From 31b4ffca8977224993a874463c33cf38a600e912 Mon Sep 17 00:00:00 2001
From: Jacky Zhao <j.zhao2k19@gmail.com>
Date: Mon, 27 Dec 2021 20:46:47 -0500
Subject: [PATCH] better char escaping

---
 parse.go |  6 ++++--
 util.go  | 40 ++++++++++++++++++++++++++++++++++++++++
 walk.go  |  2 +-
 3 files changed, 45 insertions(+), 3 deletions(-)

diff --git a/parse.go b/parse.go
index 0583c6b..e85f778 100644
--- a/parse.go
+++ b/parse.go
@@ -40,10 +40,12 @@ func parse(dir, pathPrefix string) []Link {
 		target = strings.TrimSpace(target)
 		target = strings.Replace(target, " ", "-", -1)
 
+		source := filepath.ToSlash(hugoPathTrim(trim(dir, pathPrefix, ".md")))
+
 		fmt.Printf("  '%s' => %s\n", text, target)
 		links = append(links, Link{
-			Source: filepath.ToSlash(hugoPathTrim(trim(dir, pathPrefix, ".md"))),
-			Target: target,
+			Source: UnicodeSanitize(source),
+			Target: UnicodeSanitize(target),
 			Text:   text,
 		})
 		n++
diff --git a/util.go b/util.go
index ab93bd6..70e3716 100644
--- a/util.go
+++ b/util.go
@@ -4,6 +4,7 @@ import (
 	"fmt"
 	"path/filepath"
 	"strings"
+	"unicode"
 )
 
 func trim(source, prefix, suffix string) string {
@@ -28,6 +29,45 @@ func isInternal(link string) bool {
 	return !strings.HasPrefix(link, "http")
 }
 
+// From https://golang.org/src/net/url/url.go
+func ishex(c rune) bool {
+	switch {
+	case '0' <= c && c <= '9':
+		return true
+	case 'a' <= c && c <= 'f':
+		return true
+	case 'A' <= c && c <= 'F':
+		return true
+	}
+	return false
+}
+
+// UnicodeSanitize sanitizes string to be used in Hugo URL's
+// from https://github.com/gohugoio/hugo/blob/93aad3c543828efca2adeb7f96cf50ae29878593/helpers/path.go#L94
+func UnicodeSanitize(s string) string {
+	source := []rune(s)
+	target := make([]rune, 0, len(source))
+	var prependHyphen bool
+
+	for i, r := range source {
+		isAllowed := r == '.' || r == '/' || r == '\\' || r == '_' || r == '#' || r == '+' || r == '~'
+		isAllowed = isAllowed || unicode.IsLetter(r) || unicode.IsDigit(r) || unicode.IsMark(r)
+		isAllowed = isAllowed || (r == '%' && i+2 < len(source) && ishex(source[i+1]) && ishex(source[i+2]))
+
+		if isAllowed {
+			if prependHyphen {
+				target = append(target, '-')
+				prependHyphen = false
+			}
+			target = append(target, r)
+		} else if len(target) > 0 && (r == '-' || unicode.IsSpace(r)) {
+			prependHyphen = true
+		}
+	}
+
+	return string(target)
+}
+
 // filter out certain links (e.g. to media)
 func filter(links []Link) (res []Link) {
 	for _, l := range links {
diff --git a/walk.go b/walk.go
index 2df700c..fba7523 100644
--- a/walk.go
+++ b/walk.go
@@ -47,7 +47,7 @@ func walk(root, ext string, index bool, ignorePaths map[string]struct{}) (res []
 				// check if page is private
 				if parsedPrivate, ok := frontmatter["draft"]; !ok || !parsedPrivate.(bool) {
 					info, _ := os.Stat(s)
-					adjustedPath := strings.Replace(hugoPathTrim(trim(s, root, ".md")), " ", "-", -1)
+					adjustedPath := UnicodeSanitize(strings.Replace(hugoPathTrim(trim(s, root, ".md")), " ", "-", -1))
 					i[adjustedPath] = Content{
 						LastModified: info.ModTime(),
 						Title:   title,