Crash report

What happened?

Most of the code in Modules/_sqlite/row.c assumes ->description is a tuple. However, it may be None. Since it is possible to craft a sqlite3.Row object "by hand", it is easy to provoke segfauls for paths that involve the PyTuple API and description == Py_None. Any real code would never directly instantiate a row object; it would be implicitly created by the cursor (via the .fetch*() APIs or by iterating on the cursor). However, I don't think we should let a possible segfault hang around.

import sqlite3
cx = sqlite3.connect(":memory:")
cu = cx.cursor()
row = sqlite3.Row(cu, (1,2))
row.keys()  # <= boom

CPython versions tested on:

3.14, 3.13, 3.12, CPython main branch

Operating systems tested on:

macOS

Output from running 'python -VV' on the command line:

No response

Linked PRs

• gh-129603: Don't segfault if sqlite3.Row description is None #129604
• [3.13] gh-129603: Don't segfault if sqlite3.Row description is None (#129604) #129923
• [3.12] gh-129603: Don't segfault if sqlite3.Row description is None (#129604) #129924