Speed up GetPage

```
name                             old time/op    new time/op    delta
GetPage-10                          413ns ± 0%     287ns ± 1%  -30.47%  (p=0.029 n=4+4)
GetPageRegular/From_root-10         757ns ± 1%     706ns ± 1%   -6.75%  (p=0.029 n=4+4)
GetPageRegular/Page_relative-10     838ns ± 1%     786ns ± 1%   -6.16%  (p=0.029 n=4+4)

name                             old alloc/op   new alloc/op   delta
GetPage-10                           312B ± 0%       24B ± 0%  -92.31%  (p=0.029 n=4+4)
GetPageRegular/From_root-10          328B ± 0%      200B ± 0%  -39.02%  (p=0.029 n=4+4)
GetPageRegular/Page_relative-10      360B ± 0%      232B ± 0%  -35.56%  (p=0.029 n=4+4)

name                             old allocs/op  new allocs/op  delta
GetPage-10                           8.00 ± 0%      2.00 ± 0%  -75.00%  (p=0.029 n=4+4)
GetPageRegular/From_root-10          7.00 ± 0%      5.00 ± 0%  -28.57%  (p=0.029 n=4+4)
GetPageRegular/Page_relative-10      9.00 ± 0%      7.00 ± 0%  -22.22%  (p=0.029 n=4+4)
```

2 files changed, 37 insertions, 22 deletions

diff --git a/common/paths/pathparser.go b/common/paths/pathparser.go
index 898eee341..20f5ee30d 100644
--- a/common/paths/pathparser.go
+++ b/common/paths/pathparser.go
@@ -54,15 +54,27 @@ func NormalizePathStringBasic(s string) string {
 
 // ParseIdentity parses component c with path s into a StringIdentity.
 func (pp *PathParser) ParseIdentity(c, s string) identity.StringIdentity {
+	p := pp.parsePooled(c, s)
+	defer putPath(p)
+	return identity.StringIdentity(p.IdentifierBase())
+}
+
+// ParseBaseAndBaseNameNoIdentifier parses component c with path s into a base and a base name without any identifier.
+func (pp *PathParser) ParseBaseAndBaseNameNoIdentifier(c, s string) (string, string) {
+	p := pp.parsePooled(c, s)
+	defer putPath(p)
+	return p.Base(), p.BaseNameNoIdentifier()
+}
+
+func (pp *PathParser) parsePooled(c, s string) *Path {
 	s = NormalizePathStringBasic(s)
 	p := getPath()
 	p.component = c
-	defer putPath(p)
 	p, err := pp.doParse(c, s, p)
 	if err != nil {
 		panic(err)
 	}
-	return identity.StringIdentity(p.IdentifierBase())
+	return p
 }
 
 // Parse parses component c with path s into Path using Hugo's content path rules.
@@ -259,7 +271,9 @@ type Path struct {
 
 var pathPool = &sync.Pool{
 	New: func() any {
-		return &Path{}
+		p := &Path{}
+		p.reset()
+		return p
 	},
 }
 
@@ -268,6 +282,11 @@ func getPath() *Path {
 }
 
 func putPath(p *Path) {
+	p.reset()
+	pathPool.Put(p)
+}
+
+func (p *Path) reset() {
 	p.s = ""
 	p.posContainerLow = -1
 	p.posContainerHigh = -1
@@ -279,7 +298,6 @@ func putPath(p *Path) {
 	p.disabled = false
 	p.trimLeadingSlash = false
 	p.unnormalized = nil
-	pathPool.Put(p)
 }
 
 // TrimLeadingSlash returns a copy of the Path with the leading slash removed.
diff --git a/hugolib/pagecollections.go b/hugolib/pagecollections.go
index fe7c2f1e7..5ed2a28f0 100644
--- a/hugolib/pagecollections.go
+++ b/hugolib/pagecollections.go
@@ -131,22 +131,21 @@ func (c *pageFinder) getContentNode(context page.Page, isReflink bool, ref strin
 		return c.getContentNodeForRef(context, isReflink, true, inRef, ref)
 	}
 
-	var refs []string
-
 	// We are always looking for a content file and having an extension greatly simplifies the code that follows,
 	// even in the case where the extension does not match this one.
 	if ref == "/" {
-		refs = append(refs, "/_index"+defaultContentExt)
+		if n, err := c.getContentNodeForRef(context, isReflink, false, inRef, "/_index"+defaultContentExt); n != nil || err != nil {
+			return n, err
+		}
 	} else if strings.HasSuffix(ref, "/index") {
-		refs = append(refs, ref+"/index"+defaultContentExt)
-		refs = append(refs, ref+defaultContentExt)
+		if n, err := c.getContentNodeForRef(context, isReflink, false, inRef, ref+"/index"+defaultContentExt); n != nil || err != nil {
+			return n, err
+		}
+		if n, err := c.getContentNodeForRef(context, isReflink, false, inRef, ref+defaultContentExt); n != nil || err != nil {
+			return n, err
+		}
 	} else {
-		refs = append(refs, ref+defaultContentExt)
-	}
-
-	for _, ref := range refs {
-		n, err := c.getContentNodeForRef(context, isReflink, false, inRef, ref)
-		if n != nil || err != nil {
+		if n, err := c.getContentNodeForRef(context, isReflink, false, inRef, ref+defaultContentExt); n != nil || err != nil {
 			return n, err
 		}
 	}
@@ -175,7 +174,7 @@ func (c *pageFinder) getContentNodeForRef(context page.Page, isReflink, hadExten
 
 		rel := path.Join(baseDir, ref)
 
-		relPath := contentPathParser.Parse(files.ComponentFolderContent, rel)
+		relPath, _ := contentPathParser.ParseBaseAndBaseNameNoIdentifier(files.ComponentFolderContent, rel)
 
 		n, err := c.getContentNodeFromPath(relPath, ref)
 		if n != nil || err != nil {
@@ -195,9 +194,9 @@ func (c *pageFinder) getContentNodeForRef(context page.Page, isReflink, hadExten
 		return nil, nil
 	}
 
-	refPath := contentPathParser.Parse(files.ComponentFolderContent, ref)
+	relPath, nameNoIdentifier := contentPathParser.ParseBaseAndBaseNameNoIdentifier(files.ComponentFolderContent, ref)
 
-	n, err := c.getContentNodeFromPath(refPath, ref)
+	n, err := c.getContentNodeFromPath(relPath, ref)
 
 	if n != nil || err != nil {
 		return n, err
@@ -221,7 +220,7 @@ func (c *pageFinder) getContentNodeForRef(context page.Page, isReflink, hadExten
 		return nil, nil
 	}
 
-	n = c.pageMap.pageReverseIndex.Get(refPath.BaseNameNoIdentifier())
+	n = c.pageMap.pageReverseIndex.Get(nameNoIdentifier)
 	if n == ambiguousContentNode {
 		return nil, fmt.Errorf("page reference %q is ambiguous", inRef)
 	}
@@ -254,9 +253,7 @@ func (c *pageFinder) getContentNodeFromRefReverseLookup(ref string, fi hugofs.Fi
 	return nil, nil
 }
 
-func (c *pageFinder) getContentNodeFromPath(refPath *paths.Path, ref string) (contentNodeI, error) {
-	s := refPath.Base()
-
+func (c *pageFinder) getContentNodeFromPath(s string, ref string) (contentNodeI, error) {
 	n := c.pageMap.treePages.Get(s)
 	if n != nil {
 		return n, nil