RPM source_date_epoch test should check rpmbuild version, not system rpm

[nacl]

We have a skip in this test, but we dumbly check the system "rpm" binary than the rpmbuild binary used to build the RPM in the first place. We have skipping logic here:

rules_pkg/pkg/tests/rpm/source_date_epoch/rpm_contents_vs_manifest_test.py

Lines 57 to 64 in 9748dfa

	# These tests will fail on unsupported versions of rpm(8), so skip them
	# in that case.
	if not self.rpmBinSupportsSourceDateEpoch():
	self.skipTest("RPM version too old to support SOURCE_DATE_EPOCH."
	" Must be {} or newer (is {})".format(
	version_to_string(self.SOURCE_DATE_EPOCH_MIN_VERSION),
	version_to_string(self.rpm_bin_version),
	))

But it could not be working for numerous reasons. Investigate further.

Originally posted by @nacl in #443 (comment)

[nacl]

See also #443 (comment)

[aiuto]

So... I notice that the tests fail on my machine, but my rpmbuild version is 4.17.0
The suspicious thing is

rror: Unable to open sqlite database zzz/.rpmdb/rpmdb.sqlite: unable to open database file
error: cannot open Packages index using sqlite - Operation not permitted (1)
error: cannot open Packages database in /zzz/.rpmdb
Ferror: Unable to open sqlite database /zzz/.rpmdb/rpmdb.sqlite: unable to open database file

then

  self.assertEqual(actual_epoch, EXPECTED_EPOCH)
AssertionError: '1643058161' != '1616472000'
- 1643058161
+ 1616472000

So, I am not sure if there is more to it than that.

[aiuto]

Also, sending a PR to get rpmbuild version into the toolchain.
if you want to go that route, we can make source date epoch support a condition, and use exec_compatible_with on the tests which would require it. Much like have_rpmbuild.