NEW @W-22462042@ fix: flow engine fails on Windows for any flow containing a non-ASCII character

packages/code-analyzer-flow-engine/FlowScanner/public/parse_utils.py

@@ -1415,19 +1415,30 @@ def quick_validate(flow_path: str) -> bool | None:
     has_start = False
     has_banned = False
     try:
-        with open(flow_path, 'r') as fp:
-            flow_data = fp.read()
-            for start_tag in START_ELEMS_TAGGED:
-                if start_tag in flow_data:
-                    has_start = True
-                    break
-
-            for banned_tag in BANNED_ELEMS_TAGGED:
-                if banned_tag in flow_data:
-                    has_banned = True
-                    break
-
-            return has_start and not has_banned
+        # Flow XML is authored/declared UTF-8, so read it as UTF-8 explicitly
+        # rather than relying on the platform locale encoding (e.g. cp1252 on
+        # Windows), which would raise UnicodeDecodeError on valid flows and
+        # cause them to be silently dropped. Fall back to cp1252 for legacy
+        # files, matching the pattern used in flow_scanner/__main__.py.
+        try:
+            with open(flow_path, 'r', encoding='utf-8') as fp:
+                flow_data = fp.read()
+        except UnicodeDecodeError:
+            # cp1252 is used on older Windows systems
+            with open(flow_path, 'r', encoding='cp1252') as fp:
+                flow_data = fp.read()
+
+        for start_tag in START_ELEMS_TAGGED:
+            if start_tag in flow_data:
+                has_start = True
+                break
+
+        for banned_tag in BANNED_ELEMS_TAGGED:
+            if banned_tag in flow_data:
+                has_banned = True
+                break
+
+        return has_start and not has_banned
 
     except FileNotFoundError:
         logger.critical(f"Could not find file {flow_path}")

packages/code-analyzer-flow-engine/FlowScanner/tests/test_parse_utils.py

@@ -0,0 +1,75 @@
+"""Tests for public.parse_utils.
+
+Run from the FlowScanner directory with::
+
+    python3 -m unittest tests.test_parse_utils
+"""
+import builtins
+import os
+import tempfile
+import unittest
+
+import public.parse_utils as pu
+
+
+class QuickValidateEncodingTest(unittest.TestCase):
+    """Regression tests for issue #2074.
+
+    Flow XML is authored/declared UTF-8, but quick_validate historically opened
+    files using the platform locale encoding. On Windows (cp1252) any flow
+    containing a non-cp1252 UTF-8 byte (e.g. 0x9D from U+201D smart quote)
+    raised UnicodeDecodeError, the flow was silently dropped, and the scanner
+    ultimately failed to write its results file.
+    """
+
+    #: A valid flow whose description contains a right double quotation mark
+    #: (U+201D). This encodes to byte 0x9D in UTF-8's continuation, which maps
+    #: to <undefined> in cp1252 and would raise UnicodeDecodeError.
+    UTF8_FLOW = (
+        '<?xml version="1.0" encoding="UTF-8"?>\n'
+        '<Flow xmlns="http://soap.sforce.com/2006/04/metadata">'
+        '<description>smart ”quote</description>'
+        '<start></start>'
+        '</Flow>'
+    )
+
+    def setUp(self):
+        fd, self.flow_path = tempfile.mkstemp(suffix='.flow')
+        os.close(fd)
+        with open(self.flow_path, 'w', encoding='utf-8') as fp:
+            fp.write(self.UTF8_FLOW)
+
+    def tearDown(self):
+        if os.path.exists(self.flow_path):
+            os.unlink(self.flow_path)
+
+    def test_utf8_flow_validates_under_cp1252_locale(self):
+        """quick_validate must read UTF-8 flows even when the locale is cp1252."""
+        orig_open = builtins.open
+
+        def cp1252_default_open(file, mode='r', *args, **kwargs):
+            # Emulate a Windows cp1252 locale: text reads with no explicit
+            # encoding fall back to cp1252.
+            if 'b' not in mode and 'encoding' not in kwargs:
+                kwargs['encoding'] = 'cp1252'
+            return orig_open(file, mode, *args, **kwargs)
+
+        builtins.open = cp1252_default_open
+        try:
+            result = pu.quick_validate(self.flow_path)
+        finally:
+            builtins.open = orig_open
+
+        self.assertTrue(result)
+
+    def test_utf8_flow_validates_under_utf8_locale(self):
+        """quick_validate still works under a normal UTF-8 locale."""
+        self.assertTrue(pu.quick_validate(self.flow_path))
+
+    def test_missing_file_returns_none(self):
+        """A missing file still returns None (unchanged behavior)."""
+        self.assertIsNone(pu.quick_validate(self.flow_path + '.does-not-exist'))
+
+
+if __name__ == '__main__':
+    unittest.main()

packages/code-analyzer-flow-engine/package.json

@@ -1,7 +1,7 @@
 {
   "name": "@salesforce/code-analyzer-flow-engine",
   "description": "Plugin package that adds 'Flow Scanner' as an engine into Salesforce Code Analyzer",
-  "version": "0.40.0",
+  "version": "0.40.1-SNAPSHOT",
   "author": "The Salesforce Code Analyzer Team",
   "license": "BSD-3-Clause",
   "homepage": "https://developer.salesforce.com/docs/platform/salesforce-code-analyzer/overview",