5.0.2.1 — CVE-2026-41066 security backport#1
Closed
martinPavesio wants to merge 11 commits into
Closed
Conversation
Backport of CVE-2026-41066 (GHSA-vfmq-68hx-4jfw) from lxml 6.1.0: - iterparse.__init__: resolve_entities default True -> 'internal' - ETCompatXMLParser.__init__: resolve_entities default True -> 'internal' Upstream commit: ab431ea (LP#2146291) Also includes Python 2.7 / ActiveState platform build fixes: - setupinfo.py: use AS-built libxml2/libxslt; str.format for Py2 compat - versioninfo.py: use __file__ instead of sys.argv[0] for base dir - src/lxml/etree.c: pre-generated C included to avoid Cython 3 incompatibility with Python 2.7; includes const xmlChar** cast fix for libxml2 2.15.x (fixes -Werror=incompatible-pointer-types on Rocky 9 builder) Version: 5.0.2.1
The CVE-2026-41066 fix lives entirely in the pre-generated etree.c (values[9]/values[10] default changed from Py_True to __pyx_n_s_internal). Having the .pxi files also modified caused Cython 3 to detect source changes and regenerate etree.c from scratch, discarding the fix and producing Python 3-only C code that fails to compile against Python 2.7. Rule: when a fix is applied at the C level in a pre-generated file, the Cython source (.pxi/.pyx) must NOT be modified.
…sax) setupinfo.py raises RuntimeError if any C file is absent when Cython is not available. Only etree.c was committed initially; all four remaining files are needed for a Cython-free build. Copied unmodified from lxml-5.0.2 PyPI tarball.
lxml's setupinfo.py also requires src/lxml/html/clean.c and src/lxml/html/diff.c to be present for a Cython-free build. Copied unmodified from lxml-5.0.2 PyPI tarball.
objectify.c includes etree_api.h, etree.h, lxml.etree.h, lxml.etree_api.h which are generated files gitignored by the upstream repo. Copied unmodified from lxml-5.0.2 PyPI tarball.
Change 'zlib' to 'z' in libs.extend() on Windows to use the correct library name when linking against the static zlib library.
Let Cython 3 compile from .pyx/.pxi sources at build time. The pre-generated file approach was a wrong fix for the const xmlChar** compiler error; the correct fix is in the Cython source (proxy.pxi).
…ll site xmlAttribute.defaultValue is xmlChar* (non-const) in libxml2 2.15.x while _fixThreadDictPtr expects const_xmlChar**. Add explicit Cython cast at the one non-const call site. Fixes -Werror=incompatible-pointer-types on Rocky 9 / GCC 14 when Cython regenerates etree.c from this source. Backport of CVE fix for C-level const mismatch discovered during CS-2178.
iterparse.__init__ defaulted to resolve_entities=True, allowing XXE injection via untrusted XML. Upstream fix: python-pillow/lxml@ab431ea0 (lxml 6.1.0, LP#2146291). Cython will regenerate etree.c from this source at build time.
…LParser ETCompatXMLParser.__init__ defaulted to resolve_entities=True, allowing XXE injection. Upstream fix: lxml 6.1.0 (LP#2146291).
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Backport of CVE-2026-41066 (XXE via resolve_entities default) for Python 2.7. Tag: 5.0.2.1. Ticket: CS-2178.