handle text files correctly

hunterirving · Sep 6, 2024 · a655c92 · a655c92
1 parent e1607c3
commit a655c92
Show file tree

Hide file tree

Showing 3 changed files with 15 additions and 13 deletions.
diff --git a/extensions/waybackmachine/waybackmachine.py b/extensions/waybackmachine/waybackmachine.py
@@ -246,16 +246,6 @@ def handle_request(req):
 		# Return the processed content, status code, and appropriate Content-Type
 		return processed_content, response.status_code, {'Content-Type': mime_type}
 
-	except Exception as e:
-		print("Error occurred:", str(e))
-		error_message = f"Error fetching archived page: {str(e)}"
-		return f"<html><body><p>{error_message}</p></body></html>", 500, {'Content-Type': 'text/html'}
-
-	except Exception as e:
-		print("Error occurred:", str(e))
-		error_message = f"Error fetching archived page: {str(e)}"
-		return f"<html><body><p>{error_message}</p></body></html>", 500, {'Content-Type': 'text/html'}
-
 	except Exception as e:
 		print("Error occurred:", str(e))
 		error_message = f"Error fetching archived page: {str(e)}"

diff --git a/html_utils.py b/html_utils.py
@@ -52,6 +52,7 @@
 	"⁄": b"/",
 	"°": b"*",
 	"&deg;": b"*",
+	"&amp;": b"&",
 	"′": b"'",
 	"&prime;": b"'",
 	"″": b"''",

diff --git a/proxy.py b/proxy.py
@@ -110,12 +110,23 @@ def process_response(response):
 		status_code = 200
 		headers = {}
 
-	if isinstance(content, str):
-		content = transcode_html(content, app.config["DISABLE_CHAR_CONVERSION"])
-
+	content_type = headers.get('Content-Type', '').lower()
+
+	# Transcode content unless it's explicitly text/plain
+	if not content_type.startswith('text/plain'):
+		if isinstance(content, str):
+			content = transcode_html(content, app.config["DISABLE_CHAR_CONVERSION"])
+		elif isinstance(content, bytes):
+			content = transcode_html(content.decode('utf-8', errors='replace'), app.config["DISABLE_CHAR_CONVERSION"])
+	else:
+		# For text/plain, ensure content is in bytes
+		if isinstance(content, str):
+			content = content.encode('utf-8')
+
 	response = Response(content, status_code)
 	for key, value in headers.items():
 		response.headers[key] = value
+
 	return response
 
 def handle_default_request():