disabled)
|
|
}
|
|
}
|
|
}
|
|
|
|
func TestBreakpointSave(t *testing.T) {
|
|
test.AllowRecording(t)
|
|
tests := []struct {
|
|
name string
|
|
breakpointCmds []string
|
|
}{
|
|
{
|
|
name: "basic breakpoint save",
|
|
breakpointCmds: []string{
|
|
"break main.main",
|
|
"break _fixtures/break.go:4",
|
|
},
|
|
},
|
|
{
|
|
name: "breakpoint with condition",
|
|
breakpointCmds: []string{
|
|
"break main.main:4 if i == 3",
|
|
},
|
|
},
|
|
{
|
|
name: "disabled breakpoint",
|
|
breakpointCmds: []string{
|
|
"break main main.main",
|
|
"toggle main",
|
|
},
|
|
},
|
|
{
|
|
name: "tracepoint",
|
|
breakpointCmds: []string{
|
|
"trace main.main",
|
|
},
|
|
},
|
|
{
|
|
name: "empty breakpoint list",
|
|
breakpointCmds: []string{},
|
|
},
|
|
}
|
|
for _, tc := range tests {
|
|
t.Run(tc.name, func(t *testing.T) {
|
|
withTestTerminal("break", t, func(term *FakeTerminal) {
|
|
for _, cmd := range tc.breakpointCmds {
|
|
term.MustExec(cmd)
|
|
}
|
|
|
|
// Capture breakpoints before save
|
|
bpsBefore, err := term.client.ListBreakpoints(false)
|
|
assertNoError(t, err, "ListBreakpoints before save")
|
|
|
|
f, err := os.CreateTemp("", "test*.txt")
|
|
if err != nil {
|
|
t.Fatalf("Failed to create temp file: %v", err)
|
|
}
|
|
f.Close()
|
|
defer os.Remove(f.Name())
|
|
|
|
term.MustExec("breakpoints -save " + f.Name())
|
|
term.MustExec("clearall")
|
|
term.MustExec("source " + f.Name())
|
|
|
|
// Capture breakpoints after restore
|
|
bpsAfter, err := term.client.ListBreakpoints(false)
|
|
assertNoError(t, err, "ListBreakpoints after restore")
|
|
|
|
// Compare breakpoints
|
|
compareBreakpoints(t, bpsBefore, bpsAfter)
|
|
})
|
|
})
|
|
}
|
|
|
|
// Test case for detecting modified/corrupted save files. It is the same
|
|
// idea as in the other tests but we add an additional breakpoint after
|
|
// sourcing the file
|
|
t.Run("detect additional breakpoint", func(t *testing.T) {
|
|
withTestTerminal("break", t, func(term *FakeTerminal) {
|
|
term.MustExec("break main.main")
|
|
term.MustExec("break _fixtures/break.go:4")
|
|
|
|
bpsBefore, err := term.client.ListBreakpoints(false)
|
|
assertNoError(t, err, "ListBreakpoints before save")
|
|
|
|
f, err := os.CreateTemp("", "test*.txt")
|
|
if err != nil {
|
|
t.Fatalf("Failed to create temp file: %v", err)
|
|
}
|
|
f.Close()
|
|
defer os.Remove(f.Name())
|
|
|
|
term.MustExec("breakpoints -save " + f.Name())
|
|
term.MustExec("clearall")
|
|
term.MustExec("source " + f.Name())
|
|
|
|
// Inject an extra breakpoint after sourcing
|
|
term.MustExec("break _fixtures/break.go:6")
|
|
|
|
bpsAfter, err := term.client.ListBreakpoints(false)
|
|
assertNoError(t, err, "ListBreakpoints after restore")
|
|
|
|
// Verify that we detected the injected breakpoint
|
|
// We expect 3 breakpoints (2 original + 1 injected) instead of 2
|
|
if len(bpsBefore) == len(bpsAfter) {
|
|
t.Errorf("Failed to detect injected breakpoint: before=%d, after=%d", len(bpsBefore), len(bpsAfter))
|
|
}
|
|
if len(bpsAfter) != len(bpsBefore)+1 {
|
|
t.Errorf("Unexpected breakpoint count after injection: expected=%d, got=%d", len(bpsBefore)+1, len(bpsAfter))
|
|
}
|
|
})
|
|
})
|
|
}
|
|
|
|
func TestBreakPointFailWithCond(t *testing.T) {
|
|
if runtime.GOOS == "freebsd" || runtime.GOOS == "darwin" {
|
|
t.Skip("follow exec not implemented")
|
|
}
|
|
|
|
oldYesNo := yesno
|
|
defer func() { yesno = oldYesNo }()
|
|
// always answer yes here
|
|
yesno = func(line *liner.State, question, defaultAnswer string) (bool, error) {
|
|
return true, nil
|
|
}
|
|
|
|
withTestTerminal("spawn", t, func(term *FakeTerminal) {
|
|
assertNoError(t, term.client.FollowExec(true, ""), "FollowExec")
|
|
_, err := term.Exec("break spawnchild.go:11 if i == 1")
|
|
if err != nil {
|
|
t.Fatalf("expect to set a suspended breakpoint: %v", err)
|
|
}
|
|
bp, _ := term.client.GetBreakpoint(1)
|
|
if bp.Cond != "i == 1" {
|
|
t.Errorf("expected condition to be 'i == 1', got %s", bp.Cond)
|
|
}
|
|
})
|
|
}
|
|
|
|
func TestTraceRegexpReturn(t *testing.T) {
|
|
withTestTerminal("traceret", t, func(term *FakeTerminal) {
|
|
out, err := term.Exec(`trace /main\.fncall./`)
|
|
if err != nil {
|
|
t.Errorf("error executing trace command: %v", err)
|
|
}
|
|
out, _ = term.Exec("continue")
|
|
t.Logf("continue: %q", out)
|
|
if out != "> goroutine(1): main.fncall1()\n>> goroutine(1): main.fncall1 => (1)\n> goroutine(1): main.fncall2()\n>> goroutine(1): main.fncall2 => (2)\n" {
|
|
t.Errorf("wrong output for continue")
|
|
}
|
|
})
|
|
}
|
|
|
|
func TestStarlarkOnPrefix(t *testing.T) {
|
|
withTestTerminal("testvariables2", t, func(term *FakeTerminal) {
|
|
term.MustExec("source " + findStarFile("test_allow_on"))
|
|
|
|
var customCmd *command
|
|
for i := range term.cmds.cmds {
|
|
cmd := &term.cmds.cmds[i]
|
|
if cmd.match("test_on_allowed") {
|
|
customCmd = cmd
|
|
break
|
|
}
|
|
}
|
|
|
|
if customCmd == nil {
|
|
t.Fatal("test_on_allowed command not found")
|
|
}
|
|
|
|
// All custom starlark commands should have onPrefix set
|
|
if customCmd.allowedPrefixes&onPrefix == 0 {
|
|
t.Errorf("custom starlark command should have onPrefix set, got allowedPrefixes=%v", customCmd.allowedPrefixes)
|
|
}
|
|
|
|
term.MustExec("break main.main")
|
|
|
|
bps1, _ := term.client.ListBreakpoints(false)
|
|
var targetBpID int
|
|
for _, bp := range bps1 {
|
|
if bp.FunctionName == "main.main" {
|
|
targetBpID = bp.ID
|
|
break
|
|
}
|
|
}
|
|
if targetBpID == 0 {
|
|
t.Fatal("Could not find breakpoint on main.main")
|
|
}
|
|
|
|
term.MustExec(fmt.Sprintf("on %d test_on_allowed", targetBpID))
|
|
|
|
bpListOutput := term.MustExec("breakpoints")
|
|
if !strings.Contains(bpListOutput, "test_on_allowed") {
|
|
t.Errorf("expected 'test_on_allowed' to appear in breakpoints output, got: %q", bpListOutput)
|
|
}
|
|
|
|
breakpoints, err := term.client.ListBreakpoints(false)
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
if len(breakpoints) == 0 {
|
|
t.Fatal("no breakpoints found")
|
|
}
|
|
|
|
var bp *api.Breakpoint
|
|
for _, b := range breakpoints {
|
|
if b.ID == targetBpID {
|
|
bp = b
|
|
break
|
|
}
|
|
}
|
|
if bp == nil {
|
|
t.Fatalf("Could not find breakpoint with ID %d", targetBpID)
|
|
}
|
|
|
|
if len(bp.CustomCommands) == 0 {
|
|
t.Errorf("expected custom command to be registered on breakpoint, but CustomCommands is empty")
|
|
} else {
|
|
found := slices.Contains(bp.CustomCommands, "test_on_allowed")
|
|
if !found {
|
|
t.Errorf("expected 'test_on_allowed' to be in CustomCommands, got %v", bp.CustomCommands)
|
|
}
|
|
}
|
|
|
|
term.MustExec(fmt.Sprintf("on %d test_on_allowed hello world", targetBpID))
|
|
|
|
bpListOutput2 := term.MustExec("breakpoints")
|
|
if !strings.Contains(bpListOutput2, "test_on_allowed hello world") {
|
|
t.Errorf("expected 'test_on_allowed hello world' to appear in breakpoints output, got: %q", bpListOutput2)
|
|
}
|
|
|
|
out := term.MustExec("continue")
|
|
if !strings.Contains(out, "test_on_allowed called with:") {
|
|
t.Errorf("expected custom command to be executed when breakpoint is hit, but output was: %q", out)
|
|
}
|
|
if !strings.Contains(out, "test_on_allowed called with: hello world") {
|
|
t.Errorf("expected custom command with args to be executed, but output was: %q", out)
|
|
}
|
|
})
|
|
}
|
|
|
|
func TestCustomCommandStopsOnContinue(t *testing.T) {
|
|
// This test verifies that when a custom command triggers a runCmd (like continue),
|
|
// the execution of remaining custom commands is stopped because the process state
|
|
// has been invalidated.
|
|
withTestTerminal("break", t, func(term *FakeTerminal) {
|
|
term.MustExec("source " + findStarFile("test_custom_cmd_continue"))
|
|
|
|
term.MustExec("break break.go:7")
|
|
|
|
bps, _ := term.client.ListBreakpoints(false)
|
|
var targetBpID int
|
|
for _, bp := range bps {
|
|
if strings.Contains(bp.File, "break.go") && bp.Line == 7 {
|
|
targetBpID = bp.ID
|
|
break
|
|
}
|
|
}
|
|
if targetBpID == 0 {
|
|
t.Fatal("Could not find breakpoint on break.go:7")
|
|
}
|
|
|
|
// Add three custom commands to the breakpoint:
|
|
// 1. cmd before continue (should execute)
|
|
// 2. continue (should execute and invalidate state)
|
|
// 3. cmd after continue (should NOT execute because state is invalid)
|
|
term.MustExec(fmt.Sprintf("on %d test_cmd_before_continue", targetBpID))
|
|
term.MustExec(fmt.Sprintf("on %d test_continue_cmd", targetBpID))
|
|
term.MustExec(fmt.Sprintf("on %d test_cmd_after_continue", targetBpID))
|
|
|
|
out := term.MustExec("continue")
|
|
|
|
if !strings.Contains(out, "BEFORE_CONTINUE") {
|
|
t.Errorf("expected BEFORE_CONTINUE to be printed, got: %q", out)
|
|
}
|
|
|
|
if !strings.Contains(out, "CONTINUE_CMD") {
|
|
t.Errorf("expected CONTINUE_CMD to be printed, got: %q", out)
|
|
}
|
|
|
|
if strings.Contains(out, "AFTER_CONTINUE") {
|
|
t.Errorf("AFTER_CONTINUE was printed, which means the command ran after continue, got: %q", out)
|
|
}
|
|
})
|
|
}
|
|
|
|
func TestCommandPromptExpansion(t *testing.T) {
|
|
withTestTerminal("testvariables2", t, func(term *FakeTerminal) {
|
|
term.MustExec("continue")
|
|
for _, e := range []struct{ in, tgt string }{
|
|
{"$u", "<malformed escape>"},
|
|
{"malformed at end $", "malformed at end <malformed escape>"},
|
|
{"$g", "1"},
|
|
{"$g:$f>", "1:0>"},
|
|
{"$$g", "$g"},
|
|
} {
|
|
out := term.expandPrompt(e.in)
|
|
if out != e.tgt {
|
|
t.Errorf("input %q expected %q got %q\n", e.in, e.tgt, out)
|
|
}
|
|
}
|
|
})
|
|
}
|