all: Rework page store, add a dynacache, improve partial rebuilds, and some general spring cleaningdevelop2024

There are some breaking changes in this commit, see #11455.

Closes #11455
Closes #11549

This fixes a set of bugs (see issue list) and it is also paying some technical debt accumulated over the years. We now build with Staticcheck enabled in the CI build.

The performance should be about the same as before for regular sized Hugo sites, but it should perform and scale much better to larger data sets, as objects that uses lots of memory (e.g. rendered Markdown, big JSON files read into maps with transform.Unmarshal etc.) will now get automatically garbage collected if needed. Performance on partial rebuilds when running the server in fast render mode should be the same, but the change detection should be much more accurate.

A list of the notable new features:

* A new dependency tracker that covers (almost) all of Hugo's API and is used to do fine grained partial rebuilds when running the server.
* A new and simpler tree document store which allows fast lookups and prefix-walking in all dimensions (e.g. language) concurrently.
* You can now configure an upper memory limit allowing for much larger data sets and/or running on lower specced PCs.
We have lifted the "no resources in sub folders" restriction for branch bundles (e.g. sections).
Memory Limit
* Hugos will, by default, set aside a quarter of the total system memory, but you can set this via the OS environment variable HUGO_MEMORYLIMIT (in gigabytes). This is backed by a partitioned LRU cache used throughout Hugo. A cache that gets dynamically resized in low memory situations, allowing Go's Garbage Collector to free the memory.

New Dependency Tracker: Hugo has had a rule based coarse grained approach to server rebuilds that has worked mostly pretty well, but there have been some surprises (e.g. stale content). This is now revamped with a new dependency tracker that can quickly calculate the delta given a changed resource (e.g. a content file, template, JS file etc.). This handles transitive relations, e.g. $page -> js.Build -> JS import, or $page1.Content -> render hook -> site.GetPage -> $page2.Title, or $page1.Content -> shortcode -> partial -> site.RegularPages -> $page2.Content -> shortcode ..., and should also handle changes to aggregated values (e.g. site.Lastmod) effectively.

This covers all of Hugo's API with 2 known exceptions (a list that may not be fully exhaustive):

Changes to files loaded with template func os.ReadFile may not be handled correctly. We recommend loading resources with resources.Get
Changes to Hugo objects (e.g. Page) passed in the template context to lang.Translate may not be detected correctly. We recommend having simple i18n templates without too much data context passed in other than simple types such as strings and numbers.
Note that the cachebuster configuration (when A changes then rebuild B) works well with the above, but we recommend that you revise that configuration, as it in most situations should not be needed. One example where it is still needed is with TailwindCSS and using changes to hugo_stats.json to trigger new CSS rebuilds.

Document Store: Previously, a little simplified, we split the document store (where we store pages and resources) in a tree per language. This worked pretty well, but the structure made some operations harder than they needed to be. We have now restructured it into one Radix tree for all languages. Internally the language is considered to be a dimension of that tree, and the tree can be viewed in all dimensions concurrently. This makes some operations re. language simpler (e.g. finding translations is just a slice range), but the idea is that it should also be relatively inexpensive to add more dimensions if needed (e.g. role).

Fixes #10169
Fixes #10364
Fixes #10482
Fixes #10630
Fixes #10656
Fixes #10694
Fixes #10918
Fixes #11262
Fixes #11439
Fixes #11453
Fixes #11457
Fixes #11466
Fixes #11540
Fixes #11551
Fixes #11556
Fixes #11654
Fixes #11661
Fixes #11663
Fixes #11664
Fixes #11669
Fixes #11671
Fixes #11807
Fixes #11808
Fixes #11809
Fixes #11815
Fixes #11840
Fixes #11853
Fixes #11860
Fixes #11883
Fixes #11904
Fixes #7388
Fixes #7425
Fixes #7436
Fixes #7544
Fixes #7882
Fixes #7960
Fixes #8255
Fixes #8307
Fixes #8863
Fixes #8927
Fixes #9192
Fixes #9324

1 files changed, 236 insertions, 437 deletions

diff --git a/hugolib/pages_capture.go b/hugolib/pages_capture.go
index c57c707de..acdc674e6 100644
--- a/hugolib/pages_capture.go
+++ b/hugolib/pages_capture.go
@@ -1,4 +1,4 @@
-// Copyright 2019 The Hugo Authors. All rights reserved.
+// Copyright 2021 The Hugo Authors. All rights reserved.
 //
 // Licensed under the Apache License, Version 2.0 (the "License");
 // you may not use this file except in compliance with the License.
@@ -15,190 +15,188 @@ package hugolib
 
 import (
 	"context"
+	"errors"
 	"fmt"
-	pth "path"
+	"os"
 	"path/filepath"
-	"reflect"
-
-	"github.com/gohugoio/hugo/common/herrors"
-	"github.com/gohugoio/hugo/common/loggers"
-	"github.com/gohugoio/hugo/common/maps"
-
+	"strings"
+	"sync"
+	"sync/atomic"
+	"time"
+
+	"github.com/bep/logg"
+	"github.com/gohugoio/hugo/common/paths"
+	"github.com/gohugoio/hugo/common/rungroup"
+	"github.com/gohugoio/hugo/helpers"
 	"github.com/gohugoio/hugo/parser/pageparser"
-
-	"github.com/gohugoio/hugo/hugofs/files"
+	"github.com/spf13/afero"
 
 	"github.com/gohugoio/hugo/source"
 
+	"github.com/gohugoio/hugo/common/loggers"
 	"github.com/gohugoio/hugo/hugofs"
-	"github.com/spf13/afero"
-)
-
-const (
-	walkIsRootFileMetaKey = "walkIsRootFileMetaKey"
 )
 
 func newPagesCollector(
+	ctx context.Context,
+	h *HugoSites,
 	sp *source.SourceSpec,
-	contentMap *pageMaps,
 	logger loggers.Logger,
-	contentTracker *contentChangeMap,
-	proc pagesCollectorProcessorProvider, filenames ...string) *pagesCollector {
+	infoLogger logg.LevelLogger,
+	m *pageMap,
+	ids []pathChange,
+) *pagesCollector {
 	return &pagesCollector{
-		fs:         sp.SourceFs,
-		contentMap: contentMap,
-		proc:       proc,
+		ctx:        ctx,
+		h:          h,
+		fs:         sp.BaseFs.Content.Fs,
+		m:          m,
 		sp:         sp,
 		logger:     logger,
-		filenames:  filenames,
-		tracker:    contentTracker,
+		infoLogger: infoLogger,
+		ids:        ids,
+		seenDirs:   make(map[string]bool),
 	}
 }
 
-type contentDirKey struct {
-	dirname  string
-	filename string
-	tp       bundleDirType
-}
-
-type fileinfoBundle struct {
-	header    hugofs.FileMetaInfo
-	resources []hugofs.FileMetaInfo
-}
-
-func (b *fileinfoBundle) containsResource(name string) bool {
-	for _, r := range b.resources {
-		if r.Name() == name {
-			return true
-		}
-	}
-
-	return false
-}
-
-type pageBundles map[string]*fileinfoBundle
-
 type pagesCollector struct {
-	sp     *source.SourceSpec
-	fs     afero.Fs
-	logger loggers.Logger
+	ctx        context.Context
+	h          *HugoSites
+	sp         *source.SourceSpec
+	logger     loggers.Logger
+	infoLogger logg.LevelLogger
 
-	contentMap *pageMaps
+	m *pageMap
 
-	// Ordered list (bundle headers first) used in partial builds.
-	filenames []string
+	fs afero.Fs
 
-	// Content files tracker used in partial builds.
-	tracker *contentChangeMap
+	// List of paths that have changed. Used in partial builds.
+	ids      []pathChange
+	seenDirs map[string]bool
 
-	proc pagesCollectorProcessorProvider
+	g rungroup.Group[hugofs.FileMetaInfo]
 }
 
-// isCascadingEdit returns whether the dir represents a cascading edit.
-// That is, if a front matter cascade section is removed, added or edited.
-// If this is the case we must re-evaluate its descendants.
-func (c *pagesCollector) isCascadingEdit(dir contentDirKey) (bool, string) {
-	// This is either a section or a taxonomy node. Find it.
-	prefix := cleanTreeKey(dir.dirname)
-
-	section := "/"
-	var isCascade bool
-
-	c.contentMap.walkBranchesPrefix(prefix, func(s string, n *contentNode) bool {
-		if n.fi == nil || dir.filename != n.fi.Meta().Filename {
-			return false
-		}
-
-		f, err := n.fi.Meta().Open()
-		if err != nil {
-			// File may have been removed, assume a cascading edit.
-			// Some false positives is not too bad.
-			isCascade = true
-			return true
-		}
-
-		pf, err := pageparser.ParseFrontMatterAndContent(f)
-		f.Close()
-		if err != nil {
-			isCascade = true
-			return true
-		}
-
-		if n.p == nil || n.p.bucket == nil {
-			return true
-		}
+func (c *pagesCollector) copyFile(fim hugofs.FileMetaInfo) error {
+	meta := fim.Meta()
+	f, err := meta.Open()
+	if err != nil {
+		return fmt.Errorf("copyFile: failed to open: %w", err)
+	}
 
-		section = s
+	s := c.m.s
 
-		maps.PrepareParams(pf.FrontMatter)
-		cascade1, ok := pf.FrontMatter["cascade"]
-		hasCascade := n.p.bucket.cascade != nil && len(n.p.bucket.cascade) > 0
-		if !ok {
-			isCascade = hasCascade
+	target := filepath.Join(s.PathSpec.GetTargetLanguageBasePath(), meta.PathInfo.Path())
 
-			return true
-		}
-
-		if !hasCascade {
-			isCascade = true
-			return true
-		}
+	defer f.Close()
 
-		for _, v := range n.p.bucket.cascade {
-			isCascade = !reflect.DeepEqual(cascade1, v)
-			if isCascade {
-				break
-			}
-		}
+	fs := s.PublishFsStatic
 
-		return true
-	})
+	s.PathSpec.ProcessingStats.Incr(&s.PathSpec.ProcessingStats.Files)
 
-	return isCascade, section
+	return helpers.WriteToDisk(filepath.Clean(target), f, fs)
 }
 
-// Collect.
+// Collect collects content by walking the file system and storing
+// it in the content tree.
+// It may be restricted by filenames set on the collector (partial build).
 func (c *pagesCollector) Collect() (collectErr error) {
-	c.proc.Start(context.Background())
+	var (
+		numWorkers             = c.h.numWorkers
+		numFilesProcessedTotal atomic.Uint64
+		numFilesProcessedLast  uint64
+		fileBatchTimer         = time.Now()
+		fileBatchTimerMu       sync.Mutex
+	)
+
+	l := c.infoLogger.WithField("substep", "collect")
+
+	logFilesProcessed := func(force bool) {
+		fileBatchTimerMu.Lock()
+		if force || time.Since(fileBatchTimer) > 3*time.Second {
+			numFilesProcessedBatch := numFilesProcessedTotal.Load() - numFilesProcessedLast
+			numFilesProcessedLast = numFilesProcessedTotal.Load()
+			loggers.TimeTrackf(l, fileBatchTimer,
+				logg.Fields{
+					logg.Field{Name: "files", Value: numFilesProcessedBatch},
+					logg.Field{Name: "files_total", Value: numFilesProcessedTotal.Load()},
+				},
+				"",
+			)
+			fileBatchTimer = time.Now()
+		}
+		fileBatchTimerMu.Unlock()
+	}
+
 	defer func() {
-		err := c.proc.Wait()
-		if collectErr == nil {
-			collectErr = err
-		}
+		logFilesProcessed(true)
 	}()
 
-	if len(c.filenames) == 0 {
-		// Collect everything.
-		collectErr = c.collectDir("", false, nil)
-	} else {
-		for _, pm := range c.contentMap.pmaps {
-			pm.cfg.isRebuild = true
-		}
-		dirs := make(map[contentDirKey]bool)
-		for _, filename := range c.filenames {
-			dir, btype := c.tracker.resolveAndRemove(filename)
-			dirs[contentDirKey{dir, filename, btype}] = true
-		}
-
-		for dir := range dirs {
-			for _, pm := range c.contentMap.pmaps {
-				pm.s.ResourceSpec.DeleteBySubstring(dir.dirname)
+	c.g = rungroup.Run[hugofs.FileMetaInfo](c.ctx, rungroup.Config[hugofs.FileMetaInfo]{
+		NumWorkers: numWorkers,
+		Handle: func(ctx context.Context, fi hugofs.FileMetaInfo) error {
+			if err := c.m.AddFi(fi); err != nil {
+				if errors.Is(err, pageparser.ErrPlainHTMLDocumentsNotSupported) {
+					// Reclassify this as a static file.
+					if err := c.copyFile(fi); err != nil {
+						return err
+					}
+				} else {
+					return hugofs.AddFileInfoToError(err, fi, c.fs)
+				}
 			}
+			numFilesProcessedTotal.Add(1)
+			if numFilesProcessedTotal.Load()%1000 == 0 {
+				logFilesProcessed(false)
+			}
+			return nil
+		},
+	})
 
-			switch dir.tp {
-			case bundleLeaf:
-				collectErr = c.collectDir(dir.dirname, true, nil)
-			case bundleBranch:
-				isCascading, section := c.isCascadingEdit(dir)
-
-				if isCascading {
-					c.contentMap.deleteSection(section)
-				}
-				collectErr = c.collectDir(dir.dirname, !isCascading, nil)
-			default:
+	if c.ids == nil {
+		// Collect everything.
+		collectErr = c.collectDir(nil, false, nil)
+	} else {
+		for _, s := range c.h.Sites {
+			s.pageMap.cfg.isRebuild = true
+		}
+
+		for _, id := range c.ids {
+			if id.p.IsLeafBundle() {
+				collectErr = c.collectDir(
+					id.p,
+					false,
+					func(fim hugofs.FileMetaInfo) bool {
+						return true
+					},
+				)
+			} else if id.p.IsBranchBundle() {
+				collectErr = c.collectDir(
+					id.p,
+					false,
+					func(fim hugofs.FileMetaInfo) bool {
+						if fim.IsDir() {
+							return true
+						}
+						fimp := fim.Meta().PathInfo
+						if fimp == nil {
+							return false
+						}
+
+						return strings.HasPrefix(fimp.Path(), paths.AddTrailingSlash(id.p.Dir()))
+					},
+				)
+			} else {
 				// We always start from a directory.
-				collectErr = c.collectDir(dir.dirname, true, func(fim hugofs.FileMetaInfo) bool {
-					return dir.filename == fim.Meta().Filename
+				collectErr = c.collectDir(id.p, id.isDir, func(fim hugofs.FileMetaInfo) bool {
+					if id.delete || id.isDir {
+						if id.isDir {
+							return strings.HasPrefix(fim.Meta().PathInfo.Path(), paths.AddTrailingSlash(id.p.Path()))
+						}
+
+						return id.p.Dir() == fim.Meta().PathInfo.Dir()
+					}
+					return id.p.Path() == fim.Meta().PathInfo.Path()
 				})
 			}
 
@@ -209,160 +207,51 @@ func (c *pagesCollector) Collect() (collectErr error) {
 
 	}
 
-	return
-}
-
-func (c *pagesCollector) isBundleHeader(fi hugofs.FileMetaInfo) bool {
-	class := fi.Meta().Classifier
-	return class == files.ContentClassLeaf || class == files.ContentClassBranch
-}
-
-func (c *pagesCollector) getLang(fi hugofs.FileMetaInfo) string {
-	lang := fi.Meta().Lang
-	if lang != "" {
-		return lang
-	}
-	return c.sp.Cfg.DefaultContentLanguage()
-}
-
-func (c *pagesCollector) addToBundle(info hugofs.FileMetaInfo, btyp bundleDirType, bundles pageBundles) error {
-	getBundle := func(lang string) *fileinfoBundle {
-		return bundles[lang]
+	werr := c.g.Wait()
+	if collectErr == nil {
+		collectErr = werr
 	}
 
-	cloneBundle := func(lang string) *fileinfoBundle {
-		// Every bundled content file needs a content file header.
-		// Use the default content language if found, else just
-		// pick one.
-		var (
-			source *fileinfoBundle
-			found  bool
-		)
-
-		source, found = bundles[c.sp.Cfg.DefaultContentLanguage()]
-		if !found {
-			for _, b := range bundles {
-				source = b
-				break
-			}
-		}
-
-		if source == nil {
-			panic(fmt.Sprintf("no source found, %d", len(bundles)))
-		}
-
-		clone := c.cloneFileInfo(source.header)
-		clone.Meta().Lang = lang
-
-		return &fileinfoBundle{
-			header: clone,
-		}
-	}
+	return
+}
 
-	lang := c.getLang(info)
-	bundle := getBundle(lang)
-	isBundleHeader := c.isBundleHeader(info)
-	if bundle != nil && isBundleHeader {
-		// index.md file inside a bundle, see issue 6208.
-		info.Meta().Classifier = files.ContentClassContent
-		isBundleHeader = false
-	}
-	classifier := info.Meta().Classifier
-	isContent := classifier == files.ContentClassContent
-	if bundle == nil {
-		if isBundleHeader {
-			bundle = &fileinfoBundle{header: info}
-			bundles[lang] = bundle
+func (c *pagesCollector) collectDir(dirPath *paths.Path, isDir bool, inFilter func(fim hugofs.FileMetaInfo) bool) error {
+	var dpath string
+	if dirPath != nil {
+		if isDir {
+			dpath = filepath.FromSlash(dirPath.Path())
 		} else {
-			if btyp == bundleBranch {
-				// No special logic for branch bundles.
-				// Every language needs its own _index.md file.
-				// Also, we only clone bundle headers for lonesome, bundled,
-				// content files.
-				return c.handleFiles(info)
-			}
-
-			if isContent {
-				bundle = cloneBundle(lang)
-				bundles[lang] = bundle
-			}
+			dpath = filepath.FromSlash(dirPath.Dir())
 		}
 	}
 
-	if !isBundleHeader && bundle != nil {
-		bundle.resources = append(bundle.resources, info)
-	}
-
-	if classifier == files.ContentClassFile {
-		translations := info.Meta().Translations
-
-		for lang, b := range bundles {
-			if !stringSliceContains(lang, translations...) && !b.containsResource(info.Name()) {
-
-				// Clone and add it to the bundle.
-				clone := c.cloneFileInfo(info)
-				clone.Meta().Lang = lang
-				b.resources = append(b.resources, clone)
-			}
-		}
+	if c.seenDirs[dpath] {
+		return nil
 	}
+	c.seenDirs[dpath] = true
 
-	return nil
-}
-
-func (c *pagesCollector) cloneFileInfo(fi hugofs.FileMetaInfo) hugofs.FileMetaInfo {
-	return hugofs.NewFileMetaInfo(fi, hugofs.NewFileMeta())
-}
-
-func (c *pagesCollector) collectDir(dirname string, partial bool, inFilter func(fim hugofs.FileMetaInfo) bool) error {
-	fi, err := c.fs.Stat(dirname)
+	root, err := c.fs.Stat(dpath)
 	if err != nil {
-		if herrors.IsNotExist(err) {
-			// May have been deleted.
+		if os.IsNotExist(err) {
 			return nil
 		}
 		return err
 	}
 
-	handleDir := func(
-		btype bundleDirType,
-		dir hugofs.FileMetaInfo,
-		path string,
-		readdir []hugofs.FileMetaInfo) error {
-		if btype > bundleNot && c.tracker != nil {
-			c.tracker.add(path, btype)
-		}
-
-		if btype == bundleBranch {
-			if err := c.handleBundleBranch(readdir); err != nil {
-				return err
-			}
-			// A branch bundle is only this directory level, so keep walking.
-			return nil
-		} else if btype == bundleLeaf {
-			if err := c.handleBundleLeaf(dir, path, readdir); err != nil {
-				return err
-			}
-
-			return nil
-		}
-
-		if err := c.handleFiles(readdir...); err != nil {
-			return err
-		}
+	rootm := root.(hugofs.FileMetaInfo)
 
-		return nil
+	if err := c.collectDirDir(dpath, rootm, inFilter); err != nil {
+		return err
 	}
 
-	filter := func(fim hugofs.FileMetaInfo) bool {
-		if fim.Meta().SkipDir {
-			return false
-		}
+	return nil
+}
 
+func (c *pagesCollector) collectDirDir(path string, root hugofs.FileMetaInfo, inFilter func(fim hugofs.FileMetaInfo) bool) error {
+	filter := func(fim hugofs.FileMetaInfo) bool {
 		if c.sp.IgnoreFile(fim.Meta().Filename) {
 			return false
 		}
-
 		if inFilter != nil {
 			return inFilter(fim)
 		}
@@ -370,83 +259,63 @@ func (c *pagesCollector) collectDir(dirname string, partial bool, inFilter func(
 	}
 
 	preHook := func(dir hugofs.FileMetaInfo, path string, readdir []hugofs.FileMetaInfo) ([]hugofs.FileMetaInfo, error) {
-		var btype bundleDirType
-
 		filtered := readdir[:0]
 		for _, fi := range readdir {
 			if filter(fi) {
 				filtered = append(filtered, fi)
-
-				if c.tracker != nil {
-					// Track symlinks.
-					c.tracker.addSymbolicLinkMapping(fi)
-				}
 			}
 		}
-		walkRoot := dir.Meta().IsRootFile
 		readdir = filtered
+		if len(readdir) == 0 {
+			return nil, nil
+		}
 
-		// We merge language directories, so there can be duplicates, but they
-		// will be ordered, most important first.
-		var duplicates []int
-		seen := make(map[string]bool)
-
-		for i, fi := range readdir {
-
+		// Pick the first regular file.
+		var first hugofs.FileMetaInfo
+		for _, fi := range readdir {
 			if fi.IsDir() {
 				continue
 			}
+			first = fi
+			break
+		}
 
-			meta := fi.Meta()
-			meta.IsRootFile = walkRoot
-			class := meta.Classifier
-			translationBase := meta.TranslationBaseNameWithExt
-			key := pth.Join(meta.Lang, translationBase)
-
-			if seen[key] {
-				duplicates = append(duplicates, i)
-				continue
-			}
-			seen[key] = true
+		if first == nil {
+			// Only dirs, keep walking.
+			return readdir, nil
+		}
 
-			var thisBtype bundleDirType
+		// Any bundle file will always be first.
+		firstPi := first.Meta().PathInfo
+		if firstPi == nil {
+			panic(fmt.Sprintf("collectDirDir: no path info for %q", first.Meta().Filename))
+		}
 
-			switch class {
-			case files.ContentClassLeaf:
-				thisBtype = bundleLeaf
-			case files.ContentClassBranch:
-				thisBtype = bundleBranch
+		if firstPi.IsLeafBundle() {
+			if err := c.handleBundleLeaf(dir, first, path, readdir); err != nil {
+				return nil, err
 			}
+			return nil, filepath.SkipDir
+		}
 
-			// Folders with both index.md and _index.md type of files have
-			// undefined behaviour and can never work.
-			// The branch variant will win because of sort order, but log
-			// a warning about it.
-			if thisBtype > bundleNot && btype > bundleNot && thisBtype != btype {
-				c.logger.Warnf("Content directory %q have both index.* and _index.* files, pick one.", dir.Meta().Filename)
-				// Reclassify it so it will be handled as a content file inside the
-				// section, which is in line with the <= 0.55 behaviour.
-				meta.Classifier = files.ContentClassContent
-			} else if thisBtype > bundleNot {
-				btype = thisBtype
+		for _, fi := range readdir {
+			if fi.IsDir() {
+				continue
 			}
 
-		}
-
-		if len(duplicates) > 0 {
-			for i := len(duplicates) - 1; i >= 0; i-- {
-				idx := duplicates[i]
-				readdir = append(readdir[:idx], readdir[idx+1:]...)
+			meta := fi.Meta()
+			pi := meta.PathInfo
+			if pi == nil {
+				panic(fmt.Sprintf("no path info for %q", meta.Filename))
 			}
-		}
 
-		err := handleDir(btype, dir, path, readdir)
-		if err != nil {
-			return nil, err
-		}
+			if meta.Lang == "" {
+				panic("lang not set")
+			}
 
-		if btype == bundleLeaf || partial {
-			return nil, filepath.SkipDir
+			if err := c.g.Enqueue(fi); err != nil {
+				return nil, err
+			}
 		}
 
 		// Keep walking.
@@ -454,126 +323,56 @@ func (c *pagesCollector) collectDir(dirname string, partial bool, inFilter func(
 	}
 
 	var postHook hugofs.WalkHook
-	if c.tracker != nil {
-		postHook = func(dir hugofs.FileMetaInfo, path string, readdir []hugofs.FileMetaInfo) ([]hugofs.FileMetaInfo, error) {
-			if c.tracker == nil {
-				// Nothing to do.
-				return readdir, nil
-			}
-
-			return readdir, nil
-		}
-	}
-
-	wfn := func(path string, info hugofs.FileMetaInfo, err error) error {
-		if err != nil {
-			return err
-		}
 
+	wfn := func(path string, fi hugofs.FileMetaInfo) error {
 		return nil
 	}
 
-	fim := fi.(hugofs.FileMetaInfo)
-	// Make sure the pages in this directory gets re-rendered,
-	// even in fast render mode.
-	fim.Meta().IsRootFile = true
-
-	w := hugofs.NewWalkway(hugofs.WalkwayConfig{
-		Fs:       c.fs,
-		Logger:   c.logger,
-		Root:     dirname,
-		Info:     fim,
-		HookPre:  preHook,
-		HookPost: postHook,
-		WalkFn:   wfn,
-	})
+	w := hugofs.NewWalkway(
+		hugofs.WalkwayConfig{
+			Logger:   c.logger,
+			Root:     path,
+			Info:     root,
+			Fs:       c.fs,
+			HookPre:  preHook,
+			HookPost: postHook,
+			WalkFn:   wfn,
+		})
 
 	return w.Walk()
 }
 
-func (c *pagesCollector) handleBundleBranch(readdir []hugofs.FileMetaInfo) error {
-	// Maps bundles to its language.
-	bundles := pageBundles{}
-
-	var contentFiles []hugofs.FileMetaInfo
-
-	for _, fim := range readdir {
-
-		if fim.IsDir() {
-			continue
+func (c *pagesCollector) handleBundleLeaf(dir, bundle hugofs.FileMetaInfo, inPath string, readdir []hugofs.FileMetaInfo) error {
+	bundlePi := bundle.Meta().PathInfo
+	walk := func(path string, info hugofs.FileMetaInfo) error {
+		if info.IsDir() {
+			return nil
 		}
 
-		meta := fim.Meta()
+		pi := info.Meta().PathInfo
 
-		switch meta.Classifier {
-		case files.ContentClassContent:
-			contentFiles = append(contentFiles, fim)
-		default:
-			if err := c.addToBundle(fim, bundleBranch, bundles); err != nil {
-				return err
+		if info != bundle {
+			// Everything inside a leaf bundle is a Resource,
+			// even the content pages.
+			// Note that we do allow index.md as page resources, but not in the bundle root.
+			if !pi.IsLeafBundle() || pi.Dir() != bundlePi.Dir() {
+				paths.ModifyPathBundleTypeResource(pi)
 			}
 		}
 
-	}
-
-	// Make sure the section is created before its pages.
-	if err := c.proc.Process(bundles); err != nil {
-		return err
-	}
-
-	return c.handleFiles(contentFiles...)
-}
-
-func (c *pagesCollector) handleBundleLeaf(dir hugofs.FileMetaInfo, path string, readdir []hugofs.FileMetaInfo) error {
-	// Maps bundles to its language.
-	bundles := pageBundles{}
-
-	walk := func(path string, info hugofs.FileMetaInfo, err error) error {
-		if err != nil {
-			return err
-		}
-		if info.IsDir() {
-			return nil
-		}
-
-		return c.addToBundle(info, bundleLeaf, bundles)
+		return c.g.Enqueue(info)
 	}
 
 	// Start a new walker from the given path.
-	w := hugofs.NewWalkway(hugofs.WalkwayConfig{
-		Root:       path,
-		Fs:         c.fs,
-		Logger:     c.logger,
-		Info:       dir,
-		DirEntries: readdir,
-		WalkFn:     walk,
-	})
-
-	if err := w.Walk(); err != nil {
-		return err
-	}
-
-	return c.proc.Process(bundles)
-}
-
-func (c *pagesCollector) handleFiles(fis ...hugofs.FileMetaInfo) error {
-	for _, fi := range fis {
-		if fi.IsDir() {
-			continue
-		}
+	w := hugofs.NewWalkway(
+		hugofs.WalkwayConfig{
+			Root:       inPath,
+			Fs:         c.fs,
+			Logger:     c.logger,
+			Info:       dir,
+			DirEntries: readdir,
+			WalkFn:     walk,
+		})
 
-		if err := c.proc.Process(fi); err != nil {
-			return err
-		}
-	}
-	return nil
-}
-
-func stringSliceContains(k string, values ...string) bool {
-	for _, v := range values {
-		if k == v {
-			return true
-		}
-	}
-	return false
+	return w.Walk()
 }