From cd42ba388bbdd66b9bcc587defd55e2c449450a5 Mon Sep 17 00:00:00 2001
From: Adam Saponara <as@php.net>
Date: Wed, 24 Jul 2024 22:12:24 -0400
Subject: [PATCH] send non-printable codepoints as U+FFFD

---
 termbox2.h                             | 13 +++++++------
 tests/test_invalid_utf8/expected.ansi  |  2 +-
 tests/test_invalid_utf8/test.php       |  6 +++++-
 tests/test_non_printable/expected.ansi | 24 ++++++++++++++++++++++++
 tests/test_non_printable/test.php      | 25 +++++++++++++++++++++++++
 5 files changed, 62 insertions(+), 8 deletions(-)
 create mode 100644 tests/test_non_printable/expected.ansi
 create mode 100755 tests/test_non_printable/test.php

diff --git a/termbox2.h b/termbox2.h
index 5fc791f..d8ca3ad 100644
--- a/termbox2.h
+++ b/termbox2.h
@@ -51,6 +51,7 @@ SOFTWARE.
 #include <termios.h>
 #include <unistd.h>
 #include <wchar.h>
+#include <wctype.h>
 
 #ifdef PATH_MAX
 #define TB_PATH_MAX PATH_MAX
@@ -493,6 +494,9 @@ int tb_hide_cursor(void);
  *
  * Function tb_extend_cell() is a shortcut for appending 1 codepoint to
  * cell->ech.
+ *
+ * Non-printable (`iswprint(3)`) codepoints are replaced with U+FFFD at render
+ * time.
  */
 int tb_set_cell(int x, int y, uint32_t ch, uintattr_t fg, uintattr_t bg);
 int tb_set_cell_ex(int x, int y, uint32_t *ch, size_t nch, uintattr_t fg,
@@ -3224,13 +3228,10 @@ static int send_cluster(int x, int y, uint32_t *ch, size_t nch) {
     int i;
     for (i = 0; i < (int)nch; i++) {
         uint32_t ch32 = *(ch + i);
-        int chu8_len;
-        if (ch32 == 0) { // replace null with space (from termbox 19dbee5)
-            chu8_len = 1;
-            chu8[0] = ' ';
-        } else {
-            chu8_len = tb_utf8_unicode_to_char(chu8, ch32);
+        if (!iswprint((wint_t)ch32)) {
+            ch32 = 0xfffd; // replace non-printable codepoints with U+FFFD
         }
+        int chu8_len = tb_utf8_unicode_to_char(chu8, ch32);
         if_err_return(rv, bytebuf_nputs(&global.out, chu8, (size_t)chu8_len));
     }
 
diff --git a/tests/test_invalid_utf8/expected.ansi b/tests/test_invalid_utf8/expected.ansi
index b440b63..ec78ff0 100644
--- a/tests/test_invalid_utf8/expected.ansi
+++ b/tests/test_invalid_utf8/expected.ansi
@@ -1,5 +1,5 @@
 #5foo�
-
+#5�
 
 
 
diff --git a/tests/test_invalid_utf8/test.php b/tests/test_invalid_utf8/test.php
index 76480dd..fc956cc 100755
--- a/tests/test_invalid_utf8/test.php
+++ b/tests/test_invalid_utf8/test.php
@@ -2,7 +2,11 @@
 declare(strict_types=1);
 
 $test->ffi->tb_init();
-$test->ffi->tb_print_ex(0, 0, 0, 0, NULL, "foo\xc2\x00password");
+
+$y = 0;
+$test->ffi->tb_print_ex(0, $y++, 0, 0, NULL, "foo\xc2\x00password"); // stop at NULL
+$test->ffi->tb_set_cell(0, $y++, 0xffff, 0, 0); // invalid codepoint
+
 $test->ffi->tb_present();
 
 $test->screencap();
diff --git a/tests/test_non_printable/expected.ansi b/tests/test_non_printable/expected.ansi
new file mode 100644
index 0000000..a243cb7
--- /dev/null
+++ b/tests/test_non_printable/expected.ansi
@@ -0,0 +1,24 @@
+#50x00 �
+#50x01 �
+#50x08 �
+#50x09 �
+#50x0a �
+#50x1f �
+#50x7f �
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
diff --git a/tests/test_non_printable/test.php b/tests/test_non_printable/test.php
new file mode 100755
index 0000000..e2c72da
--- /dev/null
+++ b/tests/test_non_printable/test.php
@@ -0,0 +1,25 @@
+<?php
+declare(strict_types=1);
+
+$test->ffi->tb_init();
+
+$codepoints = [
+    0x00, // NULL
+    0x01, // control code
+    0x08, // backspace
+    0x09, // tab
+    0x0a, // newline
+    0x1f, // control code
+    0x7f, // delete
+];
+
+$y = 0;
+foreach ($codepoints as $ch) {
+    $test->ffi->tb_printf(0, $y, 0, 0, "0x%02x ", $ch);
+    $test->ffi->tb_set_cell(5, $y, $ch, 0, 0);
+    $y += 1;
+}
+
+$test->ffi->tb_present();
+
+$test->screencap();
-- 
2.39.5