From b490dc534eedcc9878d4962e6d02baf4217a712f Mon Sep 17 00:00:00 2001
From: Tavian Barnes <tavianator@tavianator.com>
Date: Thu, 24 Feb 2022 14:06:05 -0500
Subject: regex: Use the encoding from the current locale

---
 tests.sh | 12 ++++++++++++
 1 file changed, 12 insertions(+)

(limited to 'tests.sh')

diff --git a/tests.sh b/tests.sh
index d7d9947..9003efd 100755
--- a/tests.sh
+++ b/tests.sh
@@ -625,6 +625,7 @@ gnu_tests=(
     test_regex
     test_regex_parens
     test_regex_error
+    test_regex_invalid_utf8
 
     test_regextype_posix_basic
     test_regextype_posix_extended
@@ -2146,6 +2147,17 @@ function test_regex_error() {
     fail quiet invoke_bfs basic -regex '['
 }
 
+function test_regex_invalid_utf8() {
+    rm -rf scratch/*
+
+    # Incomplete UTF-8 sequences
+    skip_if fail quiet touch scratch/$'\xC3'
+    skip_if fail quiet touch scratch/$'\xE2\x84'
+    skip_if fail quiet touch scratch/$'\xF0\x9F\x92'
+
+    bfs_diff scratch -regex 'scratch/..'
+}
+
 function test_E() {
     cd weirdnames
     bfs_diff -E . -regex '\./(\()'
-- 
cgit v1.2.3