Split out the puthe path/filepath functions into common/paths

So they can be used from the config package without cyclic troubles.

Updates #8654

common/paths/path.go

@@ -0,0 +1,312 @@
+// 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.
+// You may obtain a copy of the License at
+// http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+package paths
+
+import (
+	"errors"
+	"fmt"
+	"os"
+	"path"
+	"path/filepath"
+	"regexp"
+	"strings"
+)
+
+// FilePathSeparator as defined by os.Separator.
+const FilePathSeparator = string(filepath.Separator)
+
+// filepathPathBridge is a bridge for common functionality in filepath vs path
+type filepathPathBridge interface {
+	Base(in string) string
+	Clean(in string) string
+	Dir(in string) string
+	Ext(in string) string
+	Join(elem ...string) string
+	Separator() string
+}
+
+type filepathBridge struct {
+}
+
+func (filepathBridge) Base(in string) string {
+	return filepath.Base(in)
+}
+
+func (filepathBridge) Clean(in string) string {
+	return filepath.Clean(in)
+}
+
+func (filepathBridge) Dir(in string) string {
+	return filepath.Dir(in)
+}
+
+func (filepathBridge) Ext(in string) string {
+	return filepath.Ext(in)
+}
+
+func (filepathBridge) Join(elem ...string) string {
+	return filepath.Join(elem...)
+}
+
+func (filepathBridge) Separator() string {
+	return FilePathSeparator
+}
+
+var fpb filepathBridge
+
+// ToSlashTrimLeading is just a filepath.ToSlash with an added / prefix trimmer.
+func ToSlashTrimLeading(s string) string {
+	return strings.TrimPrefix(filepath.ToSlash(s), "/")
+}
+
+// MakeTitle converts the path given to a suitable title, trimming whitespace
+// and replacing hyphens with whitespace.
+func MakeTitle(inpath string) string {
+	return strings.Replace(strings.TrimSpace(inpath), "-", " ", -1)
+}
+
+// ReplaceExtension takes a path and an extension, strips the old extension
+// and returns the path with the new extension.
+func ReplaceExtension(path string, newExt string) string {
+	f, _ := fileAndExt(path, fpb)
+	return f + "." + newExt
+}
+
+func makePathRelative(inPath string, possibleDirectories ...string) (string, error) {
+	for _, currentPath := range possibleDirectories {
+		if strings.HasPrefix(inPath, currentPath) {
+			return strings.TrimPrefix(inPath, currentPath), nil
+		}
+	}
+	return inPath, errors.New("can't extract relative path, unknown prefix")
+}
+
+// Should be good enough for Hugo.
+var isFileRe = regexp.MustCompile(`.*\..{1,6}$`)
+
+// GetDottedRelativePath expects a relative path starting after the content directory.
+// It returns a relative path with dots ("..") navigating up the path structure.
+func GetDottedRelativePath(inPath string) string {
+	inPath = filepath.Clean(filepath.FromSlash(inPath))
+
+	if inPath == "." {
+		return "./"
+	}
+
+	if !isFileRe.MatchString(inPath) && !strings.HasSuffix(inPath, FilePathSeparator) {
+		inPath += FilePathSeparator
+	}
+
+	if !strings.HasPrefix(inPath, FilePathSeparator) {
+		inPath = FilePathSeparator + inPath
+	}
+
+	dir, _ := filepath.Split(inPath)
+
+	sectionCount := strings.Count(dir, FilePathSeparator)
+
+	if sectionCount == 0 || dir == FilePathSeparator {
+		return "./"
+	}
+
+	var dottedPath string
+
+	for i := 1; i < sectionCount; i++ {
+		dottedPath += "../"
+	}
+
+	return dottedPath
+}
+
+// ExtNoDelimiter takes a path and returns the extension, excluding the delimiter, i.e. "md".
+func ExtNoDelimiter(in string) string {
+	return strings.TrimPrefix(Ext(in), ".")
+}
+
+// Ext takes a path and returns the extension, including the delimiter, i.e. ".md".
+func Ext(in string) string {
+	_, ext := fileAndExt(in, fpb)
+	return ext
+}
+
+// PathAndExt is the same as FileAndExt, but it uses the path package.
+func PathAndExt(in string) (string, string) {
+	return fileAndExt(in, pb)
+}
+
+// FileAndExt takes a path and returns the file and extension separated,
+// the extension including the delimiter, i.e. ".md".
+func FileAndExt(in string) (string, string) {
+	return fileAndExt(in, fpb)
+}
+
+// FileAndExtNoDelimiter takes a path and returns the file and extension separated,
+// the extension excluding the delimiter, e.g "md".
+func FileAndExtNoDelimiter(in string) (string, string) {
+	file, ext := fileAndExt(in, fpb)
+	return file, strings.TrimPrefix(ext, ".")
+}
+
+// Filename takes a file path, strips out the extension,
+// and returns the name of the file.
+func Filename(in string) (name string) {
+	name, _ = fileAndExt(in, fpb)
+	return
+}
+
+// PathNoExt takes a path, strips out the extension,
+// and returns the name of the file.
+func PathNoExt(in string) string {
+	return strings.TrimSuffix(in, path.Ext(in))
+}
+
+// FileAndExt returns the filename and any extension of a file path as
+// two separate strings.
+//
+// If the path, in, contains a directory name ending in a slash,
+// then both name and ext will be empty strings.
+//
+// If the path, in, is either the current directory, the parent
+// directory or the root directory, or an empty string,
+// then both name and ext will be empty strings.
+//
+// If the path, in, represents the path of a file without an extension,
+// then name will be the name of the file and ext will be an empty string.
+//
+// If the path, in, represents a filename with an extension,
+// then name will be the filename minus any extension - including the dot
+// and ext will contain the extension - minus the dot.
+func fileAndExt(in string, b filepathPathBridge) (name string, ext string) {
+	ext = b.Ext(in)
+	base := b.Base(in)
+
+	return extractFilename(in, ext, base, b.Separator()), ext
+}
+
+func extractFilename(in, ext, base, pathSeparator string) (name string) {
+	// No file name cases. These are defined as:
+	// 1. any "in" path that ends in a pathSeparator
+	// 2. any "base" consisting of just an pathSeparator
+	// 3. any "base" consisting of just an empty string
+	// 4. any "base" consisting of just the current directory i.e. "."
+	// 5. any "base" consisting of just the parent directory i.e. ".."
+	if (strings.LastIndex(in, pathSeparator) == len(in)-1) || base == "" || base == "." || base == ".." || base == pathSeparator {
+		name = "" // there is NO filename
+	} else if ext != "" { // there was an Extension
+		// return the filename minus the extension (and the ".")
+		name = base[:strings.LastIndex(base, ".")]
+	} else {
+		// no extension case so just return base, which willi
+		// be the filename
+		name = base
+	}
+	return
+}
+
+// GetRelativePath returns the relative path of a given path.
+func GetRelativePath(path, base string) (final string, err error) {
+	if filepath.IsAbs(path) && base == "" {
+		return "", errors.New("source: missing base directory")
+	}
+	name := filepath.Clean(path)
+	base = filepath.Clean(base)
+
+	name, err = filepath.Rel(base, name)
+	if err != nil {
+		return "", err
+	}
+
+	if strings.HasSuffix(filepath.FromSlash(path), FilePathSeparator) && !strings.HasSuffix(name, FilePathSeparator) {
+		name += FilePathSeparator
+	}
+	return name, nil
+}
+
+// PathPrep prepares the path using the uglify setting to create paths on
+// either the form /section/name/index.html or /section/name.html.
+func PathPrep(ugly bool, in string) string {
+	if ugly {
+		return Uglify(in)
+	}
+	return PrettifyPath(in)
+}
+
+// PrettifyPath is the same as PrettifyURLPath but for file paths.
+//     /section/name.html       becomes /section/name/index.html
+//     /section/name/           becomes /section/name/index.html
+//     /section/name/index.html becomes /section/name/index.html
+func PrettifyPath(in string) string {
+	return prettifyPath(in, fpb)
+}
+
+func prettifyPath(in string, b filepathPathBridge) string {
+	if filepath.Ext(in) == "" {
+		// /section/name/  -> /section/name/index.html
+		if len(in) < 2 {
+			return b.Separator()
+		}
+		return b.Join(in, "index.html")
+	}
+	name, ext := fileAndExt(in, b)
+	if name == "index" {
+		// /section/name/index.html -> /section/name/index.html
+		return b.Clean(in)
+	}
+	// /section/name.html -> /section/name/index.html
+	return b.Join(b.Dir(in), name, "index"+ext)
+}
+
+type NamedSlice struct {
+	Name  string
+	Slice []string
+}
+
+func (n NamedSlice) String() string {
+	if len(n.Slice) == 0 {
+		return n.Name
+	}
+	return fmt.Sprintf("%s%s{%s}", n.Name, FilePathSeparator, strings.Join(n.Slice, ","))
+}
+
+// FindCWD returns the current working directory from where the Hugo
+// executable is run.
+func FindCWD() (string, error) {
+	serverFile, err := filepath.Abs(os.Args[0])
+	if err != nil {
+		return "", fmt.Errorf("can't get absolute path for executable: %v", err)
+	}
+
+	path := filepath.Dir(serverFile)
+	realFile, err := filepath.EvalSymlinks(serverFile)
+	if err != nil {
+		if _, err = os.Stat(serverFile + ".exe"); err == nil {
+			realFile = filepath.Clean(serverFile + ".exe")
+		}
+	}
+
+	if err == nil && realFile != serverFile {
+		path = filepath.Dir(realFile)
+	}
+
+	return path, nil
+}
+
+// AddTrailingSlash adds a trailing Unix styled slash (/) if not already
+// there.
+func AddTrailingSlash(path string) string {
+	if !strings.HasSuffix(path, "/") {
+		path += "/"
+	}
+	return path
+}

common/paths/path_test.go

@@ -0,0 +1,252 @@
+// 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.
+// You may obtain a copy of the License at
+// http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+package paths
+
+import (
+	"path/filepath"
+	"testing"
+
+	qt "github.com/frankban/quicktest"
+)
+
+func TestGetRelativePath(t *testing.T) {
+	tests := []struct {
+		path   string
+		base   string
+		expect interface{}
+	}{
+		{filepath.FromSlash("/a/b"), filepath.FromSlash("/a"), filepath.FromSlash("b")},
+		{filepath.FromSlash("/a/b/c/"), filepath.FromSlash("/a"), filepath.FromSlash("b/c/")},
+		{filepath.FromSlash("/c"), filepath.FromSlash("/a/b"), filepath.FromSlash("../../c")},
+		{filepath.FromSlash("/c"), "", false},
+	}
+	for i, this := range tests {
+		// ultimately a fancy wrapper around filepath.Rel
+		result, err := GetRelativePath(this.path, this.base)
+
+		if b, ok := this.expect.(bool); ok && !b {
+			if err == nil {
+				t.Errorf("[%d] GetRelativePath didn't return an expected error", i)
+			}
+		} else {
+			if err != nil {
+				t.Errorf("[%d] GetRelativePath failed: %s", i, err)
+				continue
+			}
+			if result != this.expect {
+				t.Errorf("[%d] GetRelativePath got %v but expected %v", i, result, this.expect)
+			}
+		}
+
+	}
+}
+
+func TestMakePathRelative(t *testing.T) {
+	type test struct {
+		inPath, path1, path2, output string
+	}
+
+	data := []test{
+		{"/abc/bcd/ab.css", "/abc/bcd", "/bbc/bcd", "/ab.css"},
+		{"/abc/bcd/ab.css", "/abcd/bcd", "/abc/bcd", "/ab.css"},
+	}
+
+	for i, d := range data {
+		output, _ := makePathRelative(d.inPath, d.path1, d.path2)
+		if d.output != output {
+			t.Errorf("Test #%d failed. Expected %q got %q", i, d.output, output)
+		}
+	}
+	_, error := makePathRelative("a/b/c.ss", "/a/c", "/d/c", "/e/f")
+
+	if error == nil {
+		t.Errorf("Test failed, expected error")
+	}
+}
+
+func TestGetDottedRelativePath(t *testing.T) {
+	// on Windows this will receive both kinds, both country and western ...
+	for _, f := range []func(string) string{filepath.FromSlash, func(s string) string { return s }} {
+		doTestGetDottedRelativePath(f, t)
+	}
+}
+
+func doTestGetDottedRelativePath(urlFixer func(string) string, t *testing.T) {
+	type test struct {
+		input, expected string
+	}
+	data := []test{
+		{"", "./"},
+		{urlFixer("/"), "./"},
+		{urlFixer("post"), "../"},
+		{urlFixer("/post"), "../"},
+		{urlFixer("post/"), "../"},
+		{urlFixer("tags/foo.html"), "../"},
+		{urlFixer("/tags/foo.html"), "../"},
+		{urlFixer("/post/"), "../"},
+		{urlFixer("////post/////"), "../"},
+		{urlFixer("/foo/bar/index.html"), "../../"},
+		{urlFixer("/foo/bar/foo/"), "../../../"},
+		{urlFixer("/foo/bar/foo"), "../../../"},
+		{urlFixer("foo/bar/foo/"), "../../../"},
+		{urlFixer("foo/bar/foo/bar"), "../../../../"},
+		{"404.html", "./"},
+		{"404.xml", "./"},
+		{"/404.html", "./"},
+	}
+	for i, d := range data {
+		output := GetDottedRelativePath(d.input)
+		if d.expected != output {
+			t.Errorf("Test %d failed. Expected %q got %q", i, d.expected, output)
+		}
+	}
+}
+
+func TestMakeTitle(t *testing.T) {
+	type test struct {
+		input, expected string
+	}
+	data := []test{
+		{"Make-Title", "Make Title"},
+		{"MakeTitle", "MakeTitle"},
+		{"make_title", "make_title"},
+	}
+	for i, d := range data {
+		output := MakeTitle(d.input)
+		if d.expected != output {
+			t.Errorf("Test %d failed. Expected %q got %q", i, d.expected, output)
+		}
+	}
+}
+
+// Replace Extension is probably poorly named, but the intent of the
+// function is to accept a path and return only the file name with a
+// new extension. It's intentionally designed to strip out the path
+// and only provide the name. We should probably rename the function to
+// be more explicit at some point.
+func TestReplaceExtension(t *testing.T) {
+	type test struct {
+		input, newext, expected string
+	}
+	data := []test{
+		// These work according to the above definition
+		{"/some/random/path/file.xml", "html", "file.html"},
+		{"/banana.html", "xml", "banana.xml"},
+		{"./banana.html", "xml", "banana.xml"},
+		{"banana/pie/index.html", "xml", "index.xml"},
+		{"../pies/fish/index.html", "xml", "index.xml"},
+		// but these all fail
+		{"filename-without-an-ext", "ext", "filename-without-an-ext.ext"},
+		{"/filename-without-an-ext", "ext", "filename-without-an-ext.ext"},
+		{"/directory/mydir/", "ext", ".ext"},
+		{"mydir/", "ext", ".ext"},
+	}
+
+	for i, d := range data {
+		output := ReplaceExtension(filepath.FromSlash(d.input), d.newext)
+		if d.expected != output {
+			t.Errorf("Test %d failed. Expected %q got %q", i, d.expected, output)
+		}
+	}
+}
+
+func TestExtNoDelimiter(t *testing.T) {
+	c := qt.New(t)
+	c.Assert(ExtNoDelimiter(filepath.FromSlash("/my/data.json")), qt.Equals, "json")
+}
+
+func TestFilename(t *testing.T) {
+	type test struct {
+		input, expected string
+	}
+	data := []test{
+		{"index.html", "index"},
+		{"./index.html", "index"},
+		{"/index.html", "index"},
+		{"index", "index"},
+		{"/tmp/index.html", "index"},
+		{"./filename-no-ext", "filename-no-ext"},
+		{"/filename-no-ext", "filename-no-ext"},
+		{"filename-no-ext", "filename-no-ext"},
+		{"directory/", ""}, // no filename case??
+		{"directory/.hidden.ext", ".hidden"},
+		{"./directory/../~/banana/gold.fish", "gold"},
+		{"../directory/banana.man", "banana"},
+		{"~/mydir/filename.ext", "filename"},
+		{"./directory//tmp/filename.ext", "filename"},
+	}
+
+	for i, d := range data {
+		output := Filename(filepath.FromSlash(d.input))
+		if d.expected != output {
+			t.Errorf("Test %d failed. Expected %q got %q", i, d.expected, output)
+		}
+	}
+}
+
+func TestFileAndExt(t *testing.T) {
+	type test struct {
+		input, expectedFile, expectedExt string
+	}
+	data := []test{
+		{"index.html", "index", ".html"},
+		{"./index.html", "index", ".html"},
+		{"/index.html", "index", ".html"},
+		{"index", "index", ""},
+		{"/tmp/index.html", "index", ".html"},
+		{"./filename-no-ext", "filename-no-ext", ""},
+		{"/filename-no-ext", "filename-no-ext", ""},
+		{"filename-no-ext", "filename-no-ext", ""},
+		{"directory/", "", ""}, // no filename case??
+		{"directory/.hidden.ext", ".hidden", ".ext"},
+		{"./directory/../~/banana/gold.fish", "gold", ".fish"},
+		{"../directory/banana.man", "banana", ".man"},
+		{"~/mydir/filename.ext", "filename", ".ext"},
+		{"./directory//tmp/filename.ext", "filename", ".ext"},
+	}
+
+	for i, d := range data {
+		file, ext := fileAndExt(filepath.FromSlash(d.input), fpb)
+		if d.expectedFile != file {
+			t.Errorf("Test %d failed. Expected filename %q got %q.", i, d.expectedFile, file)
+		}
+		if d.expectedExt != ext {
+			t.Errorf("Test %d failed. Expected extension %q got %q.", i, d.expectedExt, ext)
+		}
+	}
+}
+
+func TestFindCWD(t *testing.T) {
+	type test struct {
+		expectedDir string
+		expectedErr error
+	}
+
+	// cwd, _ := os.Getwd()
+	data := []test{
+		//{cwd, nil},
+		// Commenting this out. It doesn't work properly.
+		// There's a good reason why we don't use os.Getwd(), it doesn't actually work the way we want it to.
+		// I really don't know a better way to test this function. - SPF 2014.11.04
+	}
+	for i, d := range data {
+		dir, err := FindCWD()
+		if d.expectedDir != dir {
+			t.Errorf("Test %d failed. Expected %q but got %q", i, d.expectedDir, dir)
+		}
+		if d.expectedErr != err {
+			t.Errorf("Test %d failed. Expected %q but got %q", i, d.expectedErr, err)
+		}
+	}
+}

common/paths/url.go

@@ -0,0 +1,212 @@
+// 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.
+// You may obtain a copy of the License at
+// http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+package paths
+
+import (
+	"fmt"
+	"net/url"
+	"path"
+	"strings"
+
+	"github.com/PuerkitoBio/purell"
+)
+
+type pathBridge struct {
+}
+
+func (pathBridge) Base(in string) string {
+	return path.Base(in)
+}
+
+func (pathBridge) Clean(in string) string {
+	return path.Clean(in)
+}
+
+func (pathBridge) Dir(in string) string {
+	return path.Dir(in)
+}
+
+func (pathBridge) Ext(in string) string {
+	return path.Ext(in)
+}
+
+func (pathBridge) Join(elem ...string) string {
+	return path.Join(elem...)
+}
+
+func (pathBridge) Separator() string {
+	return "/"
+}
+
+var pb pathBridge
+
+func sanitizeURLWithFlags(in string, f purell.NormalizationFlags) string {
+	s, err := purell.NormalizeURLString(in, f)
+	if err != nil {
+		return in
+	}
+
+	// Temporary workaround for the bug fix and resulting
+	// behavioral change in purell.NormalizeURLString():
+	// a leading '/' was inadvertently added to relative links,
+	// but no longer, see #878.
+	//
+	// I think the real solution is to allow Hugo to
+	// make relative URL with relative path,
+	// e.g. "../../post/hello-again/", as wished by users
+	// in issues #157, #622, etc., without forcing
+	// relative URLs to begin with '/'.
+	// Once the fixes are in, let's remove this kludge
+	// and restore SanitizeURL() to the way it was.
+	//                         -- @anthonyfok, 2015-02-16
+	//
+	// Begin temporary kludge
+	u, err := url.Parse(s)
+	if err != nil {
+		panic(err)
+	}
+	if len(u.Path) > 0 && !strings.HasPrefix(u.Path, "/") {
+		u.Path = "/" + u.Path
+	}
+	return u.String()
+	// End temporary kludge
+
+	// return s
+
+}
+
+// SanitizeURL sanitizes the input URL string.
+func SanitizeURL(in string) string {
+	return sanitizeURLWithFlags(in, purell.FlagsSafe|purell.FlagRemoveTrailingSlash|purell.FlagRemoveDotSegments|purell.FlagRemoveDuplicateSlashes|purell.FlagRemoveUnnecessaryHostDots|purell.FlagRemoveEmptyPortSeparator)
+}
+
+// SanitizeURLKeepTrailingSlash is the same as SanitizeURL, but will keep any trailing slash.
+func SanitizeURLKeepTrailingSlash(in string) string {
+	return sanitizeURLWithFlags(in, purell.FlagsSafe|purell.FlagRemoveDotSegments|purell.FlagRemoveDuplicateSlashes|purell.FlagRemoveUnnecessaryHostDots|purell.FlagRemoveEmptyPortSeparator)
+}
+
+// MakePermalink combines base URL with content path to create full URL paths.
+// Example
+//    base:   http://spf13.com/
+//    path:   post/how-i-blog
+//    result: http://spf13.com/post/how-i-blog
+func MakePermalink(host, plink string) *url.URL {
+	base, err := url.Parse(host)
+	if err != nil {
+		panic(err)
+	}
+
+	p, err := url.Parse(plink)
+	if err != nil {
+		panic(err)
+	}
+
+	if p.Host != "" {
+		panic(fmt.Errorf("can't make permalink from absolute link %q", plink))
+	}
+
+	base.Path = path.Join(base.Path, p.Path)
+
+	// path.Join will strip off the last /, so put it back if it was there.
+	hadTrailingSlash := (plink == "" && strings.HasSuffix(host, "/")) || strings.HasSuffix(p.Path, "/")
+	if hadTrailingSlash && !strings.HasSuffix(base.Path, "/") {
+		base.Path = base.Path + "/"
+	}
+
+	return base
+}
+
+// IsAbsURL determines whether the given path points to an absolute URL.
+func IsAbsURL(path string) bool {
+	url, err := url.Parse(path)
+	if err != nil {
+		return false
+	}
+
+	return url.IsAbs() || strings.HasPrefix(path, "//")
+}
+
+// AddContextRoot adds the context root to an URL if it's not already set.
+// For relative URL entries on sites with a base url with a context root set (i.e. http://example.com/mysite),
+// relative URLs must not include the context root if canonifyURLs is enabled. But if it's disabled, it must be set.
+func AddContextRoot(baseURL, relativePath string) string {
+	url, err := url.Parse(baseURL)
+	if err != nil {
+		panic(err)
+	}
+
+	newPath := path.Join(url.Path, relativePath)
+
+	// path strips trailing slash, ignore root path.
+	if newPath != "/" && strings.HasSuffix(relativePath, "/") {
+		newPath += "/"
+	}
+	return newPath
+}
+
+// URLizeAn
+
+// PrettifyURL takes a URL string and returns a semantic, clean URL.
+func PrettifyURL(in string) string {
+	x := PrettifyURLPath(in)
+
+	if path.Base(x) == "index.html" {
+		return path.Dir(x)
+	}
+
+	if in == "" {
+		return "/"
+	}
+
+	return x
+}
+
+// PrettifyURLPath takes a URL path to a content and converts it
+// to enable pretty URLs.
+//     /section/name.html       becomes /section/name/index.html
+//     /section/name/           becomes /section/name/index.html
+//     /section/name/index.html becomes /section/name/index.html
+func PrettifyURLPath(in string) string {
+	return prettifyPath(in, pb)
+}
+
+// Uglify does the opposite of PrettifyURLPath().
+//     /section/name/index.html becomes /section/name.html
+//     /section/name/           becomes /section/name.html
+//     /section/name.html       becomes /section/name.html
+func Uglify(in string) string {
+	if path.Ext(in) == "" {
+		if len(in) < 2 {
+			return "/"
+		}
+		// /section/name/  -> /section/name.html
+		return path.Clean(in) + ".html"
+	}
+
+	name, ext := fileAndExt(in, pb)
+	if name == "index" {
+		// /section/name/index.html -> /section/name.html
+		d := path.Dir(in)
+		if len(d) > 1 {
+			return d + ext
+		}
+		return in
+	}
+	// /.xml -> /index.xml
+	if name == "" {
+		return path.Dir(in) + "index" + ext
+	}
+	// /section/name.html -> /section/name.html
+	return path.Clean(in)
+}

common/paths/url_test.go

@@ -0,0 +1,129 @@
+// 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.
+// You may obtain a copy of the License at
+// http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+package paths
+
+import (
+	"strings"
+	"testing"
+
+	qt "github.com/frankban/quicktest"
+)
+
+func TestSanitizeURL(t *testing.T) {
+	tests := []struct {
+		input    string
+		expected string
+	}{
+		{"http://foo.bar/", "http://foo.bar"},
+		{"http://foo.bar", "http://foo.bar"},          // issue #1105
+		{"http://foo.bar/zoo/", "http://foo.bar/zoo"}, // issue #931
+	}
+
+	for i, test := range tests {
+		o1 := SanitizeURL(test.input)
+		o2 := SanitizeURLKeepTrailingSlash(test.input)
+
+		expected2 := test.expected
+
+		if strings.HasSuffix(test.input, "/") && !strings.HasSuffix(expected2, "/") {
+			expected2 += "/"
+		}
+
+		if o1 != test.expected {
+			t.Errorf("[%d] 1: Expected %#v, got %#v\n", i, test.expected, o1)
+		}
+		if o2 != expected2 {
+			t.Errorf("[%d] 2: Expected %#v, got %#v\n", i, expected2, o2)
+		}
+	}
+}
+
+func TestMakePermalink(t *testing.T) {
+	type test struct {
+		host, link, output string
+	}
+
+	data := []test{
+		{"http://abc.com/foo", "post/bar", "http://abc.com/foo/post/bar"},
+		{"http://abc.com/foo/", "post/bar", "http://abc.com/foo/post/bar"},
+		{"http://abc.com", "post/bar", "http://abc.com/post/bar"},
+		{"http://abc.com", "bar", "http://abc.com/bar"},
+		{"http://abc.com/foo/bar", "post/bar", "http://abc.com/foo/bar/post/bar"},
+		{"http://abc.com/foo/bar", "post/bar/", "http://abc.com/foo/bar/post/bar/"},
+	}
+
+	for i, d := range data {
+		output := MakePermalink(d.host, d.link).String()
+		if d.output != output {
+			t.Errorf("Test #%d failed. Expected %q got %q", i, d.output, output)
+		}
+	}
+}
+
+func TestAddContextRoot(t *testing.T) {
+	tests := []struct {
+		baseURL  string
+		url      string
+		expected string
+	}{
+		{"http://example.com/sub/", "/foo", "/sub/foo"},
+		{"http://example.com/sub/", "/foo/index.html", "/sub/foo/index.html"},
+		{"http://example.com/sub1/sub2", "/foo", "/sub1/sub2/foo"},
+		{"http://example.com", "/foo", "/foo"},
+		// cannot guess that the context root is already added int the example below
+		{"http://example.com/sub/", "/sub/foo", "/sub/sub/foo"},
+		{"http://example.com/тря", "/трям/", "/тря/трям/"},
+		{"http://example.com", "/", "/"},
+		{"http://example.com/bar", "//", "/bar/"},
+	}
+
+	for _, test := range tests {
+		output := AddContextRoot(test.baseURL, test.url)
+		if output != test.expected {
+			t.Errorf("Expected %#v, got %#v\n", test.expected, output)
+		}
+	}
+}
+
+func TestPretty(t *testing.T) {
+	c := qt.New(t)
+	c.Assert("/section/name/index.html", qt.Equals, PrettifyURLPath("/section/name.html"))
+	c.Assert("/section/sub/name/index.html", qt.Equals, PrettifyURLPath("/section/sub/name.html"))
+	c.Assert("/section/name/index.html", qt.Equals, PrettifyURLPath("/section/name/"))
+	c.Assert("/section/name/index.html", qt.Equals, PrettifyURLPath("/section/name/index.html"))
+	c.Assert("/index.html", qt.Equals, PrettifyURLPath("/index.html"))
+	c.Assert("/name/index.xml", qt.Equals, PrettifyURLPath("/name.xml"))
+	c.Assert("/", qt.Equals, PrettifyURLPath("/"))
+	c.Assert("/", qt.Equals, PrettifyURLPath(""))
+	c.Assert("/section/name", qt.Equals, PrettifyURL("/section/name.html"))
+	c.Assert("/section/sub/name", qt.Equals, PrettifyURL("/section/sub/name.html"))
+	c.Assert("/section/name", qt.Equals, PrettifyURL("/section/name/"))
+	c.Assert("/section/name", qt.Equals, PrettifyURL("/section/name/index.html"))
+	c.Assert("/", qt.Equals, PrettifyURL("/index.html"))
+	c.Assert("/name/index.xml", qt.Equals, PrettifyURL("/name.xml"))
+	c.Assert("/", qt.Equals, PrettifyURL("/"))
+	c.Assert("/", qt.Equals, PrettifyURL(""))
+}
+
+func TestUgly(t *testing.T) {
+	c := qt.New(t)
+	c.Assert("/section/name.html", qt.Equals, Uglify("/section/name.html"))
+	c.Assert("/section/sub/name.html", qt.Equals, Uglify("/section/sub/name.html"))
+	c.Assert("/section/name.html", qt.Equals, Uglify("/section/name/"))
+	c.Assert("/section/name.html", qt.Equals, Uglify("/section/name/index.html"))
+	c.Assert("/index.html", qt.Equals, Uglify("/index.html"))
+	c.Assert("/name.xml", qt.Equals, Uglify("/name.xml"))
+	c.Assert("/", qt.Equals, Uglify("/"))
+	c.Assert("/", qt.Equals, Uglify(""))
+}