From c0f2a5ef9125f2b44b32cab92e1b0b1bfe0e3fdc Mon Sep 17 00:00:00 2001 From: Batuhan Taskaya Date: Sat, 29 Oct 2022 15:28:20 +0300 Subject: [PATCH] gh-98744: Prevent column-level decoding crashes on traceback module (#98824) --- Lib/test/test_traceback.py | 50 +++++++++++++++++++ Lib/traceback.py | 25 ++++------ ...2-10-28-23-44-17.gh-issue-98744.sGHDWm.rst | 2 + 3 files changed, 63 insertions(+), 14 deletions(-) create mode 100644 Misc/NEWS.d/next/Library/2022-10-28-23-44-17.gh-issue-98744.sGHDWm.rst diff --git a/Lib/test/test_traceback.py b/Lib/test/test_traceback.py index cf52d17ff8f..56b168735d1 100644 --- a/Lib/test/test_traceback.py +++ b/Lib/test/test_traceback.py @@ -804,6 +804,56 @@ class TracebackErrorLocationCaretTestBase: ] self.assertEqual(actual, expected) + def test_wide_characters_unicode_with_problematic_byte_offset(self): + def f(): + width + + actual = self.get_exception(f) + expected = [ + f"Traceback (most recent call last):", + f" File \"{__file__}\", line {self.callable_line}, in get_exception", + f" callable()", + f" File \"{__file__}\", line {f.__code__.co_firstlineno + 1}, in f", + f" width", + ] + self.assertEqual(actual, expected) + + + def test_byte_offset_with_wide_characters_middle(self): + def f(): + width = 1 + raise ValueError(width) + + actual = self.get_exception(f) + expected = [ + f"Traceback (most recent call last):", + f" File \"{__file__}\", line {self.callable_line}, in get_exception", + f" callable()", + f" File \"{__file__}\", line {f.__code__.co_firstlineno + 2}, in f", + f" raise ValueError(width)", + ] + self.assertEqual(actual, expected) + + def test_byte_offset_multiline(self): + def f(): + www = 1 + th = 0 + + print(1, www( + th)) + + actual = self.get_exception(f) + expected = [ + f"Traceback (most recent call last):", + f" File \"{__file__}\", line {self.callable_line}, in get_exception", + f" callable()", + f" File \"{__file__}\", line {f.__code__.co_firstlineno + 4}, in f", + f" print(1, www(", + f" ^^^^", + ] + self.assertEqual(actual, expected) + + @requires_debug_ranges() class PurePythonTracebackErrorCaretTests( diff --git a/Lib/traceback.py b/Lib/traceback.py index 6270100348a..0f0f2b317de 100644 --- a/Lib/traceback.py +++ b/Lib/traceback.py @@ -476,32 +476,32 @@ class StackSummary(list): frame_summary.colno is not None and frame_summary.end_colno is not None ): - colno = _byte_offset_to_character_offset( - frame_summary._original_line, frame_summary.colno) - end_colno = _byte_offset_to_character_offset( - frame_summary._original_line, frame_summary.end_colno) + start_offset = _byte_offset_to_character_offset( + frame_summary._original_line, frame_summary.colno) + 1 + end_offset = _byte_offset_to_character_offset( + frame_summary._original_line, frame_summary.end_colno) + 1 anchors = None if frame_summary.lineno == frame_summary.end_lineno: with suppress(Exception): anchors = _extract_caret_anchors_from_line_segment( - frame_summary._original_line[colno - 1:end_colno - 1] + frame_summary._original_line[start_offset - 1:end_offset - 1] ) else: - end_colno = stripped_characters + len(stripped_line) + end_offset = stripped_characters + len(stripped_line) # show indicators if primary char doesn't span the frame line - if end_colno - colno < len(stripped_line) or ( + if end_offset - start_offset < len(stripped_line) or ( anchors and anchors.right_start_offset - anchors.left_end_offset > 0): row.append(' ') - row.append(' ' * (colno - stripped_characters)) + row.append(' ' * (start_offset - stripped_characters)) if anchors: row.append(anchors.primary_char * (anchors.left_end_offset)) row.append(anchors.secondary_char * (anchors.right_start_offset - anchors.left_end_offset)) - row.append(anchors.primary_char * (end_colno - colno - anchors.right_start_offset)) + row.append(anchors.primary_char * (end_offset - start_offset - anchors.right_start_offset)) else: - row.append('^' * (end_colno - colno)) + row.append('^' * (end_offset - start_offset)) row.append('\n') @@ -561,10 +561,7 @@ class StackSummary(list): def _byte_offset_to_character_offset(str, offset): as_utf8 = str.encode('utf-8') - if offset > len(as_utf8): - offset = len(as_utf8) - - return len(as_utf8[:offset + 1].decode("utf-8")) + return len(as_utf8[:offset].decode("utf-8", errors="replace")) _Anchors = collections.namedtuple( diff --git a/Misc/NEWS.d/next/Library/2022-10-28-23-44-17.gh-issue-98744.sGHDWm.rst b/Misc/NEWS.d/next/Library/2022-10-28-23-44-17.gh-issue-98744.sGHDWm.rst new file mode 100644 index 00000000000..cf99ea51bf7 --- /dev/null +++ b/Misc/NEWS.d/next/Library/2022-10-28-23-44-17.gh-issue-98744.sGHDWm.rst @@ -0,0 +1,2 @@ +Prevent crashing in :mod:`traceback` when retrieving the byte-offset for +some source files that contain certain unicode characters.