From 1996bea540d1da302113e952a34f1f420f4c9ad2fedf60075027f86ce210b7a7 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Tom=C3=A1=C5=A1=20Chv=C3=A1tal?= Date: Mon, 1 Apr 2019 09:04:03 +0000 Subject: [PATCH] Accepting request 690091 from home:jayvdb:branches:devel:languages:python - Activate and fix test suite, removing bcond 'tests' - Add Requires sgmllib3k to fix exception in Python 3 when using the loose parser: UnboundLocalError: local variable 'feedparser' referenced before assignment and less clear but similarly broken result with the strict parser. - Add catch-gzip-error.patch to fix fetch crash in Python 3 - Add non-ascii-entity-hiding.patch to workaround failure to properly parse non-ASCII based encodings on Python 3.5+ - Add py37.patch to fix parsing crash on Python 3.7 OBS-URL: https://build.opensuse.org/request/show/690091 OBS-URL: https://build.opensuse.org/package/show/devel:languages:python/python-feedparser?expand=0&rev=40 --- catch-gzip-error.patch | 30 +++++++++++++++++++++++++ non-ascii-entity-hiding.patch | 17 ++++++++++++++ py37.patch | 16 +++++++++++++ python-feedparser.changes | 14 ++++++++++++ python-feedparser.spec | 42 ++++++++++++++++++++++++++--------- 5 files changed, 109 insertions(+), 10 deletions(-) create mode 100644 catch-gzip-error.patch create mode 100644 non-ascii-entity-hiding.patch create mode 100644 py37.patch diff --git a/catch-gzip-error.patch b/catch-gzip-error.patch new file mode 100644 index 0000000..41e42cc --- /dev/null +++ b/catch-gzip-error.patch @@ -0,0 +1,30 @@ +diff -ur feedparser-5.2.1-orig/feedparser/feedparser.py feedparser-5.2.1/feedparser/feedparser.py +--- feedparser-5.2.1-orig/feedparser/feedparser.py 2015-07-24 12:00:04.000000000 +0700 ++++ feedparser-5.2.1/feedparser/feedparser.py 2019-03-31 02:26:41.242238792 +0700 +@@ -3865,12 +3869,12 @@ + if gzip and 'gzip' in http_headers.get('content-encoding', ''): + try: + data = gzip.GzipFile(fileobj=_StringIO(data)).read() +- except (IOError, struct.error), e: ++ except (EOFError, IOError, struct.error), e: + # IOError can occur if the gzip header is bad. + # struct.error can occur if the data is damaged. + result['bozo'] = 1 + result['bozo_exception'] = e +- if isinstance(e, struct.error): ++ if isinstance(e, (EOFError, struct.error)): + # A gzip header was found but the data is corrupt. + # Ideally, we should re-request the feed without the + # 'Accept-encoding: gzip' header, but we don't. +diff -ur feedparser-5.2.1-orig/feedparser/feedparsertest.py feedparser-5.2.1/feedparser/feedparsertest.py +--- feedparser-5.2.1-orig/feedparser/feedparsertest.py 2015-07-24 11:45:09.000000000 +0700 ++++ feedparser-5.2.1/feedparser/feedparsertest.py 2019-03-31 04:02:14.470216076 +0700 +@@ -448,7 +456,7 @@ + def test_gzip_struct_error(self): + f = feedparser.parse('http://localhost:8097/tests/compression/gzip-struct-error.gz') + self.assertEqual(f.bozo, 1) +- self.assertTrue(isinstance(f.bozo_exception, struct.error)) ++ self.assertTrue(isinstance(f.bozo_exception, (EOFError, struct.error))) + def test_zlib_good(self): + f = feedparser.parse('http://localhost:8097/tests/compression/deflate.z') + self.assertEqual(f.version, 'atom10') diff --git a/non-ascii-entity-hiding.patch b/non-ascii-entity-hiding.patch new file mode 100644 index 0000000..86e15c2 --- /dev/null +++ b/non-ascii-entity-hiding.patch @@ -0,0 +1,17 @@ +diff -ur feedparser-5.2.1-orig/feedparser/feedparsertest.py feedparser-5.2.1/feedparser/feedparsertest.py +--- feedparser-5.2.1-orig/feedparser/feedparsertest.py 2015-07-24 11:45:09.000000000 +0700 ++++ feedparser-5.2.1/feedparser/feedparsertest.py 2019-03-31 04:02:14.470216076 +0700 +@@ -273,6 +273,13 @@ + &exponential3;""" + doc = codecs.BOM_UTF16_BE + doc.encode('utf-16be') + result = feedparser.parse(doc) ++ if sys.version_info >= (3, 5): ++ # While this isnt the expected result, it is verification that ++ # non-ASCII-compatible encodings cant hide nasty stuff. ++ # Instead, nothing is loaded, which is a bug in a different ++ # area. ++ self.assertEqual(result['feed'], {}) ++ return + self.assertEqual(result['feed']['title'], u'&exponential3') + def test_gb2312_converted_to_gb18030_in_xml_encoding(self): + # \u55de was chosen because it exists in gb18030 but not gb2312 diff --git a/py37.patch b/py37.patch new file mode 100644 index 0000000..c50d5a3 --- /dev/null +++ b/py37.patch @@ -0,0 +1,16 @@ +diff -ur feedparser-5.2.1-orig/feedparser/feedparser.py feedparser-5.2.1/feedparser/feedparser.py +--- feedparser-5.2.1-orig/feedparser/feedparser.py 2015-07-24 12:00:04.000000000 +0700 ++++ feedparser-5.2.1/feedparser/feedparser.py 2019-03-31 02:26:41.242238792 +0700 +@@ -3764,7 +3764,11 @@ + latlons = itertools.imap(float, value.strip().replace(',', ' ').split()) + nxt = latlons.next + while True: +- t = [nxt(), nxt()][::swap and -1 or 1] ++ # Python 3.7 support ++ try: ++ t = [nxt(), nxt()][::swap and -1 or 1] ++ except StopIteration: ++ return + if dims == 3: + t.append(nxt()) + yield tuple(t) diff --git a/python-feedparser.changes b/python-feedparser.changes index 4eb9843..9f1b6d7 100644 --- a/python-feedparser.changes +++ b/python-feedparser.changes @@ -1,3 +1,17 @@ +------------------------------------------------------------------- +Sat Mar 30 22:06:01 UTC 2019 - John Vandenberg + +- Activate test suite +- Add Requires sgmllib3k to fix exception in Python 3 when + using the loose parser: + UnboundLocalError: local variable 'feedparser' referenced before assignment + and less clear but similarly broken result with the strict + parser. +- Add catch-gzip-error.patch to fix fetch crash in Python 3 +- Add non-ascii-entity-hiding.patch to workaround failure + to properly parse non-ASCII based encodings on Python 3.5+ +- Add py37.patch to fix parsing crash on Python 3.7 + ------------------------------------------------------------------- Tue Dec 4 12:47:57 UTC 2018 - Matej Cepl diff --git a/python-feedparser.spec b/python-feedparser.spec index fc21a82..c422833 100644 --- a/python-feedparser.spec +++ b/python-feedparser.spec @@ -1,7 +1,7 @@ # # spec file for package python-feedparser # -# Copyright (c) 2018 SUSE LINUX GmbH, Nuernberg, Germany. +# Copyright (c) 2019 SUSE LINUX GmbH, Nuernberg, Germany. # # All modifications and additions to the file contributed by third parties # remain the property of their copyright owners, unless otherwise agreed @@ -17,8 +17,6 @@ %{?!python_module:%define python_module() python-%{**} python3-%{**}} -# Tests require a network connection -%bcond_without tests Name: python-feedparser Version: 5.2.1 Release: 0 @@ -27,12 +25,26 @@ License: BSD-2-Clause Group: Development/Libraries/Python URL: https://github.com/kurtmckee/feedparser Source: https://files.pythonhosted.org/packages/source/f/feedparser/feedparser-%{version}.tar.bz2 +Patch0: py37.patch +Patch1: non-ascii-entity-hiding.patch +# Similar to https://github.com/kurtmckee/feedparser/commit/b3d9463.patch +# However the "gets overwritten as xml.sax.SAXException later" is not +# happening in version 5.2.1. +Patch2: catch-gzip-error.patch BuildRequires: %{python_module setuptools} BuildRequires: %{python_module xml} BuildRequires: fdupes BuildRequires: python-rpm-macros -BuildRequires: python3-2to3 +BuildRequires: python3-sgmllib3k +# Tests fail in Python 2 and 3 when chardet is installed +#BuildRequires: python3-chardet Requires: python-xml +# chardet is an optional dependency, but some tests fail when it is present +Recommends: python-chardet +%ifpython3 +# If sgmllib is not available, the parser can fail with undefined local variable +Requires: python3-sgmllib3k +%endif BuildArch: noarch %python_subpackages @@ -42,8 +54,18 @@ A universal feed parser module for Python that handles RSS 0.9x, RSS 1.0, RSS %prep %setup -q -n feedparser-%{version} +%autopatch -p1 + find . -type f -exec chmod 0644 {} \; # 5.2.1 had executable bit set on almost all files +# In version 5.2.1, it contains only test data, so move it out of install +mv feedparser/tests . + +# Move the test module out of the runtime install, into top level directory +# so it can find the test data, and manually run 2to3. +python3 -m lib2to3 -w -n -o . --add-suffix=-%{python3_bin_suffix} --no-diffs feedparser/feedparsertest.py +mv feedparser/feedparsertest.py feedparsertest.py-%{python2_bin_suffix} + %build %python_build @@ -51,13 +73,13 @@ find . -type f -exec chmod 0644 {} \; # 5.2.1 had executable bit set on almost %python_install %python_expand %fdupes %{buildroot}%{$python_sitelib} -%if %{with tests} %check -# Tests only work on Python 2.x -pushd feedparser -python2 feedparsertest.py -popd -%endif +%{python_expand export PYTHONPATH=%{buildroot}%{$python_sitelib}:${PWD} +cp feedparsertest.py-%{$python_bin_suffix} feedparsertest.py +$python -c 'import feedparser; assert feedparser._XML_AVAILABLE == 1; assert feedparser._SGML_AVAILABLE == 1' +$python -c 'import feedparsertest; assert feedparsertest._UTF32_AVAILABLE == 1' +$python feedparsertest.py +} %files %{python_files} %license LICENSE