From f858ec383b3d17f146fb765e5b7385895da5bdd1 Mon Sep 17 00:00:00 2001
From: mntn <85877297+mntn-xyz@users.noreply.github.com>
Date: Sat, 25 Dec 2021 11:51:26 -0500
Subject: [PATCH] Allow non-latin characters in paths

This makes gmnhg use [\pL\d\-_] to match paths instead of \w,
allowing letter characters from any Unicode language in file
paths.

Fixes #48.
---
 cmd/gmnhg/main.go      | 6 +++---
 internal/gmnhg/post.go | 2 +-
 2 files changed, 4 insertions(+), 4 deletions(-)

diff --git a/cmd/gmnhg/main.go b/cmd/gmnhg/main.go
index e274c1c..31c0e47 100644
--- a/cmd/gmnhg/main.go
+++ b/cmd/gmnhg/main.go
@@ -149,9 +149,9 @@ const (
 )
 
 var (
-	tmplNameRegex  = regexp.MustCompile("^" + templateBase + `([\w-_ /]+)\.gotmpl$`)
-	leafIndexRegex = regexp.MustCompile("^" + contentBase + `([\w-_ /]+)/index\.[\w]+$`)
-	pagePathRegex  = regexp.MustCompile("^" + contentBase + `([\w-_ /]+)/([\w-_ ]+)\.md$`)
+	tmplNameRegex  = regexp.MustCompile("^" + templateBase + `([\pL\d\-_ /]+)\.gotmpl$`)
+	leafIndexRegex = regexp.MustCompile("^" + contentBase + `([\pL\d\-_ /]+)/index\.[\pL\d_]+$`)
+	pagePathRegex  = regexp.MustCompile("^" + contentBase + `([\pL\d\-_ /]+)/([\pL\d\-_ ]+)\.md$`)
 )
 
 var hugoConfigFiles = []string{"config.toml", "config.yaml", "config.json"}
diff --git a/internal/gmnhg/post.go b/internal/gmnhg/post.go
index 4747574..4408673 100644
--- a/internal/gmnhg/post.go
+++ b/internal/gmnhg/post.go
@@ -62,7 +62,7 @@ var (
 	yamlDelimiter   = []byte("---\n")
 	tomlDelimiter   = []byte("+++\n")
 	jsonObjectRegex = regexp.MustCompile(`\A(\{[\s\S]*\})\n\n`)
-	orgModeRegex    = regexp.MustCompile(`\A((?:#\+\w+\[?\]?: ?[^\n\r]*\n)+)`)
+	orgModeRegex    = regexp.MustCompile(`\A((?:#\+[\pL\d_]+\[?\]?: ?[^\n\r]*\n)+)`)
 )
 
 // ParseMetadata extracts TOML/JSON/YAML/org-mode format front matter