xml.etree.ElementTree.iterparse() should emit ResourceWarning when not explicitly closed

osamakader

Feature or enhancement

Proposal:

Description

The iterparse() function in xml.etree.ElementTree has a TODO comment (line 1270) requesting that a ResourceWarning be emitted when the iterator is garbage collected without being explicitly closed:

def __del__(self):
    # TODO: Emit a ResourceWarning if it was not explicitly closed.
    # (When the close() method will be supported in all maintained Python versions.)
    if close_source:
        source.close()

The close() method has been available for many years now, so this TODO can be implemented.

Problem

When iterparse() opens a file by filename (rather than accepting a pre-opened file object), it manages the file internally. If the iterator is not explicitly closed, the file handle remains open until garbage collection, which can lead to:

File descriptor leaks
Files remaining locked on Windows
Resource exhaustion in long-running programs

Currently, no warning is issued to alert developers of this problem.

Proposed Solution

Implement the TODO by:

Adding a _closed flag to track whether close() has been called
Emitting a ResourceWarning in __del__() if the iterator was not closed and had opened a file internally (close_source == True)
The warning should only apply when iterparse() opened the file itself (filename passed), not when a file object was provided

Example

Current behavior (no warning):

import xml.etree.ElementTree as ET

# This leaks a file handle but gives no warning
context = ET.iterparse('data.xml')
next(context)
# Iterator goes out of scope without close()

Expected behavior (with fix):

import xml.etree.ElementTree as ET

# This should emit ResourceWarning
context = ET.iterparse('data.xml')
next(context)
# ResourceWarning: unclosed file <_io.BufferedReader name='data.xml'>

# Proper usage (no warning):
context = ET.iterparse('data.xml')
next(context)
context.close()  # Explicit close