From e11dfbc7539e31be5e4381e416874ad22e408502 Mon Sep 17 00:00:00 2001
From: Daniel Garcia Moreno <daniel.garcia@suse.com>
Date: Thu, 11 Apr 2024 13:35:03 +0200
Subject: [PATCH] python: use utf_%d_be encoding on bigendian archs

Fix https://github.com/liblouis/liblouis/issues/1551
---
 python/louis/__init__.py.in |  5 +++--
 python/tests/test_louis.py  | 16 ++++++++++++++++
 2 files changed, 19 insertions(+), 2 deletions(-)

Index: liblouis-3.25.0/python/louis/__init__.py.in
===================================================================
--- liblouis-3.25.0.orig/python/louis/__init__.py.in
+++ liblouis-3.25.0/python/louis/__init__.py.in
@@ -39,7 +39,7 @@ function for information about how liblo
 @author: Andre-Abush Clause <dev@andreabc.net>
 """
 
-from sys import getfilesystemencoding, platform, version_info
+from sys import byteorder, getfilesystemencoding, platform, version_info
 from atexit import register
 from ctypes import (
     c_ushort,
@@ -60,6 +60,7 @@ except ImportError:  # Unix/Cygwin
     _loader, _functype = cdll, CFUNCTYPE
 liblouis = _loader["###LIBLOUIS_SONAME###"]
 _is_windows = platform == "win32"
+_endianness = "be" if byteorder == "big" else "le"
 
 # { Module Configuration
 #: Specifies the charSize (in bytes) used by liblouis.
@@ -78,7 +79,7 @@ outlenMultiplier = 4 + wideCharBytes * 2
 fileSystemEncoding = "mbcs" if _is_windows else getfilesystemencoding()
 #: Specifies the encoding to use when converting from byte strings to unicode strings.
 #: @type: str
-conversionEncoding = "utf_%d_le" % (wideCharBytes * 8)
+conversionEncoding = "utf_%d_%s" % (wideCharBytes * 8, _endianness)
 # }
 
 # Some general utility functions
Index: liblouis-3.25.0/python/tests/test_louis.py
===================================================================
--- liblouis-3.25.0.orig/python/tests/test_louis.py
+++ liblouis-3.25.0/python/tests/test_louis.py
@@ -96,5 +96,21 @@ class TestUnicodeDecomposed(unittest.Tes
     def test_14(self):
         self.assertEqual(louis.translateString(["en-ueb-g1.ctb", "tests/test.cti"], "a \ud83e\udd23 b"),
                          'a "<rolling on the floor laughing"> b')
+
+class TestEndianness(unittest.TestCase):
+    def test_1(self):
+        self.assertEqual(louis.translate(["unicode.dis","en-chardefs.cti"], "abcdefghijklmnopqrstuvwxyz")[0],
+                         "⠁⠃⠉⠙⠑⠋⠛⠓⠊⠚⠅⠇⠍⠝⠕⠏⠟⠗⠎⠞⠥⠧⠺⠭⠽⠵")
+
+    def test_2(self):
+        # invert encoding
+        _encoding = louis.conversionEncoding
+        _endianness = "le" if louis._endianness == "be" else "be"
+        louis.conversionEncoding = "utf_%d_%s" % (louis.wideCharBytes * 8, _endianness)
+        with self.assertRaises(UnicodeDecodeError) as context:
+            self.assertEqual(louis.translate(["unicode.dis","en-chardefs.cti"], "abcdefghijklmnopqrstuvwxyz")[0],
+                             "⠁⠃⠉⠙⠑⠋⠛⠓⠊⠚⠅⠇⠍⠝⠕⠏⠟⠗⠎⠞⠥⠧⠺⠭⠽⠵")
+        louis.conversionEncoding = _encoding
+
 if __name__ == '__main__':
     unittest.main()