Use --hardlink-dupes in %py_byte_compile and brp-python-bytecompile

(for Python 3.9+)

Resolves: rhbz#1977895

brp-python-bytecompile

@@ -29,7 +29,6 @@ fi
 # Python < 3.4 (inc. Python 2) uses compileall module from stdlib with some hacks
 # When we drop support for Python 2, we'd be able to use all compileall2 features like:
 # - -s and -p options to manipulate with a path baked into pyc files instead of $real_libdir
-# - -o 0 -o 1 to produce multiple files in one run - each with a different optimization level - instead of $options
 # - removed useless $depth - both compileall and compileall2 are limited by sys.getrecursionlimit()
 # These changes will make this script much simpler
 # In Python >= 3.9, compileall2 was merged back to standard library (compileall) so we can use it directly again.
@@ -49,6 +48,12 @@ function python_bytecompile()
 	#
 	if [ "$python_version" -ge 39 ]; then
 
+		# For Python 3.9+, we compile all opt levels in one go: only
+		# when $options is empty.
+		if [ -n "$options" ]; then
+		    return
+		fi
+
 		[ ! -z $exclude ] && exclude="-x '$exclude'"
 		# -q disables verbose output
 		# -f forces the process to overwrite existing compiled files
@@ -57,7 +62,7 @@ function python_bytecompile()
 		# -x and -e together implements the same functionality as the Filter class below
 		# -s strips $RPM_BUILD_ROOT from the path
 		# -p prepends the leading slash to the path to make it absolute
-		$python_binary -B $options -m compileall -q -f $exclude -s "$RPM_BUILD_ROOT" -p / -e "$RPM_BUILD_ROOT" "$python_libdir"
+		$python_binary -B -m compileall -o 0 -o 1 -q -f $exclude -s "$RPM_BUILD_ROOT" -p / --hardlink-dupes -e "$RPM_BUILD_ROOT" "$python_libdir"
 
 	#
 	# Python 3.4 and higher
@@ -133,6 +138,7 @@ do
 	fi
 
 	# Generate optimized (.pyo) byte-compiled files.
+	# N.B. For Python 3.9+, this call does nothing
 	python_bytecompile "-O" "$python_binary" "" "$python_libdir" "$depth" "$real_libdir"
 	if [ $? -ne 0 -a 0$errors_terminate -ne 0 ]; then
 		# One or more of the files had a syntax error

macros.pybytecompile

@@ -34,7 +34,7 @@ py3_byte_compile () {\
 py39_byte_compile () {\
     python_binary="env PYTHONHASHSEED=0 %1"\
     bytecode_compilation_path="%2"\
-    $python_binary -s -B -m compileall -o 0 -o 1 -s $RPM_BUILD_ROOT -p / $bytecode_compilation_path \
+    $python_binary -s -B -m compileall -o 0 -o 1 -s $RPM_BUILD_ROOT -p / --hardlink-dupes $bytecode_compilation_path \
 }\
 \
 # Path to intepreter should not contain any arguments \

python-rpm-macros.spec

@@ -47,7 +47,7 @@ elseif posix.stat('macros.python-srpm') then
 end
 }
 Version:        %{__default_python3_version}
-Release:        7%{?dist}
+Release:        8%{?dist}
 
 BuildArch:      noarch
 
@@ -141,6 +141,11 @@ install -m 755 brp-* %{buildroot}%{_rpmconfigdir}/redhat/
 
 
 %changelog
+* Thu Sep 09 2021 Petr Viktorin <pviktori@redhat.com> - 3.10-8
+- Use --hardlink-dupes in %%py_byte_compile and brp-python-bytecompile
+  (for Python 3.9+)
+- Resolves: rhbz#1977895
+
 * Fri Jul 23 2021 Fedora Release Engineering <releng@fedoraproject.org> - 3.9-7
 - Rebuilt for https://fedoraproject.org/wiki/Fedora_35_Mass_Rebuild
 

tests/pythontest.spec

@@ -19,15 +19,35 @@ echo "print()" > %{buildroot}%{basedir}/directory/to/test/recursion/file_in_dir.
 %py_byte_compile %{python3} %{buildroot}%{basedir}/file.py
 %py_byte_compile %{python3} %{buildroot}%{basedir}/directory
 
+# Files in sitelib are compiled automatically by brp-python-bytecompile
+mkdir -p %{buildroot}%{python3_sitelib}/directory/
+echo "print()" > %{buildroot}%{python3_sitelib}/directory/file.py
+
 %check
+LOCATIONS="%{buildroot}%{basedir} %{buildroot}%{python3_sitelib}/directory/"
+
 # Count .py and .pyc files
-PY=$(find %{buildroot}%{basedir} -name "*.py" | wc -l)
+PY=$(find $LOCATIONS -name "*.py" | wc -l)
-PYC=$(find %{buildroot}%{basedir} -name "*.pyc" | wc -l)
+PYC=$(find $LOCATIONS -name "*.pyc" | wc -l)
+
+# We should have 3 .py files
+test $PY -eq 3
 
 # Every .py file should be byte-compiled to two .pyc files (optimization level 0 and 1)
 # so we should have two times more .pyc files than .py files
 test $(expr $PY \* 2) -eq $PYC
 
+# In this case the .pyc files should be identical across omtimization levels
+# (they don't use docstrings and assert staements)
+# So they should be hardlinked; the number of distinct inodes should match the
+# number of source files. (Or be smaller, if the dupe detection is done
+# across all files.)
+
+INODES=$(stat --format %i $(find $LOCATIONS -name "*.pyc") | sort -u | wc -l)
+test $PY -ge $INODES
+
+
 %files
 %pycached %{basedir}/file.py
 %pycached %{basedir}/directory/to/test/recursion/file_in_dir.py
+%pycached %{python3_sitelib}/directory/file.py