From 19675337cc6b5209f5fc74e8aea5363dd4a0fa9b Mon Sep 17 00:00:00 2001
From: Saeed Rasooli <saeed.gnu@gmail.com>
Date: Sun, 14 Apr 2024 04:58:41 +0330
Subject: [PATCH] DSL: detect utf-16-le and utf-16-be encodings, #555

---
 pyglossary/plugins/dsl/__init__.py | 9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)

diff --git a/pyglossary/plugins/dsl/__init__.py b/pyglossary/plugins/dsl/__init__.py
index 77fc91244..6e7e4c241 100644
--- a/pyglossary/plugins/dsl/__init__.py
+++ b/pyglossary/plugins/dsl/__init__.py
@@ -230,7 +230,12 @@ def open(
 			self.processHeaderLine(line)
 
 	def detectEncoding(self) -> str:
-		for testEncoding in ("utf-8", "utf-16"):
+		for testEncoding in (
+			"utf-8",
+			"utf-16",
+			"utf-16-le",
+			"utf-16-be",
+		):
 			with compressionOpen(
 				self._filename,
 				dz=True,
@@ -240,7 +245,7 @@ def detectEncoding(self) -> str:
 				try:
 					for _ in range(10):
 						fileObj.readline()
-				except UnicodeDecodeError:
+				except (UnicodeDecodeError, UnicodeError):
 					log.info(f"Encoding of DSL file is not {testEncoding}")
 					continue
 				else: