scan error formatting (#5628)

pkg/scan/scan_context.go

@@ -8,6 +8,7 @@ import (
 
 	"github.com/projectdiscovery/nuclei/v3/pkg/output"
 	"github.com/projectdiscovery/nuclei/v3/pkg/protocols/common/contextargs"
+	"github.com/projectdiscovery/utils/errkit"
 )
 
 type ScanContextOption func(*ScanContext)
@@ -30,7 +31,7 @@ type ScanContext struct {
 	OnWarning func(string)
 
 	// unexported state fields
-	errors   []error
+	error    error
 	warnings []string
 	events   []*output.InternalWrappedEvent
 	results  []*output.ResultEvent
@@ -52,8 +53,8 @@ func (s *ScanContext) Context() context.Context {
 	return s.ctx
 }
 
-func (s *ScanContext) GenerateErrorMessage() string {
-	return joinErrors(s.errors)
+func (s *ScanContext) GenerateErrorMessage() error {
+	return s.error
 }
 
 // GenerateResult returns final results slice from all events
@@ -94,13 +95,16 @@ func (s *ScanContext) LogError(err error) {
 	if err == nil {
 		return
 	}
-
 	if s.OnError != nil {
 		s.OnError(err)
 	}
-	s.errors = append(s.errors, err)
+	if s.error == nil {
+		s.error = err
+	} else {
+		s.error = errkit.Append(s.error, err)
+	}
 
-	errorMessage := s.GenerateErrorMessage()
+	errorMessage := s.GenerateErrorMessage().Error()
 
 	for _, result := range s.results {
 		result.Error = errorMessage
@@ -129,14 +133,3 @@ func (s *ScanContext) LogWarning(format string, args ...any) {
 		}
 	}
 }
-
-// joinErrors joins multiple errors and returns a single error string
-func joinErrors(errors []error) string {
-	var errorMessages []string
-	for _, e := range errors {
-		if e != nil {
-			errorMessages = append(errorMessages, e.Error())
-		}
-	}
-	return strings.Join(errorMessages, "; ")
-}

pkg/tmplexec/exec.go

@@ -20,7 +20,6 @@ import (
 	"github.com/projectdiscovery/nuclei/v3/pkg/tmplexec/flow"
 	"github.com/projectdiscovery/nuclei/v3/pkg/tmplexec/generic"
 	"github.com/projectdiscovery/nuclei/v3/pkg/tmplexec/multiproto"
-	"github.com/projectdiscovery/nuclei/v3/pkg/types/nucleierr"
 	"github.com/projectdiscovery/utils/errkit"
 )
 
@@ -207,7 +206,7 @@ func (e *TemplateExecuter) Execute(ctx *scan.ScanContext) (bool, error) {
 	ctx.LogError(errx)
 
 	if lastMatcherEvent != nil {
-		lastMatcherEvent.InternalEvent["error"] = tryParseCause(fmt.Errorf("%s", ctx.GenerateErrorMessage()))
+		lastMatcherEvent.InternalEvent["error"] = getErrorCause(ctx.GenerateErrorMessage())
 		writeFailureCallback(lastMatcherEvent, e.options.Options.MatcherStatus)
 	}
 
@@ -222,7 +221,7 @@ func (e *TemplateExecuter) Execute(ctx *scan.ScanContext) (bool, error) {
 					Info:       e.options.TemplateInfo,
 					Type:       e.getTemplateType(),
 					Host:       ctx.Input.MetaInput.Input,
-					Error:      tryParseCause(fmt.Errorf("%s", ctx.GenerateErrorMessage())),
+					Error:      getErrorCause(ctx.GenerateErrorMessage()),
 				},
 			},
 			OperatorsResult: &operators.Result{
@@ -235,31 +234,27 @@ func (e *TemplateExecuter) Execute(ctx *scan.ScanContext) (bool, error) {
 	return executed.Load() || matched.Load(), errx
 }
 
-// tryParseCause tries to parse the cause of given error
+// getErrorCause tries to parse the cause of given error
 // this is legacy support due to use of errorutil in existing libraries
 // but this should not be required once all libraries are updated
-func tryParseCause(err error) string {
-	errStr := ""
-	errX := errkit.FromError(err)
-	if errX != nil {
-		var errCause error
-
-		if len(errX.Errors()) > 1 {
-			errCause = errX.Errors()[0]
+func getErrorCause(err error) string {
+	if err == nil {
+		return ""
 	}
-		if errCause == nil {
-			errCause = errX
+	errx := errkit.FromError(err)
+	var cause error
+	for _, e := range errx.Errors() {
+		if e != nil && strings.Contains(e.Error(), "context deadline exceeded") {
+			continue
 		}
-
-		msg := strings.Trim(errCause.Error(), "{} ")
-		parts := strings.Split(msg, ":")
-		errCause = errkit.New("%s", parts[len(parts)-1])
-		errKind := errkit.GetErrorKind(err, nucleierr.ErrTemplateLogic).String()
-		errStr = errCause.Error()
-		errStr = strings.TrimSpace(strings.Replace(errStr, "errKind="+errKind, "", -1))
+		cause = e
+		break
 	}
-
-	return errStr
+	if cause == nil {
+		cause = errkit.Append(errkit.New("could not get error cause"), errx)
+	}
+	// parseScanError prettifies the error message and removes everything except the cause
+	return parseScanError(cause.Error())
 }
 
 // ExecuteWithResults executes the protocol requests and returns results instead of writing them.

pkg/tmplexec/interface.go

@@ -1,10 +1,15 @@
 package tmplexec
 
 import (
+	"errors"
+	"regexp"
+	"strings"
+
 	"github.com/projectdiscovery/nuclei/v3/pkg/scan"
 	"github.com/projectdiscovery/nuclei/v3/pkg/tmplexec/flow"
 	"github.com/projectdiscovery/nuclei/v3/pkg/tmplexec/generic"
 	"github.com/projectdiscovery/nuclei/v3/pkg/tmplexec/multiproto"
+	"github.com/projectdiscovery/utils/errkit"
 )
 
 var (
@@ -30,3 +35,37 @@ type TemplateEngine interface {
 	// Name returns name of template engine
 	Name() string
 }
+
+var (
+	// A temporary fix to remove errKind from error message
+	// this is because errkit is not used everywhere yet
+	reNoKind = regexp.MustCompile(`([\[][^][]+[\]]|errKind=[^ ]+) `)
+)
+
+// parseScanError parses given scan error and only returning the cause
+// instead of inefficient one
+func parseScanError(msg string) string {
+	if msg == "" {
+		return ""
+	}
+	if strings.HasPrefix(msg, "ReadStatusLine:") {
+		// last index is actual error (from rawhttp)
+		parts := strings.Split(msg, ":")
+		msg = strings.TrimSpace(parts[len(parts)-1])
+	}
+	if strings.Contains(msg, "read ") {
+		// same here
+		parts := strings.Split(msg, ":")
+		msg = strings.TrimSpace(parts[len(parts)-1])
+	}
+	e := errkit.FromError(errors.New(msg))
+	for _, err := range e.Errors() {
+		if err != nil && strings.Contains(err.Error(), "context deadline exceeded") {
+			continue
+		}
+		msg = reNoKind.ReplaceAllString(err.Error(), "")
+		return msg
+	}
+	wrapped := errkit.Append(errkit.New("failed to get error cause"), e).Error()
+	return reNoKind.ReplaceAllString(wrapped, "")
+}