Only drop the license-problematic file from ISO-Schematron, not the entire module

Resolves: RHEL-5570

.gitignore

@@ -1,3 +1,4 @@
 /lxml-4.9.1.tar.gz
 /lxml-4.9.2.tar.gz
 /lxml-4.9.2-no-isoschematron.tar.gz
+/lxml-4.9.2-no-isoschematron-rng.tar.gz

380.patch

@@ -0,0 +1,24 @@
+From d18f2f22218ea0e0b5327b5a2bda789afdf16e41 Mon Sep 17 00:00:00 2001
+From: =?UTF-8?q?Miro=20Hron=C4=8Dok?= <miro@hroncok.cz>
+Date: Fri, 14 Jul 2023 12:18:25 +0200
+Subject: [PATCH] Skip test_isoschematron.test_schematron_invalid_schema_empty
+ without the RNG file
+
+The expected SchematronParseError only happens when validate_schema is true.
+---
+ src/lxml/tests/test_isoschematron.py | 2 ++
+ 1 file changed, 2 insertions(+)
+
+diff --git a/src/lxml/tests/test_isoschematron.py b/src/lxml/tests/test_isoschematron.py
+index 6d2aa3fb6..900f257c3 100644
+--- a/src/lxml/tests/test_isoschematron.py
++++ b/src/lxml/tests/test_isoschematron.py
+@@ -55,6 +55,8 @@ def test_schematron_empty_pattern(self):
+         schema = isoschematron.Schematron(schema)
+         self.assertTrue(schema)
+ 
++    @unittest.skipIf(not isoschematron.schematron_schema_valid_supported,
++                     'SchematronParseError is risen only when validate_schema is true')
+     def test_schematron_invalid_schema_empty(self):
+         schema = self.parse('''\
+ <schema xmlns="http://purl.oclc.org/dsdl/schematron" />

Make-the-validation-of-ISO-Schematron-files-optional.patch

@@ -0,0 +1,116 @@
+From 4ac96ce046e9f58141bd66639ba8cb1fad9deefb Mon Sep 17 00:00:00 2001
+From: Stefan Behnel <stefan_ml@behnel.de>
+Date: Wed, 12 Jul 2023 16:59:07 +0200
+Subject: [PATCH] Make the validation of ISO-Schematron files optional in lxml,
+ depending on the availability of the RNG validation file. Some lxml
+ distributions discard the validation schema file due to licensing issues.
+
+See https://bugs.launchpad.net/lxml/+bug/2024343
+---
+ CHANGES.txt                        | 11 +++++++++++
+ doc/validation.txt                 |  9 +++++++++
+ src/lxml/isoschematron/__init__.py | 24 +++++++++++++++++++-----
+ 3 files changed, 39 insertions(+), 5 deletions(-)
+
+diff --git a/CHANGES.txt b/CHANGES.txt
+index c684ad5..40e32cd 100644
+--- a/CHANGES.txt
++++ b/CHANGES.txt
+@@ -2,6 +2,17 @@
+ lxml changelog
+ ==============
+ 
++4.9.2+
++======
++
++* LP#2024343: The validation of the schema file itself is now optional in the
++  ISO-Schematron implementation.  This was done because some lxml distributions
++  discard the RNG validation schema file due to licensing issues.  The validation
++  can now always be disabled with ``Schematron(..., validate_schema=False)``.
++  It is enabled by default if available and disabled otherwise.  The module
++  constant ``lxml.isoschematron.schematron_schema_valid_supported`` can be used
++  to detect whether schema file validation is available.
++
+ 4.9.2 (2022-12-13)
+ ==================
+ 
+diff --git a/doc/validation.txt b/doc/validation.txt
+index af9d007..27c0ccd 100644
+--- a/doc/validation.txt
++++ b/doc/validation.txt
+@@ -615,6 +615,15 @@ The usage of validation phases is a unique feature of ISO-Schematron and can be
+ a very powerful tool e.g. for establishing validation stages or to provide 
+ different validators for different "validation audiences".
+ 
++Note: Some lxml distributions exclude the validation schema file due to licensing issues.
++Since lxml 4.9.2-8, the validation of the user provided schema can be disabled with
++``Schematron(..., validate_schema=False)``.
++It is enabled by default if available and disabled otherwise.  Previous versions of
++lxml always had it enabled and failed at import time if the file was not available.
++Thus, some distributions chose to remove the entire ISO-Schematron support.
++The module constant ``lxml.isoschematron.schematron_schema_valid_supported`` can be used
++since lxml 4.9.2-8 to detect whether schema file validation is available.
++
+ (Pre-ISO-Schematron)
+ --------------------
+ 
+diff --git a/src/lxml/isoschematron/__init__.py b/src/lxml/isoschematron/__init__.py
+index 5967b10..2846a66 100644
+--- a/src/lxml/isoschematron/__init__.py
++++ b/src/lxml/isoschematron/__init__.py
+@@ -61,10 +61,16 @@ iso_svrl_for_xslt1 = _etree.XSLT(_etree.parse(
+ svrl_validation_errors = _etree.XPath(
+     '//svrl:failed-assert', namespaces={'svrl': SVRL_NS})
+ 
+-
+ # RelaxNG validator for schematron schemas
+-schematron_schema_valid = _etree.RelaxNG(
+-    file=os.path.join(_resources_dir, 'rng', 'iso-schematron.rng'))
++schematron_schema_valid_supported = False
++try:
++    schematron_schema_valid = _etree.RelaxNG(
++        file=os.path.join(_resources_dir, 'rng', 'iso-schematron.rng'))
++    schematron_schema_valid_supported = True
++except _etree.RelaxNGParseError:
++    # Some distributions delete the file due to licensing issues.
++    def schematron_schema_valid(arg):
++        raise NotImplementedError("Validating the ISO schematron requires iso-schematron.rng")
+ 
+ 
+ def stylesheet_params(**kwargs):
+@@ -153,6 +159,13 @@ class Schematron(_etree._Validator):
+     report document gets stored and can be accessed as the ``validation_report``
+     property.
+ 
++    If ``validate_schema`` is set to False, the validation of the schema file
++    itself is disabled.  Validation happens by default after building the full
++    schema, unless the schema validation file cannot be found at import time,
++    in which case the validation gets disabled.  Some lxml distributions exclude
++    this file due to licensing issues.  ISO-Schematron validation can then still
++    be used normally, but the schemas themselves cannot be validated.
++
+     Here is a usage example::
+ 
+       >>> from lxml import etree
+@@ -234,7 +247,8 @@ class Schematron(_etree._Validator):
+     def __init__(self, etree=None, file=None, include=True, expand=True,
+                  include_params={}, expand_params={}, compile_params={},
+                  store_schematron=False, store_xslt=False, store_report=False,
+-                 phase=None, error_finder=ASSERTS_ONLY):
++                 phase=None, error_finder=ASSERTS_ONLY,
++                 validate_schema=schematron_schema_valid_supported):
+         super(Schematron, self).__init__()
+ 
+         self._store_report = store_report
+@@ -273,7 +287,7 @@ class Schematron(_etree._Validator):
+             schematron = self._include(schematron, **include_params)
+         if expand:
+             schematron = self._expand(schematron, **expand_params)
+-        if not schematron_schema_valid(schematron):
++        if validate_schema and not schematron_schema_valid(schematron):
+             raise _etree.SchematronParseError(
+                 "invalid schematron schema: %s" %
+                 schematron_schema_valid.error_log)
+-- 
+2.40.1
+

get-lxml-source.sh

@@ -13,17 +13,16 @@ fi
 
 versionedname=lxml-${version}
 orig_archive=${versionedname}.tar.gz
-new_archive=${versionedname}-no-isoschematron.tar.gz
+new_archive=${versionedname}-no-isoschematron-rng.tar.gz
 
 if [ ! -e ${orig_archive} ]; then
     wget -N https://files.pythonhosted.org/packages/source/l/lxml/${orig_archive}
 fi
 
-deleted_module=lxml-${version}/src/lxml/isoschematron/
-deleted_test=lxml-${version}/src/lxml/tests/test_isoschematron.py
+deleted_directory=lxml-${version}/src/lxml/isoschematron/resources/rng
 
 # tar --delete does not operate on compressed archives, so do
 # gz decompression explicitly
 gzip --decompress ${orig_archive}
-tar -v --delete -f ${orig_archive//.gz} {$deleted_module,$deleted_test}
+tar -v --delete -f ${orig_archive//.gz} ${deleted_directory}
 gzip -cf ${orig_archive//.gz} > ${new_archive}

python3.11-lxml.spec

@@ -3,7 +3,7 @@
 
 Name:           python%{python3_pkgversion}-lxml
 Version:        4.9.2
-Release:        3%{?dist}
+Release:        4%{?dist}
 Summary:        XML processing library combining libxml2/libxslt with the ElementTree API
 
 # The lxml project is licensed under BSD-3-Clause
@@ -12,13 +12,21 @@ Summary:        XML processing library combining libxml2/libxslt with the Elemen
 # .xsl schematron files are under the MIT license
 License:        BSD and MIT
 URL:            https://github.com/lxml/lxml
+
 # We use the get-lxml-source.sh script to generate the tarball
-# without the isoschematron submodule as it contains a problematic
-# license.
+# without the isoschematron RNG validation file under a problematic license.
 # See: https://gitlab.com/fedora/legal/fedora-license-data/-/issues/154
-Source0:         lxml-%{version}-no-isoschematron.tar.gz
+Source0:         lxml-%{version}-no-isoschematron-rng.tar.gz
 Source1:         get-lxml-source.sh
 
+# Make the validation of ISO-Schematron files optional in lxml,
+# depending on the availability of the RNG validation file
+# Rebased from https://github.com/lxml/lxml/commit/4bfab2c821961fb4c5ed8a04e329778c9b09a1df
+# Will be included in lxml 5.0
+Patch:          Make-the-validation-of-ISO-Schematron-files-optional.patch
+# Skip test_isoschematron.test_schematron_invalid_schema_empty without the RNG file
+Patch:          https://github.com/lxml/lxml/pull/380.patch
+
 BuildRequires:  gcc
 BuildRequires:  libxml2-devel
 BuildRequires:  libxslt-devel
@@ -39,12 +47,6 @@ XML Schema, XSLT, C14N and much more.
 %prep
 %autosetup -n lxml-%{version} -p1
 
-# Remove isoschematron module due to problematic license
-sed -i "s/, 'lxml.isoschematron'//" setup.py
-# Remove the doctests for it (the documentation is not shipped)
-# The command [d]eletes all lines from the first pattern to the second
-sed -Ei '/^Schematron$/,/^\(Pre-ISO-Schematron\)$/d' doc/validation.txt
-
 # Remove pregenerated Cython C sources
 # We need to do this after %%pyproject_buildrequires because setup.py errors
 # without Cython and without the .c files.
@@ -70,6 +72,11 @@ cp -a build/lib.%{python3_platform}-*/* src/
 %{python3_sitearch}/lxml-*.egg-info/
 
 %changelog
+* Tue Aug 15 2023 Tomas Orsava <torsava@redhat.com> - 4.9.2-4
+- Bring back the isoschematron submodule,
+  but without the validation of the schema file itself
+- Resolves: RHEL-5570
+
 * Thu Feb 16 2023 Charalampos Stratakis <cstratak@redhat.com> - 4.9.2-3
 - Remove the isoschematron submodule
 

sources

@@ -1 +1 @@
-SHA512 (lxml-4.9.2-no-isoschematron.tar.gz) = 3ee80967dfa69c840a27ad99546727346f8bb058ed64bf93d2aa8959fb3000e78c1227f4b8709d25aaadeb9a4ccdf97a03eb53a1171de8feda93fbeaa15a8b39
+SHA512 (lxml-4.9.2-no-isoschematron-rng.tar.gz) = 7de5a626bb87f2f56f90a3c308199ed79d78796eeb075b5d42df4bccf3606d3fb2e2981fa9206412b675304dd487fce2dfb90b3b3e0793e8d63be2216b325c47