fix: block positions

Author: yuin

ast/ast.go

@@ -57,6 +57,11 @@ type Node interface {
 	// If this node position is not defined, Pos returns -1.
 	Pos() int
 
+	// SetPos sets a position of this node in a source.
+	// Some node may ignore this method. For example, Paragraph node ignores this method because
+	// it calculates its position from its lines.
+	SetPos(v int)
+
 	// NextSibling returns a next sibling node of this node.
 	NextSibling() Node
 
@@ -179,6 +184,23 @@ type Node interface {
 	RemoveAttributes()
 }
 
+type pos struct {
+	has   bool
+	value int
+}
+
+func (p *pos) Pos() int {
+	if p.has {
+		return p.value
+	}
+	return -1
+}
+
+func (p *pos) SetPos(v int) {
+	p.has = true
+	p.value = v
+}
+
 // A BaseNode struct implements the Node interface partialliy.
 type BaseNode struct {
 	firstChild Node
@@ -188,6 +210,7 @@ type BaseNode struct {
 	prev       Node
 	childCount int
 	attributes []Attribute
+	pos        pos
 }
 
 func ensureIsolated(v Node) {
@@ -196,6 +219,16 @@ func ensureIsolated(v Node) {
 	}
 }
 
+// Pos implements Node.Pos .
+func (n *BaseNode) Pos() int {
+	return n.pos.Pos()
+}
+
+// SetPos implements Node.SetPos .
+func (n *BaseNode) SetPos(v int) {
+	n.pos.SetPos(v)
+}
+
 // HasChildren implements Node.HasChildren .
 func (n *BaseNode) HasChildren() bool {
 	return n.firstChild != nil

ast/block.go

@@ -19,14 +19,6 @@ func (b *BaseBlock) Type() NodeType {
 	return TypeBlock
 }
 
-// Pos implements Node.Pos.
-func (b *BaseBlock) Pos() int {
-	if b.lines.Len() == 0 {
-		return -1
-	}
-	return b.lines.At(0).Start
-}
-
 // IsRaw implements Node.IsRaw.
 func (b *BaseBlock) IsRaw() bool {
 	return false
@@ -132,6 +124,14 @@ func (n *TextBlock) Dump(source []byte, level int) {
 	DumpHelper(n, source, level, nil, nil)
 }
 
+// Pos implements Node.Pos.
+func (n *TextBlock) Pos() int {
+	if n.lines.Len() == 0 {
+		return -1
+	}
+	return n.lines.At(0).Start
+}
+
 // KindTextBlock is a NodeKind of the TextBlock node.
 var KindTextBlock = NewNodeKind("TextBlock")
 
@@ -164,6 +164,14 @@ func (n *Paragraph) Dump(source []byte, level int) {
 	DumpHelper(n, source, level, nil, nil)
 }
 
+// Pos implements Node.Pos.
+func (n *Paragraph) Pos() int {
+	if n.lines.Len() == 0 {
+		return -1
+	}
+	return n.lines.At(0).Start
+}
+
 // KindParagraph is a NodeKind of the Paragraph node.
 var KindParagraph = NewNodeKind("Paragraph")
 
@@ -512,6 +520,7 @@ func (n *HTMLBlock) Dump(source []byte, level int) {
 	indent := strings.Repeat("    ", level)
 	fmt.Printf("%s%s {\n", indent, "HTMLBlock")
 	indent2 := strings.Repeat("    ", level+1)
+	fmt.Printf("%sPos: %d\n", indent2, n.Pos())
 	fmt.Printf("%sRawText: \"", indent2)
 	for i := range n.Lines().Len() {
 		s := n.Lines().At(i)
@@ -576,6 +585,14 @@ func (l *LinkReferenceDefinition) IsRaw() bool {
 	return true
 }
 
+// Pos implements Node.Pos.
+func (l *LinkReferenceDefinition) Pos() int {
+	if l.lines.Len() == 0 {
+		return -1
+	}
+	return l.lines.At(0).Start
+}
+
 // Dump implements Node.Dump.
 func (l *LinkReferenceDefinition) Dump(source []byte, level int) {
 	m := map[string]string{

ast/inline.go

@@ -11,30 +11,13 @@ import (
 // A BaseInline struct implements the Node interface partialliy.
 type BaseInline struct {
 	BaseNode
-
-	pos    int
-	hasPos bool
 }
 
 // Type implements Node.Type.
 func (b *BaseInline) Type() NodeType {
 	return TypeInline
 }
 
-// Pos implements Node.Pos.
-func (b *BaseInline) Pos() int {
-	if !b.hasPos {
-		return -1
-	}
-	return b.pos
-}
-
-// SetPos sets a position of this node.
-func (b *BaseInline) SetPos(pos int) {
-	b.pos = pos
-	b.hasPos = true
-}
-
 // IsRaw implements Node.IsRaw.
 func (b *BaseInline) IsRaw() bool {
 	return false

ast_test.go

@@ -6,6 +6,7 @@ import (
 
 	. "github.com/yuin/goldmark"
 	"github.com/yuin/goldmark/ast"
+	"github.com/yuin/goldmark/extension"
 	"github.com/yuin/goldmark/parser"
 	"github.com/yuin/goldmark/testutil"
 	"github.com/yuin/goldmark/text"
@@ -334,3 +335,16 @@ aaaa **b**
 		t.Error("unexpected position for 1st image")
 	}
 }
+
+func TestBlockPos(t *testing.T) {
+	markdown := New(
+		WithExtensions(extension.GFM, extension.Footnote, extension.DefinitionList),
+	)
+
+	source := []byte(`
+- [ ] hoge
+`)
+	c := parser.NewContext()
+	n := markdown.Parser().Parse(text.NewReader(source), parser.WithContext(c))
+	n.Dump(source, 0)
+}

extension/ast/definition_list.go

@@ -17,6 +17,14 @@ func (n *DefinitionList) Dump(source []byte, level int) {
 	gast.DumpHelper(n, source, level, nil, nil)
 }
 
+// Pos implements Node.Pos.
+func (n *DefinitionList) Pos() int {
+	if n.FirstChild() != nil {
+		return n.FirstChild().Pos()
+	}
+	return -1
+}
+
 // KindDefinitionList is a NodeKind of the DefinitionList node.
 var KindDefinitionList = gast.NewNodeKind("DefinitionList")
 
@@ -44,6 +52,14 @@ func (n *DefinitionTerm) Dump(source []byte, level int) {
 	gast.DumpHelper(n, source, level, nil, nil)
 }
 
+// Pos implements Node.Pos.
+func (n *DefinitionTerm) Pos() int {
+	if n.Lines().Len() == 0 {
+		return -1
+	}
+	return n.Lines().At(0).Start
+}
+
 // KindDefinitionTerm is a NodeKind of the DefinitionTerm node.
 var KindDefinitionTerm = gast.NewNodeKind("DefinitionTerm")
 

extension/ast/table.go

@@ -123,6 +123,7 @@ func (n *TableHeader) Dump(source []byte, level int) {
 // NewTableHeader returns a new TableHeader node.
 func NewTableHeader(row *TableRow) *TableHeader {
 	n := &TableHeader{}
+	n.SetPos(row.Pos())
 	for c := row.FirstChild(); c != nil; {
 		next := c.NextSibling()
 		n.AppendChild(n, c)

extension/table.go

@@ -154,6 +154,7 @@ func NewTableParagraphTransformer() parser.ParagraphTransformer {
 }
 
 func (b *tableParagraphTransformer) Transform(node *gast.Paragraph, reader text.Reader, pc parser.Context) {
+	ppos := node.Pos()
 	lines := node.Lines()
 	if lines.Len() < 2 {
 		return
@@ -169,6 +170,7 @@ func (b *tableParagraphTransformer) Transform(node *gast.Paragraph, reader text.
 		}
 		table := ast.NewTable()
 		table.Alignments = alignments
+		table.SetPos(ppos)
 		table.AppendChild(table, ast.NewTableHeader(header))
 		for j := i + 1; j < lines.Len(); j++ {
 			table.AppendChild(table, b.parseRow(lines.At(j), alignments, false, reader, pc))
@@ -187,13 +189,15 @@ func (b *tableParagraphTransformer) Transform(node *gast.Paragraph, reader text.
 
 func (b *tableParagraphTransformer) parseRow(segment text.Segment,
 	alignments []ast.Alignment, isHeader bool, reader text.Reader, pc parser.Context) *ast.TableRow {
+	npos := segment
 	source := reader.Source()
 	segment = segment.TrimLeftSpace(source)
 	segment = segment.TrimRightSpace(source)
 	line := segment.Value(source)
 	pos := 0
 	limit := len(line)
 	row := ast.NewTableRow(alignments)
+	row.SetPos(npos.Start)
 	if len(line) > 0 && line[pos] == '|' {
 		pos++
 	}
@@ -213,6 +217,7 @@ func (b *tableParagraphTransformer) parseRow(segment text.Segment,
 
 		var escapedCell *escapedPipeCell
 		node := ast.NewTableCell()
+		node.SetPos(npos.Start + pos - npos.Padding)
 		node.Alignment = alignment
 		hasBacktick := false
 		closure := pos

parser/parser.go

@@ -985,13 +985,17 @@ retry:
 		if continuable && result == noBlocksOpened && !bp.CanInterruptParagraph() {
 			continue
 		}
+
 		if w > 3 && !bp.CanAcceptIndentedLine() {
 			continue
 		}
 		lastBlock = pc.LastOpenedBlock()
 		last := lastBlock.Node
+		_, blockPos := reader.Position()
 		node, state := bp.Open(parent, reader, pc)
 		if node != nil {
+			node.SetPos(blockPos.Start + max(pc.BlockOffset(), 0))
+
 			// Parser requires last node to be a paragraph.
 			// With table extension:
 			//