[fix] Permit to display the output even in case of unicode errors

doc/whatsnew/fragments/8736.bugfix

@@ -0,0 +1,5 @@
+When displaying unicode with surrogates (or other potential ``UnicodeEncodeError``),
+pylint will now display a '?' character (using ``encode(encoding="utf-8", errors="replace")``)
+instead of crashing. The functional tests classes are also updated to handle this case.
+
+Closes #8736.

pylint/reporters/base_reporter.py

@@ -42,7 +42,14 @@ def handle_message(self, msg: Message) -> None:
 
     def writeln(self, string: str = "") -> None:
         """Write a line in the output buffer."""
-        print(string, file=self.out)
+        try:
+            print(string, file=self.out)
+        except UnicodeEncodeError:
+            print(self.reencode_output_after_unicode_error(string), file=self.out)
+
+    @staticmethod
+    def reencode_output_after_unicode_error(string: str) -> str:
+        return string.encode(encoding="utf-8", errors="replace").decode("utf8")
 
     def display_reports(self, layout: Section) -> None:
         """Display results encapsulated in the layout tree."""

pylint/testutils/functional/lint_module_output_update.py

@@ -40,4 +40,4 @@ def _check_output_text(
         with open(self._test_file.expected_output, "w", encoding="utf-8") as f:
             writer = csv.writer(f, dialect="test")
             for line in actual_output:
-                writer.writerow(line.to_csv())
+                self.safe_write_output_line(writer, line)

pylint/testutils/lint_module_test.py

@@ -11,7 +11,7 @@
 from collections import Counter
 from io import StringIO
 from pathlib import Path
-from typing import TextIO
+from typing import TYPE_CHECKING, TextIO
 
 import pytest
 from _pytest.config import Config
@@ -20,6 +20,7 @@
 from pylint.config.config_initialization import _config_initialization
 from pylint.lint import PyLinter
 from pylint.message.message import Message
+from pylint.reporters import BaseReporter
 from pylint.testutils.constants import _EXPECTED_RE, _OPERATORS, UPDATE_OPTION
 
 # need to import from functional.test_file to avoid cyclic import
@@ -31,6 +32,8 @@
 from pylint.testutils.output_line import OutputLine
 from pylint.testutils.reporter_for_tests import FunctionalTestReporter
 
+if TYPE_CHECKING:
+    import _csv
 MessageCounter = Counter[tuple[int, str]]
 
 PYLINTRC = Path(__file__).parent / "testing_pylintrc"
@@ -303,10 +306,22 @@ def error_msg_for_unequal_output(
             expected_csv = StringIO()
             writer = csv.writer(expected_csv, dialect="test")
             for line in sorted(received_lines, key=sort_by_line_number):
-                writer.writerow(line.to_csv())
+                self.safe_write_output_line(writer, line)
             error_msg += expected_csv.getvalue()
         return error_msg
 
+    def safe_write_output_line(self, writer: _csv._writer, line: OutputLine) -> None:
+        """Write an OutputLine to the CSV writer, handling UnicodeEncodeError."""
+        try:
+            writer.writerow(line.to_csv())
+        except UnicodeEncodeError:
+            writer.writerow(
+                [
+                    BaseReporter.reencode_output_after_unicode_error(s)
+                    for s in line.to_csv()
+                ]
+            )
+
     def _check_output_text(
         self,
         _: MessageCounter,

tests/functional/r/regression_02/regression_8736.py

@@ -0,0 +1,3 @@
+"""This does not crash in the functional tests, but it did when called directly."""
+
+assert "\U00010000" == "\ud800\udc00"  # [comparison-of-constants]

tests/functional/r/regression_02/regression_8736.txt

@@ -0,0 +1 @@
+comparison-of-constants:3:7:3:37::"Comparison between constants: '𐀀 == ??' has a constant value":HIGH