Split parse and render for Goldmark

This also speeds up situations where you only need the fragments/toc and not the rendered content, e.g. Related
with fragments type indexing:

```bash

name            old time/op    new time/op    delta
RelatedSite-10    12.3ms ± 2%    10.7ms ± 1%  -12.95%  (p=0.029 n=4+4)

name            old alloc/op   new alloc/op   delta
RelatedSite-10    38.6MB ± 0%    38.2MB ± 0%   -1.08%  (p=0.029 n=4+4)

name            old allocs/op  new allocs/op  delta
RelatedSite-10      117k ± 0%      115k ± 0%   -1.36%  (p=0.029 n=4+4)
```

Fixes #10750

hugolib/content_render_hooks_test.go

@@ -427,3 +427,52 @@ Image:
 <p>html-image: image.jpg|Text: Hello<br> Goodbye|Plain: Hello GoodbyeEND</p>
 `)
 }
+
+func TestRenderHookContentFragmentsOnSelf(t *testing.T) {
+	files := `
+-- hugo.toml --
+baseURL = "https://example.org"
+disableKinds = ["taxonomy", "term", "RSS", "sitemap", "robotsTXT"]
+-- content/p1.md --
+---
+title: "p1"
+---
+
+## A {#z}
+## B
+## C
+
+-- content/p2.md --
+---
+title: "p2"
+---
+
+## D
+## E
+## F
+
+-- layouts/_default/_markup/render-heading.html --
+Heading: {{ .Text }}|
+Self Fragments: {{ .Page.Fragments.Identifiers }}|
+P1 Fragments: {{ (site.GetPage "p1.md").Fragments.Identifiers }}|
+-- layouts/_default/single.html --
+{{ .Content}}
+`
+
+	b := NewIntegrationTestBuilder(
+		IntegrationTestConfig{
+			T:           t,
+			TxtarString: files,
+		},
+	).Build()
+
+	b.AssertFileContent("public/p1/index.html", `
+Self Fragments: [b c z]
+P1 Fragments: [b c z]
+	`)
+	b.AssertFileContent("public/p2/index.html", `
+Self Fragments: [d e f]
+P1 Fragments: [b c z]
+	`)
+
+}

hugolib/integrationtest_builder.go

@@ -32,6 +32,8 @@ import (
 func NewIntegrationTestBuilder(conf IntegrationTestConfig) *IntegrationTestBuilder {
 	// Code fences.
 	conf.TxtarString = strings.ReplaceAll(conf.TxtarString, "§§§", "```")
+	// Multiline strings.
+	conf.TxtarString = strings.ReplaceAll(conf.TxtarString, "§§", "`")
 
 	data := txtar.Parse([]byte(conf.TxtarString))
 

hugolib/page__per_output.go

@@ -115,14 +115,7 @@ func newPageContentOutput(p *pageState, po *pageOutput) (*pageContentOutput, err
 		isHTML := cp.p.m.markup == "html"
 
 		if !isHTML {
-			r, err := po.contentRenderer.RenderContent(ctx, cp.workContent, true)
+			createAndSetToC := func(tocProvider converter.TableOfContentsProvider) {
-			if err != nil {
-				return err
-			}
-
-			cp.workContent = r.Bytes()
-
-			if tocProvider, ok := r.(converter.TableOfContentsProvider); ok {
 				cfg := p.s.ContentSpec.Converters.GetMarkupConfig()
 				cp.tableOfContents = tocProvider.TableOfContents()
 				cp.tableOfContentsHTML = template.HTML(
@@ -132,6 +125,31 @@ func newPageContentOutput(p *pageState, po *pageOutput) (*pageContentOutput, err
 						cfg.TableOfContents.Ordered,
 					),
 				)
+			}
+			// If the converter supports doing the parsing separately, we do that.
+			parseResult, ok, err := po.contentRenderer.ParseContent(ctx, cp.workContent)
+			if err != nil {
+				return err
+			}
+			if ok {
+				// This is Goldmark.
+				// Store away the parse result for later use.
+				createAndSetToC(parseResult)
+				cp.astDoc = parseResult.Doc()
+
+				return nil
+			}
+
+			// This is Asciidoctor etc.
+			r, err := po.contentRenderer.ParseAndRenderContent(ctx, cp.workContent, true)
+			if err != nil {
+				return err
+			}
+
+			cp.workContent = r.Bytes()
+
+			if tocProvider, ok := r.(converter.TableOfContentsProvider); ok {
+				createAndSetToC(tocProvider)
 			} else {
 				tmpContent, tmpTableOfContents := helpers.ExtractTOC(cp.workContent)
 				cp.tableOfContentsHTML = helpers.BytesToHTML(tmpTableOfContents)
@@ -153,6 +171,19 @@ func newPageContentOutput(p *pageState, po *pageOutput) (*pageContentOutput, err
 			return nil
 		}
 
+		if cp.astDoc != nil {
+			// The content is parsed, but not rendered.
+			r, ok, err := po.contentRenderer.RenderContent(ctx, cp.workContent, cp.astDoc)
+			if err != nil {
+				return err
+			}
+			if !ok {
+				return errors.New("invalid state: astDoc is set but RenderContent returned false")
+			}
+
+			cp.workContent = r.Bytes()
+		}
+
 		if p.cmap.hasNonMarkdownShortcode || cp.placeholdersEnabled {
 			// There are one or more replacement tokens to be replaced.
 			var hasShortcodeVariants bool
@@ -210,7 +241,7 @@ func newPageContentOutput(p *pageState, po *pageOutput) (*pageContentOutput, err
 				}
 			}
 		} else if cp.p.m.summary != "" {
-			b, err := po.contentRenderer.RenderContent(ctx, []byte(cp.p.m.summary), false)
+			b, err := po.contentRenderer.ParseAndRenderContent(ctx, []byte(cp.p.m.summary), false)
 			if err != nil {
 				return err
 			}
@@ -282,6 +313,8 @@ type pageContentOutput struct {
 	summary             template.HTML
 	tableOfContents     *tableofcontents.Fragments
 	tableOfContentsHTML template.HTML
+	// For Goldmark we split Parse and Render.
+	astDoc any
 
 	truncated bool
 
@@ -682,15 +715,66 @@ func (p *pageContentOutput) setAutoSummary() error {
 	return nil
 }
 
-func (cp *pageContentOutput) RenderContent(ctx context.Context, content []byte, renderTOC bool) (converter.Result, error) {
+func (cp *pageContentOutput) getContentConverter() (converter.Converter, error) {
 	if err := cp.initRenderHooks(); err != nil {
 		return nil, err
 	}
-	c := cp.p.getContentConverter()
+	return cp.p.getContentConverter(), nil
+}
+
+func (cp *pageContentOutput) ParseAndRenderContent(ctx context.Context, content []byte, renderTOC bool) (converter.ResultRender, error) {
+	c, err := cp.getContentConverter()
+	if err != nil {
+		return nil, err
+	}
 	return cp.renderContentWithConverter(ctx, c, content, renderTOC)
 }
 
-func (cp *pageContentOutput) renderContentWithConverter(ctx context.Context, c converter.Converter, content []byte, renderTOC bool) (converter.Result, error) {
+func (cp *pageContentOutput) ParseContent(ctx context.Context, content []byte) (converter.ResultParse, bool, error) {
+	c, err := cp.getContentConverter()
+	if err != nil {
+		return nil, false, err
+	}
+	p, ok := c.(converter.ParseRenderer)
+	if !ok {
+		return nil, ok, nil
+	}
+	rctx := converter.RenderContext{
+		Src:         content,
+		RenderTOC:   true,
+		GetRenderer: cp.renderHooks.getRenderer,
+	}
+	r, err := p.Parse(rctx)
+	return r, ok, err
+
+}
+func (cp *pageContentOutput) RenderContent(ctx context.Context, content []byte, doc any) (converter.ResultRender, bool, error) {
+	c, err := cp.getContentConverter()
+	if err != nil {
+		return nil, false, err
+	}
+	p, ok := c.(converter.ParseRenderer)
+	if !ok {
+		return nil, ok, nil
+	}
+	rctx := converter.RenderContext{
+		Src:         content,
+		RenderTOC:   true,
+		GetRenderer: cp.renderHooks.getRenderer,
+	}
+	r, err := p.Render(rctx, doc)
+	if err == nil {
+		if ids, ok := r.(identity.IdentitiesProvider); ok {
+			for _, v := range ids.GetIdentities() {
+				cp.trackDependency(v)
+			}
+		}
+	}
+
+	return r, ok, err
+}
+
+func (cp *pageContentOutput) renderContentWithConverter(ctx context.Context, c converter.Converter, content []byte, renderTOC bool) (converter.ResultRender, error) {
 	r, err := c.Convert(
 		converter.RenderContext{
 			Src:         content,

hugolib/shortcode.go

@@ -422,7 +422,7 @@ func doRenderShortcode(
 		// shortcode.
 		if sc.doMarkup && (level > 0 || sc.configVersion() == 1) {
 			var err error
-			b, err := p.pageOutput.contentRenderer.RenderContent(ctx, []byte(inner), false)
+			b, err := p.pageOutput.contentRenderer.ParseAndRenderContent(ctx, []byte(inner), false)
 			if err != nil {
 				return zeroShortcode, err
 			}

markup/asciidocext/convert.go

@@ -52,7 +52,7 @@ func (p provider) New(cfg converter.ProviderConfig) (converter.Provider, error)
 }
 
 type asciidocResult struct {
-	converter.Result
+	converter.ResultRender
 	toc *tableofcontents.Fragments
 }
 
@@ -65,7 +65,7 @@ type asciidocConverter struct {
 	cfg converter.ProviderConfig
 }
 
-func (a *asciidocConverter) Convert(ctx converter.RenderContext) (converter.Result, error) {
+func (a *asciidocConverter) Convert(ctx converter.RenderContext) (converter.ResultRender, error) {
 	b, err := a.getAsciidocContent(ctx.Src, a.ctx)
 	if err != nil {
 		return nil, err
@@ -75,8 +75,8 @@ func (a *asciidocConverter) Convert(ctx converter.RenderContext) (converter.Resu
 		return nil, err
 	}
 	return asciidocResult{
-		Result: converter.Bytes(content),
+		ResultRender: converter.Bytes(content),
-		toc:    toc,
+		toc:          toc,
 	}, nil
 }
 

markup/converter/converter.go

@@ -74,7 +74,7 @@ var NopConverter = new(nopConverter)
 
 type nopConverter int
 
-func (nopConverter) Convert(ctx RenderContext) (Result, error) {
+func (nopConverter) Convert(ctx RenderContext) (ResultRender, error) {
 	return &bytes.Buffer{}, nil
 }
 
@@ -85,15 +85,29 @@ func (nopConverter) Supports(feature identity.Identity) bool {
 // Converter wraps the Convert method that converts some markup into
 // another format, e.g. Markdown to HTML.
 type Converter interface {
-	Convert(ctx RenderContext) (Result, error)
+	Convert(ctx RenderContext) (ResultRender, error)
 	Supports(feature identity.Identity) bool
 }
 
-// Result represents the minimum returned from Convert.
+// ParseRenderer is an optional interface.
-type Result interface {
+// The Goldmark converter implements this, and this allows us
+// to extract the ToC without having to render the content.
+type ParseRenderer interface {
+	Parse(RenderContext) (ResultParse, error)
+	Render(RenderContext, any) (ResultRender, error)
+}
+
+// ResultRender represents the minimum returned from Convert and Render.
+type ResultRender interface {
 	Bytes() []byte
 }
 
+// ResultParse represents the minimum returned from Parse.
+type ResultParse interface {
+	Doc() any
+	TableOfContents() *tableofcontents.Fragments
+}
+
 // DocumentInfo holds additional information provided by some converters.
 type DocumentInfo interface {
 	AnchorSuffix() string

markup/goldmark/convert.go

@@ -18,6 +18,7 @@ import (
 	"bytes"
 
 	"github.com/gohugoio/hugo/identity"
+
 	"github.com/gohugoio/hugo/markup/goldmark/codeblocks"
 	"github.com/gohugoio/hugo/markup/goldmark/images"
 	"github.com/gohugoio/hugo/markup/goldmark/internal/extensions/attributes"
@@ -26,6 +27,7 @@ import (
 	"github.com/gohugoio/hugo/markup/converter"
 	"github.com/gohugoio/hugo/markup/tableofcontents"
 	"github.com/yuin/goldmark"
+	"github.com/yuin/goldmark/ast"
 	"github.com/yuin/goldmark/extension"
 	"github.com/yuin/goldmark/parser"
 	"github.com/yuin/goldmark/renderer"
@@ -158,26 +160,41 @@ func newMarkdown(pcfg converter.ProviderConfig) goldmark.Markdown {
 
 var _ identity.IdentitiesProvider = (*converterResult)(nil)
 
-type converterResult struct {
+type parserResult struct {
-	converter.Result
+	doc any
 	toc *tableofcontents.Fragments
+}
+
+func (p parserResult) Doc() any {
+	return p.doc
+}
+
+func (p parserResult) TableOfContents() *tableofcontents.Fragments {
+	return p.toc
+}
+
+type renderResult struct {
+	converter.ResultRender
 	ids identity.Identities
 }
 
-func (c converterResult) TableOfContents() *tableofcontents.Fragments {
+func (r renderResult) GetIdentities() identity.Identities {
-	return c.toc
+	return r.ids
 }
 
-func (c converterResult) GetIdentities() identity.Identities {
+type converterResult struct {
-	return c.ids
+	converter.ResultRender
+	tableOfContentsProvider
+	identity.IdentitiesProvider
+}
+
+type tableOfContentsProvider interface {
+	TableOfContents() *tableofcontents.Fragments
 }
 
 var converterIdentity = identity.KeyValueIdentity{Key: "goldmark", Value: "converter"}
 
-func (c *goldmarkConverter) Convert(ctx converter.RenderContext) (result converter.Result, err error) {
+func (c *goldmarkConverter) Parse(ctx converter.RenderContext) (converter.ResultParse, error) {
-
-	buf := &render.BufWriter{Buffer: &bytes.Buffer{}}
-	result = buf
 	pctx := c.newParserContext(ctx)
 	reader := text.NewReader(ctx.Src)
 
@@ -186,6 +203,16 @@ func (c *goldmarkConverter) Convert(ctx converter.RenderContext) (result convert
 		parser.WithContext(pctx),
 	)
 
+	return parserResult{
+		doc: doc,
+		toc: pctx.TableOfContents(),
+	}, nil
+
+}
+func (c *goldmarkConverter) Render(ctx converter.RenderContext, doc any) (converter.ResultRender, error) {
+	n := doc.(ast.Node)
+	buf := &render.BufWriter{Buffer: &bytes.Buffer{}}
+
 	rcx := &render.RenderContextDataHolder{
 		Rctx: ctx,
 		Dctx: c.ctx,
@@ -197,15 +224,32 @@ func (c *goldmarkConverter) Convert(ctx converter.RenderContext) (result convert
 		ContextData: rcx,
 	}
 
-	if err := c.md.Renderer().Render(w, ctx.Src, doc); err != nil {
+	if err := c.md.Renderer().Render(w, ctx.Src, n); err != nil {
 		return nil, err
 	}
 
-	return converterResult{
+	return renderResult{
-		Result: buf,
+		ResultRender: buf,
-		ids:    rcx.IDs.GetIdentities(),
+		ids:          rcx.IDs.GetIdentities(),
-		toc:    pctx.TableOfContents(),
 	}, nil
+
+}
+
+func (c *goldmarkConverter) Convert(ctx converter.RenderContext) (converter.ResultRender, error) {
+	parseResult, err := c.Parse(ctx)
+	if err != nil {
+		return nil, err
+	}
+	renderResult, err := c.Render(ctx, parseResult.Doc())
+	if err != nil {
+		return nil, err
+	}
+	return converterResult{
+		ResultRender:            renderResult,
+		tableOfContentsProvider: parseResult,
+		IdentitiesProvider:      renderResult.(identity.IdentitiesProvider),
+	}, nil
+
 }
 
 var featureSet = map[identity.Identity]bool{

markup/goldmark/convert_test.go

@@ -34,7 +34,7 @@ import (
 	qt "github.com/frankban/quicktest"
 )
 
-func convert(c *qt.C, mconf markup_config.Config, content string) converter.Result {
+func convert(c *qt.C, mconf markup_config.Config, content string) converter.ResultRender {
 	p, err := Provider.New(
 		converter.ProviderConfig{
 			MarkupConfig: mconf,

markup/org/convert.go

@@ -43,7 +43,7 @@ type orgConverter struct {
 	cfg converter.ProviderConfig
 }
 
-func (c *orgConverter) Convert(ctx converter.RenderContext) (converter.Result, error) {
+func (c *orgConverter) Convert(ctx converter.RenderContext) (converter.ResultRender, error) {
 	logger := c.cfg.Logger
 	config := org.New()
 	config.Log = logger.Warn()

markup/pandoc/convert.go

@@ -43,7 +43,7 @@ type pandocConverter struct {
 	cfg converter.ProviderConfig
 }
 
-func (c *pandocConverter) Convert(ctx converter.RenderContext) (converter.Result, error) {
+func (c *pandocConverter) Convert(ctx converter.RenderContext) (converter.ResultRender, error) {
 	b, err := c.getPandocContent(ctx.Src, c.ctx)
 	if err != nil {
 		return nil, err

markup/rst/convert.go

@@ -47,7 +47,7 @@ type rstConverter struct {
 	cfg converter.ProviderConfig
 }
 
-func (c *rstConverter) Convert(ctx converter.RenderContext) (converter.Result, error) {
+func (c *rstConverter) Convert(ctx converter.RenderContext) (converter.ResultRender, error) {
 	b, err := c.getRstContent(ctx.Src, c.ctx)
 	if err != nil {
 		return nil, err

resources/page/page.go

@@ -109,9 +109,13 @@ type ContentProvider interface {
 
 // ContentRenderer provides the content rendering methods for some content.
 type ContentRenderer interface {
-	// RenderContent renders the given content.
+	// ParseAndRenderContent renders the given content.
 	// For internal use only.
-	RenderContent(ctx context.Context, content []byte, renderTOC bool) (converter.Result, error)
+	ParseAndRenderContent(ctx context.Context, content []byte, enableTOC bool) (converter.ResultRender, error)
+	// For internal use only.
+	ParseContent(ctx context.Context, content []byte) (converter.ResultParse, bool, error)
+	// For internal use only.
+	RenderContent(ctx context.Context, content []byte, doc any) (converter.ResultRender, bool, error)
 }
 
 // FileProvider provides the source file.

resources/page/page_lazy_contentprovider.go

@@ -133,7 +133,16 @@ func (lcp *LazyContentProvider) TableOfContents(ctx context.Context) template.HT
 	return lcp.cp.TableOfContents(ctx)
 }
 
-func (lcp *LazyContentProvider) RenderContent(ctx context.Context, content []byte, renderTOC bool) (converter.Result, error) {
+func (lcp *LazyContentProvider) ParseAndRenderContent(ctx context.Context, content []byte, renderTOC bool) (converter.ResultRender, error) {
 	lcp.init.Do(ctx)
-	return lcp.cp.RenderContent(ctx, content, renderTOC)
+	return lcp.cp.ParseAndRenderContent(ctx, content, renderTOC)
+}
+
+func (lcp *LazyContentProvider) ParseContent(ctx context.Context, content []byte) (converter.ResultParse, bool, error) {
+	lcp.init.Do(ctx)
+	return lcp.cp.ParseContent(ctx, content)
+}
+func (lcp *LazyContentProvider) RenderContent(ctx context.Context, content []byte, doc any) (converter.ResultRender, bool, error) {
+	lcp.init.Do(ctx)
+	return lcp.cp.RenderContent(ctx, content, doc)
 }

resources/page/page_nop.go

@@ -538,7 +538,14 @@ func (p *nopPage) HeadingsFiltered(context.Context) tableofcontents.Headings {
 
 type nopContentRenderer int
 
-func (r *nopContentRenderer) RenderContent(ctx context.Context, content []byte, renderTOC bool) (converter.Result, error) {
+func (r *nopContentRenderer) ParseAndRenderContent(ctx context.Context, content []byte, renderTOC bool) (converter.ResultRender, error) {
 	b := &bytes.Buffer{}
 	return b, nil
 }
+
+func (r *nopContentRenderer) ParseContent(ctx context.Context, content []byte) (converter.ResultParse, bool, error) {
+	return nil, false, nil
+}
+func (r *nopContentRenderer) RenderContent(ctx context.Context, content []byte, doc any) (converter.ResultRender, bool, error) {
+	return nil, false, nil
+}