Remove remaining references to lxml python dependency, which is no longer required (#20356)

changelog.d/20356.misc

@@ -0,0 +1 @@
+Remove unused `lxml` packaging and type-stub dependencies, and update URL-preview tests to check for BeautifulSoup instead.

debian/build_virtualenv

@@ -53,7 +53,6 @@ dh_virtualenv \
     --builtin-venv \
     --python "$SNAKE" \
     --upgrade-pip \
-    --preinstall="lxml" \
     --preinstall="mock" \
     --preinstall="wheel" \
     --extra-pip-arg="--no-deps" \

debian/changelog

@@ -1,3 +1,9 @@
+matrix-synapse-py3 (1.163.0+unreleased1) UNRELEASED; urgency=medium
+
+  * Stop preinstalling the unused Python lxml package.
+
+ -- Synapse Packaging team <packages@matrix.org>  Wed, 14 Oct 2026 11:30:04 +0100
+
 matrix-synapse-py3 (1.163.0) stable; urgency=medium
 
   * New synapse release 1.163.0.

poetry.lock

@@ -1247,22 +1247,6 @@ files = [
     {file = "librt-0.8.1.tar.gz", hash = "sha256:be46a14693955b3bd96014ccbdb8339ee8c9346fbe11c1b78901b55125f14c73"},
 ]
 
-[[package]]
-name = "lxml-stubs"
-version = "0.5.1"
-description = "Type annotations for the lxml package"
-optional = true
-python-versions = "*"
-groups = ["main"]
-markers = "extra == \"saml2\" or extra == \"all\""
-files = [
-    {file = "lxml-stubs-0.5.1.tar.gz", hash = "sha256:e0ec2aa1ce92d91278b719091ce4515c12adc1d564359dfaf81efa7d4feab79d"},
-    {file = "lxml_stubs-0.5.1-py3-none-any.whl", hash = "sha256:1f689e5dbc4b9247cb09ae820c7d34daeb1fdbd1db06123814b856dae7787272"},
-]
-
-[package.extras]
-test = ["coverage[toml] (>=7.2.5)", "mypy (>=1.2.0)", "pytest (>=7.3.0)", "pytest-mypy-plugins (>=1.10.1)"]
-
 [[package]]
 name = "markdown-it-py"
 version = "4.0.0"
@@ -1972,7 +1956,7 @@ files = [
     {file = "pycparser-3.0-py3-none-any.whl", hash = "sha256:b727414169a36b7d524c1c3e31839a521725078d7b2ff038656844266160a992"},
     {file = "pycparser-3.0.tar.gz", hash = "sha256:600f49d217304a5902ac3c37e1281c9fe94e4d0489de643a9504c5cdfdfc6b29"},
 ]
-markers = {main = "implementation_name != \"PyPy\"", dev = "implementation_name != \"PyPy\" and platform_python_implementation != \"PyPy\""}
+markers = {main = "implementation_name != \"PyPy\"", dev = "platform_python_implementation != \"PyPy\" and implementation_name != \"PyPy\""}
 
 [[package]]
 name = "pydantic"
@@ -3645,7 +3629,7 @@ docs = ["Sphinx", "repoze.sphinx.autointerface"]
 test = ["zope.i18nmessageid", "zope.testing", "zope.testrunner (>=6.4)"]
 
 [extras]
-all = ["authlib", "beautifulsoup4", "defusedxml", "hiredis", "jaeger-client", "lxml-stubs", "matrix-synapse-ldap3", "opentracing", "psycopg2", "psycopg2cffi", "psycopg2cffi-compat", "pympler", "pysaml2", "pytz", "sentry-sdk", "thrift", "tornado", "txredisapi"]
+all = ["authlib", "beautifulsoup4", "defusedxml", "hiredis", "jaeger-client", "matrix-synapse-ldap3", "opentracing", "psycopg2", "psycopg2cffi", "psycopg2cffi-compat", "pympler", "pysaml2", "pytz", "sentry-sdk", "thrift", "tornado", "txredisapi"]
 cache-memory = ["pympler"]
 jwt = ["authlib"]
 matrix-synapse-ldap3 = ["matrix-synapse-ldap3"]
@@ -3653,7 +3637,7 @@ oidc = ["authlib"]
 opentracing = ["jaeger-client", "opentracing", "thrift", "tornado"]
 postgres = ["psycopg2", "psycopg2cffi", "psycopg2cffi-compat"]
 redis = ["hiredis", "txredisapi"]
-saml2 = ["defusedxml", "lxml-stubs", "pysaml2", "pytz"]
+saml2 = ["defusedxml", "pysaml2", "pytz"]
 sentry = ["sentry-sdk"]
 test = ["idna", "parameterized"]
 url-preview = ["beautifulsoup4"]
@@ -3661,4 +3645,4 @@ url-preview = ["beautifulsoup4"]
 [metadata]
 lock-version = "2.1"
 python-versions = ">=3.10.0,<4.0.0"
-content-hash = "6e8f0d223b61ab7c100236797f0ec1eed7f23b819c01c117fe445c6b1e91bbfc"
+content-hash = "ae436463df1256bff83762dfd3871e58682c3d6e67f622b3c2b5f5cfa25ce475"

pyproject.toml

@@ -128,7 +128,6 @@ postgres = [
 ]
 saml2 = [
     "pysaml2>=4.5.0",
-    "lxml-stubs>=0.4.0",
 
     # Transitive dependencies from pysaml2
     # These dependencies aren't directly required by Synapse.
@@ -187,7 +186,6 @@ all = [
     "psycopg2cffi-compat==1.1;platform_python_implementation == 'PyPy'",
     # saml2
     "pysaml2>=4.5.0",
-    "lxml-stubs>=0.4.0",
     # oidc and jwt
     "authlib>=0.15.1",
     # url-preview

tests/rest/client/test_capabilities.py

@@ -32,9 +32,9 @@ from tests import unittest
 from tests.unittest import override_config, skip_unless
 
 try:
-    import lxml
+    import bs4
 except ImportError:
-    lxml = None  # type: ignore[assignment]
+    bs4 = None  # type: ignore[assignment]
 
 
 class CapabilitiesTestCase(unittest.HomeserverTestCase):
@@ -336,7 +336,7 @@ class CapabilitiesTestCase(unittest.HomeserverTestCase):
         self.assertEqual(channel.code, HTTPStatus.OK)
         self.assertFalse(capabilities["m.preview_url"]["enabled"])
 
-    @skip_unless(lxml is not None, "Requires lxml")
+    @skip_unless(bs4 is not None, "Requires beautifulsoup4")
     @override_config(
         {
             "url_preview_enabled": True,
@@ -372,7 +372,7 @@ class CapabilitiesTestCase(unittest.HomeserverTestCase):
         self.assertFalse(capabilities["io.element.msc4452.preview_url"]["enabled"])
         self.assertFalse(capabilities["m.preview_url"]["enabled"])
 
-    @skip_unless(lxml is not None, "Requires lxml")
+    @skip_unless(bs4 is not None, "Requires beautifulsoup4")
     @override_config(
         {
             "url_preview_enabled": True,

tests/rest/client/test_media.py

@@ -1605,7 +1605,7 @@ class URLPreviewTests(unittest.HomeserverTestCase):
 
 
 # We test this here because this endpoint must still work
-# even if lxml is not installed.
+# even if beautifulsoup4 is not installed.
 class URLPreviewDisabledTests(unittest.HomeserverTestCase):
     servlets = [
         admin.register_servlets,