Allow to get source position of unmarshal errors (enhanced version)

decode.go

@@ -22,7 +22,6 @@ import (
 	"io"
 	"math"
 	"reflect"
-	"strconv"
 	"time"
 )
 
@@ -110,7 +109,6 @@ func (p *parser) peek() yaml_event_type_t {
 }
 
 func (p *parser) fail() {
-	var where string
 	var line int
 	if p.parser.context_mark.line != 0 {
 		line = p.parser.context_mark.line
@@ -125,16 +123,13 @@ func (p *parser) fail() {
 			line++
 		}
 	}
-	if line != 0 {
-		where = "line " + strconv.Itoa(line) + ": "
-	}
 	var msg string
 	if len(p.parser.problem) > 0 {
 		msg = p.parser.problem
 	} else {
 		msg = "unknown problem parsing YAML content"
 	}
-	failf("%s%s", where, msg)
+	fail(&ParserError{msg, line})
 }
 
 func (p *parser) anchor(n *Node, anchor []byte) {
@@ -313,7 +308,7 @@ func (p *parser) mapping() *Node {
 type decoder struct {
 	doc     *Node
 	aliases map[*Node]bool
-	terrors []string
+	terrors []UnmarshalError
 
 	stringMapType  reflect.Type
 	generalMapType reflect.Type
@@ -359,7 +354,11 @@ func (d *decoder) terror(n *Node, tag string, out reflect.Value) {
 			value = " `" + value + "`"
 		}
 	}
-	d.terrors = append(d.terrors, fmt.Sprintf("line %d: cannot unmarshal %s%s into %s", n.Line, shortTag(tag), value, out.Type()))
+	d.terrors = append(d.terrors, UnmarshalError{
+		Message: fmt.Sprintf("cannot unmarshal %s%s into %s", shortTag(tag), value, out.Type()),
+		Line:    n.Line,
+		Column:  n.Column,
+	})
 }
 
 func (d *decoder) callUnmarshaler(n *Node, u Unmarshaler) (good bool) {
@@ -773,7 +772,11 @@ func (d *decoder) mapping(n *Node, out reflect.Value) (good bool) {
 			for j := i + 2; j < l; j += 2 {
 				nj := n.Content[j]
 				if ni.Kind == nj.Kind && ni.Value == nj.Value {
-					d.terrors = append(d.terrors, fmt.Sprintf("line %d: mapping key %#v already defined at line %d", nj.Line, nj.Value, ni.Line))
+					d.terrors = append(d.terrors, UnmarshalError{
+						Message: fmt.Sprintf("mapping key %#v already defined at line %d", nj.Value, ni.Line),
+						Line:    nj.Line,
+						Column:  nj.Column,
+					})
 				}
 			}
 		}
@@ -921,7 +924,11 @@ func (d *decoder) mappingStruct(n *Node, out reflect.Value) (good bool) {
 		if info, ok := sinfo.FieldsMap[sname]; ok {
 			if d.uniqueKeys {
 				if doneFields[info.Id] {
-					d.terrors = append(d.terrors, fmt.Sprintf("line %d: field %s already set in type %s", ni.Line, name.String(), out.Type()))
+					d.terrors = append(d.terrors, UnmarshalError{
+						Message: fmt.Sprintf("field %s already set in type %s", name.String(), out.Type()),
+						Line:    ni.Line,
+						Column:  ni.Column,
+					})
 					continue
 				}
 				doneFields[info.Id] = true
@@ -941,7 +948,11 @@ func (d *decoder) mappingStruct(n *Node, out reflect.Value) (good bool) {
 			d.unmarshal(n.Content[i+1], value)
 			inlineMap.SetMapIndex(name, value)
 		} else if d.knownFields {
-			d.terrors = append(d.terrors, fmt.Sprintf("line %d: field %s not found in type %s", ni.Line, name.String(), out.Type()))
+			d.terrors = append(d.terrors, UnmarshalError{
+				Message: fmt.Sprintf("field %s not found in type %s", name.String(), out.Type()),
+				Line:    ni.Line,
+				Column:  ni.Column,
+			})
 		}
 	}
 

decode_test.go

@@ -980,6 +980,18 @@ func (s *S) TestDecoderErrors(c *C) {
 	}
 }
 
+func (s *S) TestParserError(c *C) {
+	var v struct {
+		A, B int
+	}
+	data := "a: 1\n=\nb: 2"
+	err := yaml.Unmarshal([]byte(data), &v)
+	c.Assert(err, DeepEquals, &yaml.ParserError{
+		Message: "could not find expected ':'",
+		Line:    2,
+	})
+}
+
 var unmarshalerTests = []struct {
 	data, tag string
 	value     interface{}
@@ -1112,8 +1124,8 @@ func (s *S) TestUnmarshalerWholeDocument(c *C) {
 }
 
 func (s *S) TestUnmarshalerTypeError(c *C) {
-	unmarshalerResult[2] = &yaml.TypeError{[]string{"foo"}}
-	unmarshalerResult[4] = &yaml.TypeError{[]string{"bar"}}
+	unmarshalerResult[2] = &yaml.TypeError{[]yaml.UnmarshalError{{"foo", 1, 1}}}
+	unmarshalerResult[4] = &yaml.TypeError{[]yaml.UnmarshalError{{"bar", 1, 1}}}
 	defer func() {
 		delete(unmarshalerResult, 2)
 		delete(unmarshalerResult, 4)
@@ -1130,8 +1142,8 @@ func (s *S) TestUnmarshalerTypeError(c *C) {
 	c.Assert(err, ErrorMatches, ""+
 		"yaml: unmarshal errors:\n"+
 		"  line 1: cannot unmarshal !!str `A` into int\n"+
-		"  foo\n"+
-		"  bar\n"+
+		"  line 1: foo\n"+
+		"  line 1: bar\n"+
 		"  line 1: cannot unmarshal !!str `B` into int")
 	c.Assert(v.M["abc"], NotNil)
 	c.Assert(v.M["def"], IsNil)
@@ -1143,8 +1155,8 @@ func (s *S) TestUnmarshalerTypeError(c *C) {
 }
 
 func (s *S) TestObsoleteUnmarshalerTypeError(c *C) {
-	unmarshalerResult[2] = &yaml.TypeError{[]string{"foo"}}
-	unmarshalerResult[4] = &yaml.TypeError{[]string{"bar"}}
+	unmarshalerResult[2] = &yaml.TypeError{[]yaml.UnmarshalError{{"foo", 1, 1}}}
+	unmarshalerResult[4] = &yaml.TypeError{[]yaml.UnmarshalError{{"bar", 1, 1}}}
 	defer func() {
 		delete(unmarshalerResult, 2)
 		delete(unmarshalerResult, 4)
@@ -1161,8 +1173,8 @@ func (s *S) TestObsoleteUnmarshalerTypeError(c *C) {
 	c.Assert(err, ErrorMatches, ""+
 		"yaml: unmarshal errors:\n"+
 		"  line 1: cannot unmarshal !!str `A` into int\n"+
-		"  foo\n"+
-		"  bar\n"+
+		"  line 1: foo\n"+
+		"  line 1: bar\n"+
 		"  line 1: cannot unmarshal !!str `B` into int")
 	c.Assert(v.M["abc"], NotNil)
 	c.Assert(v.M["def"], IsNil)

yaml.go

@@ -26,6 +26,7 @@ import (
 	"fmt"
 	"io"
 	"reflect"
+	"strconv"
 	"strings"
 	"sync"
 	"unicode/utf8"
@@ -288,7 +289,7 @@ func (e *Encoder) Close() (err error) {
 
 func handleErr(err *error) {
 	if v := recover(); v != nil {
-		if e, ok := v.(yamlError); ok {
+		if e, ok := v.(*yamlError); ok {
 			*err = e.err
 		} else {
 			panic(v)
@@ -300,24 +301,69 @@ type yamlError struct {
 	err error
 }
 
+func (e *yamlError) Unwrap() error {
+	return e.err
+}
+
 func fail(err error) {
-	panic(yamlError{err})
+	panic(&yamlError{err})
 }
 
 func failf(format string, args ...interface{}) {
-	panic(yamlError{fmt.Errorf("yaml: "+format, args...)})
+	panic(&yamlError{fmt.Errorf("yaml: "+format, args...)})
+}
+
+// ParserError is each error with a source line position found by the parser.
+// Unlike UnmarshalErrors, it is only used for terminal failures.
+type ParserError struct {
+	Message string
+	Line    int
+}
+
+func (e *ParserError) Error() string {
+	var b strings.Builder
+	b.WriteString("yaml: ")
+	if e.Line != 0 {
+		b.WriteString("line " + strconv.Itoa(e.Line) + ": ")
+	}
+	b.WriteString(e.Message)
+	return b.String()
+}
+
+// UnmarshalError is each error with a source position found by Unmarshal.
+type UnmarshalError struct {
+	Message string
+	Line    int
+	Column  int
+}
+
+func (e UnmarshalError) Error() string {
+	return fmt.Sprintf("line %d: %s", e.Line, e.Message)
 }
 
 // A TypeError is returned by Unmarshal when one or more fields in
 // the YAML document cannot be properly decoded into the requested
 // types. When this error is returned, the value is still
 // unmarshaled partially.
 type TypeError struct {
-	Errors []string
+	Errors []UnmarshalError
 }
 
 func (e *TypeError) Error() string {
-	return fmt.Sprintf("yaml: unmarshal errors:\n  %s", strings.Join(e.Errors, "\n  "))
+	var b strings.Builder
+	b.WriteString("yaml: unmarshal errors:")
+	for _, err := range e.Errors {
+		b.WriteString("\n  " + err.Error())
+	}
+	return b.String()
+}
+
+func (e *TypeError) Unwrap() []error {
+	errs := make([]error, 0, len(e.Errors))
+	for _, err := range e.Errors {
+		errs = append(errs, err)
+	}
+	return errs
 }
 
 type Kind uint32