Bug report

Bug description:

When several elements have the same ID, xml.dom.minidom.Document.getElementById() first returns the last of the siblings, and a later lookup of a different ID, which continues the incremental search, overwrites the cached entry. The same call can then return a different element on an unmodified document.

from xml.dom import minidom

doc = minidom.parseString(
    '<!DOCTYPE r [<!ATTLIST e id ID #IMPLIED>]>'
    '<r><e id="a" n="1"/><e id="a" n="2"/></r>')
print(doc.getElementById("a").getAttribute("n"))  # 2
doc.getElementById("b")                            # None
print(doc.getElementById("a").getAttribute("n"))  # 1

Both calls should return the same element, the first one in document order (as the WHATWG DOM and libxml2 do).

CPython versions tested on:

CPython main branch

Operating systems tested on:

Linux

Linked PRs

• gh-161110: Make minidom getElementById() stable for duplicate IDs #161111