From 5ac8e73c24537b0a311137438f2cffc4e2696039 Mon Sep 17 00:00:00 2001
From: "Benjamin A. Beasley" <code@musicinmybrain.net>
Date: Sat, 25 Jun 2022 12:53:26 -0400
Subject: [PATCH] Drop test data with problematic license status

---
 .gitignore          |  1 +
 get_source          | 34 ++++++++++++++++++++++++++++++++++
 python-chardet.spec | 44 ++++++++++++++++++++++----------------------
 sources             |  2 +-
 4 files changed, 58 insertions(+), 23 deletions(-)
 create mode 100755 get_source

diff --git a/.gitignore b/.gitignore
index bb18b7a..4a4974b 100644
--- a/.gitignore
+++ b/.gitignore
@@ -1,2 +1,3 @@
 /chardet-*.tar.gz
 /chardet-*/
+/chardet-4.0.0-filtered.tar.xz
diff --git a/get_source b/get_source
new file mode 100755
index 0000000..8268d9e
--- /dev/null
+++ b/get_source
@@ -0,0 +1,34 @@
+#!/bin/sh
+set -o errexit
+set -o nounset
+
+if [ "$#" != '1' ]
+then
+	cat 1>&2 <<EOF
+Usage: $0 VERSION
+
+Downloads the requested version and creates a filtered source tarball.
+EOF
+	exit 1
+fi
+
+VERSION="${1}"
+OUTDIR="${PWD}"
+TMPDIR="$(mktemp -d)"
+trap "rm -rf '${TMPDIR}'" INT TERM EXIT
+
+cd "${TMPDIR}"
+URL="$(rpm -E "%{pypi_source chardet ${VERSION}}")"
+echo "--> Downloading: ${URL}" 1>&2
+curl -L -O "${URL}"
+
+ARCHIVE="$(find . -mindepth 1 -maxdepth 1 -type f -name '*.tar.gz' -print -quit)"
+echo "--> Extracting: $(basename "${ARCHIVE}")" 1>&2
+tar -xzf "${ARCHIVE}"
+echo '--> Removing tests due to licensing issues' 1>&2
+rm -rvf */tests/
+FILTERED="$(basename "${ARCHIVE}" .tar.gz)-filtered.tar.xz"
+echo "--> Re-archiving: ${FILTERED}" 1>&2
+tar -c "${FILTERED}" */ | xz -9e > "${FILTERED}"
+mv -v "${FILTERED}.xz" "${OUTDIR}"
+echo 'Done.' 1>&2
diff --git a/python-chardet.spec b/python-chardet.spec
index 56dc09d..2e38257 100644
--- a/python-chardet.spec
+++ b/python-chardet.spec
@@ -1,6 +1,3 @@
-# Circular dependency on pytest
-%bcond_without tests
-
 # Sphinx-generated HTML documentation is not suitable for packaging; see
 # https://bugzilla.redhat.com/show_bug.cgi?id=2006555 for discussion.
 #
@@ -12,31 +9,34 @@ Version:        4.0.0
 Release:        %autorelease
 Summary:        Python character encoding detector
 
-# The entire source is LGPLv2+, except:
-#
-# - tests/ascii/howto.diveintomark.org.xml is CC-BY-SA (but is not installed)
+# The entire source, after tests/ are filtered, is LGPLv2+. See the comment
+# above Source0 for more details on tests/.
 #
 # See also:
-# clarify Trove license classifier
-# https://github.com/chardet/chardet/issues/162
-# Clarify LGPL version in License trove classifier
-# https://github.com/chardet/chardet/pull/255
+#   clarify Trove license classifier
+#   https://github.com/chardet/chardet/issues/162
+#   Clarify LGPL version in License trove classifier
+#   https://github.com/chardet/chardet/pull/255
 License:        LGPLv2+
 URL:            https://github.com/chardet/chardet
-Source0:        %{pypi_source chardet}
+# A filtered source tarball, obtained by (see Source1):
+#
+#   ./get_source %%{version}
+#
+# is required because the contents of tests/ are under various undocumented
+# licenses and are, for the most part, not freely redistributable. See:
+#
+#   problematic licensing of /tests?
+#   https://github.com/chardet/chardet/issues/231
+Source0:        chardet-%{version}-filtered.tar.xz
+Source1:        get_source
 # Hand-written for Fedora in groff_man(7) format based on --help output
-Source1:        chardetect.1
+Source2:        chardetect.1
 
 BuildArch:      noarch
 
 BuildRequires:  python3-devel
 
-%if %{with tests}
-BuildRequires:  python3dist(pytest)
-# Optional; enables additional tests (see test.py)
-BuildRequires:  python3dist(hypothesis)
-%endif
-
 %if %{with doc_pdf}
 BuildRequires:  make
 BuildRequires:  python3dist(sphinx)
@@ -112,13 +112,13 @@ PYTHONPATH="${PWD}" %make_build -C docs latex SPHINXOPTS='%{?_smp_mflags}'
 %install
 %pyproject_install
 %pyproject_save_files chardet
-install -t '%{buildroot}%{_mandir}/man1' -D -p -m 0644 '%{SOURCE1}'
+install -t '%{buildroot}%{_mandir}/man1' -D -p -m 0644 '%{SOURCE2}'
 
 
-%if %{with tests}
 %check
-%pytest -v
-%endif
+# We cannot run the upstream tests because they would require data files with
+# problematic license status.
+%pyproject_check_import
 
 
 %files -n python3-chardet -f %{pyproject_files}
diff --git a/sources b/sources
index dad62a1..f73c80e 100644
--- a/sources
+++ b/sources
@@ -1 +1 @@
-SHA512 (chardet-4.0.0.tar.gz) = ebd7f420e1094445270db993f6373ffe7370419e002b0bb13299dc6c9b0f7c4e77b0f44f871fba6371e6869e7c86728514367db377e3137487a3acf50cb81e96
+SHA512 (chardet-4.0.0-filtered.tar.xz) = ac510d460dc65233bda8d11964d6d9a130a53ef64035393d05c52ab6c4e0b7a7ca6f34108c21771fb40236d71e7de62dfe6c6f9fa6cd1722a355116930b5e12b