linter: Expect rules to be in NFKC (publicsuffix#725)

For more info read
https://docs.python.org/3/library/unicodedata.html
under 'unicodedata.normalize'.

See publicsuffix#715 (comment)

linter/pslint.py

@@ -25,6 +25,7 @@
 
 import sys
 import codecs
+import unicodedata
 
 nline = 0
 line = ""
@@ -104,7 +105,7 @@ def lint_psl(infile):
 	for line in lines:
 		nline += 1
 
-		# check for leadind/trailing whitespace
+		# check for leading/trailing whitespace
 		stripped = line.strip()
 		if stripped != line:
 			line = line.replace('\t','\\t')
@@ -168,6 +169,10 @@ def lint_psl(infile):
 			error('Invalid UTF-8 character')
 			continue
 
+		# rules must be NFC coded (Unicode's Normal Form Kanonical Composition)
+		if unicodedata.normalize("NFKC", line) != line:
+			error('Rule must be NFKC')
+
 		# each rule must be lowercase (or more exactly: not uppercase and not titlecase)
 		if line != line.lower():
 			error('Rule must be lowercase')

linter/test_NFKC.expected

@@ -0,0 +1,2 @@
+9: error: Rule must be NFKC: 'südtirol.it'
+11: warning: No PRIVATE section found

linter/test_NFKC.input

@@ -0,0 +1,11 @@
+// test:
+// - label contains non-NFKC character(s)
+//
+// best viewed with 'LC_ALL=C.UTF-8 vi <filename>' (or any other UTF-8 locale)
+
+// ===BEGIN ICANN DOMAINS===
+
+südtirol.it
+südtirol.it
+
+// ===END ICANN DOMAINS===