feat: carry the byte offset and column in SyntaxError

Assisted-by: GLM 5.3 Flash

CHANGELOG.md

@@ -51,6 +51,12 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
   failure alike. `Valid(data)` reports whether a document parses, nil on
   success and the parse error on failure, the library call the `-validate`
   mode of interpres-decode is built on.
+- `SyntaxError` carries the byte `Offset` the scan stopped at and the 1-based
+  `Column` on the line, beside the line it always had, and `SourceLine(src)`
+  renders that line with a caret under the position, for messages shown under
+  the input. An input that is not valid UTF-8 names the offset of the first
+  invalid byte in its message. The new fields are additive: a `SyntaxError`
+  built from a line and a message alone is unchanged.
 - `Decoder.UseNumber()` decodes the integers and floats of the document into
   `Number`, which carries the literal the document wrote, so `0x1f`, `1_000`,
   `+1.0` and `inf` survive a round trip with their spelling instead of the

decode_test.go

@@ -1220,3 +1220,56 @@ func TestNumberMethods(t *testing.T) {
 		}
 	}
 }
+
+func TestSyntaxErrorPosition(t *testing.T) {
+	src := []byte("alpha = 1\nbeta x = 2\n")
+	_, err := ParseMap(src)
+	se, ok := errors.AsType[*SyntaxError](err)
+	if !ok {
+		t.Fatalf("err = %v, want a SyntaxError", err)
+	}
+	if se.Line != 2 {
+		t.Errorf("Line = %d, want 2", se.Line)
+	}
+	if want := strings.Index(string(src), "x"); se.Offset != want {
+		t.Errorf("Offset = %d, want %d", se.Offset, want)
+	}
+	if se.Column != 6 {
+		t.Errorf("Column = %d, want 6", se.Column)
+	}
+	want := "beta x = 2\n     ^"
+	if got := se.SourceLine(src); got != want {
+		t.Errorf("SourceLine =\n%s\nwant:\n%s", got, want)
+	}
+}
+
+func TestSyntaxErrorUTF8Offset(t *testing.T) {
+	src := []byte("a = \"ok\"\nb = \"\xff\xfe\"\n")
+	_, err := ParseMap(src)
+	se, ok := errors.AsType[*SyntaxError](err)
+	if !ok {
+		t.Fatalf("err = %v, want a SyntaxError", err)
+	}
+	if !strings.Contains(se.Msg, "byte offset 14") {
+		t.Errorf("Msg = %q, want it to name byte offset 14", se.Msg)
+	}
+	if se.Offset != 14 {
+		t.Errorf("Offset = %d, want 14", se.Offset)
+	}
+	want := "b = \"\xff\xfe\"\n     ^"
+	if got := se.SourceLine(src); got != want {
+		t.Errorf("SourceLine =\n%q\nwant:\n%q", got, want)
+	}
+}
+
+func TestSourceLineEdgePositions(t *testing.T) {
+	src := []byte("a = 1\n")
+	e := &SyntaxError{Line: 1, Msg: "no position"}
+	if got, want := e.SourceLine(src), "a = 1\n^"; got != want {
+		t.Errorf("SourceLine(zero offset) =\n%q\nwant:\n%q", got, want)
+	}
+	e = &SyntaxError{Line: 2, Offset: 100, Msg: "past the end"}
+	if got, want := e.SourceLine(src), "\n^"; got != want {
+		t.Errorf("SourceLine(offset past end) =\n%q\nwant:\n%q", got, want)
+	}
+}

docs/API.md

@@ -642,20 +642,27 @@ sequenceDiagram
 
 ## Types
 
-### `type SyntaxError struct{ Line int; Msg string }`
+### `type SyntaxError struct{ Line, Offset, Column int; Msg string }`
 
-Describes a document the parser rejected, with the 1-based `Line` at which it
+Describes a document the parser rejected: the 1-based `Line` at which it gave
-gave up and `Error()` rendering as `interpres: line N: msg`. A malformed
+up, the `Offset` in bytes the scan stopped at, the 1-based `Column` on that
-document is the usual cause; the nesting limit and an input that is not valid
+line, and `Error()` rendering as `interpres: line N: msg`. A malformed document
-UTF-8 report through the same type. Read the structured fields with a type
+is the usual cause; the nesting limit and an input that is not valid UTF-8
-assertion or `errors.AsType`:
+report through the same type, with the UTF-8 message naming the offset of the
+first invalid byte. Read the structured fields with a type assertion or
+`errors.AsType`:
 
 ```go
 if se, ok := errors.AsType[*interpres.SyntaxError](err); ok {
-	fmt.Println(se.Line, se.Msg)
+	fmt.Println(se.Line, se.Offset, se.Column, se.Msg)
+	fmt.Println(se.SourceLine(data)) // the line, with a caret under Offset
 }
 ```
 
+`SourceLine(src)` renders the source line the error points at from `src`,
+followed by a caret line marking the column, for messages the reader sees
+under the input.
+
 ### `type DecodeError struct{ Path []string; Err error }`
 
 Wraps a decoding failure with the key path at which it happened. `Path` lists

interpres.go

@@ -21,17 +21,23 @@
 package interpres
 
 import (
+	"bytes"
 	"context"
 	"errors"
 	"fmt"
 	"os"
 	"slices"
+	"strings"
 )
 
-// A SyntaxError describes a malformed TOML document, including the 1-based
+// A SyntaxError describes a malformed TOML document. Line is the 1-based line
-// line on which the problem was detected.
+// the problem was detected on. Offset is the byte offset in the input the scan
+// stopped at, and Column is the 1-based column on that line; both are new in
+// 2.0 and a struct literal that names Line and Msg alone still builds.
 type SyntaxError struct {
 	Line   int
+	Offset int
+	Column int
 	Msg    string
 }
 
@@ -39,6 +45,28 @@ func (e *SyntaxError) Error() string {
 	return fmt.Sprintf("interpres: line %d: %s", e.Line, e.Msg)
 }
 
+// SourceLine returns the source line the error points at, rendered from src,
+// followed by a caret line marking the column. It is meant for a message the
+// reader sees under the input:
+//
+//	port = = 8080
+//	       ^
+//
+// The caret sits at Offset when it falls inside src, and at the start of the
+// line when the error carries no position.
+func (e *SyntaxError) SourceLine(src []byte) string {
+	off := min(e.Offset, len(src))
+	start := 0
+	if i := bytes.LastIndexByte(src[:off], '\n'); i >= 0 {
+		start = i + 1
+	}
+	end := len(src)
+	if i := bytes.IndexByte(src[start:], '\n'); i >= 0 {
+		end = start + i
+	}
+	return string(src[start:end]) + "\n" + strings.Repeat(" ", off-start) + "^"
+}
+
 // A DecodeError wraps a decoding failure with the key path at which it
 // happened. Path lists one segment per level from the document root, the
 // outermost key first: a key contributes its name and an array element its

parser.go

@@ -4,6 +4,7 @@
 package interpres
 
 import (
+	"bytes"
 	"context"
 	"fmt"
 	"strconv"
@@ -652,7 +653,7 @@ func (p *parser) parseKeyComponent() (string, error) {
 		// does not decode names that, before any grammar message can.
 		if !p.eof() && p.peek() >= utf8.RuneSelf {
 			if r, size := utf8.DecodeRune(p.src[p.pos:]); r == utf8.RuneError && size == 1 {
-				return "", p.errf("invalid UTF-8 in key")
+				return "", p.errf("invalid UTF-8 in key at byte offset %d", p.pos)
 			}
 		}
 		if p.pos == start {
@@ -707,7 +708,7 @@ func (p *parser) parseAtom() (any, error) {
 		return nil, p.errf("expected a value")
 	}
 	if hasHighByte(tok) && !utf8.ValidString(tok) {
-		return nil, p.errf("invalid UTF-8 in value")
+		return nil, p.errf("invalid UTF-8 in value at byte offset %d", p.pos)
 	}
 	// A date may be followed by a space and a time, forming one date-time.
 	if isDateToken(tok) && !p.eof() && p.peek() == ' ' {
@@ -885,7 +886,7 @@ func (p *parser) writeContentRune(b *strings.Builder) error {
 	}
 	r, size := utf8.DecodeRune(p.src[p.pos:])
 	if r == utf8.RuneError && size == 1 {
-		return p.errf("invalid UTF-8 in string")
+		return p.errf("invalid UTF-8 in string at byte offset %d", p.pos)
 	}
 	p.pos += size
 	b.WriteRune(r)
@@ -1393,7 +1394,7 @@ func (p *parser) skipComment() (string, error) {
 		default:
 			r, size := utf8.DecodeRune(p.src[p.pos:])
 			if r == utf8.RuneError && size == 1 {
-				return "", p.errf("invalid UTF-8 in comment")
+				return "", p.errf("invalid UTF-8 in comment at byte offset %d", p.pos)
 			}
 			p.pos += size
 		}
@@ -1451,13 +1452,21 @@ func (p *parser) expectLineEnd() error {
 	}
 	r, size := utf8.DecodeRune(p.src[p.pos:])
 	if r == utf8.RuneError && size == 1 {
-		return p.errf("invalid UTF-8 after value")
+		return p.errf("invalid UTF-8 after value at byte offset %d", p.pos)
 	}
 	return p.errf("unexpected %q after value", string(r))
 }
 
+// errf builds the SyntaxError with the position the scan stopped at: the line,
+// the byte offset in the input, and the 1-based column on that line. The
+// offset is the cursor, which on an escape or a delimiter run sits just after
+// the bytes that caused the complaint; SourceLine renders the caret there.
 func (p *parser) errf(format string, args ...any) error {
-	return &SyntaxError{Line: p.line, Msg: fmt.Sprintf(format, args...)}
+	col := p.pos + 1
+	if start := bytes.LastIndexByte(p.src[:p.pos], '\n'); start >= 0 {
+		col = p.pos - start
+	}
+	return &SyntaxError{Line: p.line, Offset: p.pos, Column: col, Msg: fmt.Sprintf(format, args...)}
 }
 
 // pathKey joins key components with a NUL separator so a dotted path can be