From 0d66adbc178ea4bcdf5ad2b91e2580c17e13b659 Mon Sep 17 00:00:00 2001
From: Jan Kaluza <jkaluza@redhat.com>
Date: Wed, 5 Sep 2018 08:39:48 +0200
Subject: [PATCH] Add list of built RPMs to architecture specific modulemd
 files in CG builds.

---
 Vagrantfile                                   |   2 +
 docker/Dockerfile-tests                       |   1 +
 .../builder/KojiContentGenerator.py           | 144 +++++++++++-
 tests/test_content_generator.py               | 220 +++++++++++++++++-
 4 files changed, 362 insertions(+), 5 deletions(-)

diff --git a/Vagrantfile b/Vagrantfile
index c1f4aee2..156bab76 100644
--- a/Vagrantfile
+++ b/Vagrantfile
@@ -30,9 +30,11 @@ $script = <<SCRIPT
         python-sqlalchemy \
         python-virtualenv \
         python-futures \
+        python2-pungi \
         python3 \
         python3-devel \
         python3-docutils \
+        python3-pungi \
         redhat-rpm-config \
         redhat-rpm-config \
         rpm-build \
diff --git a/docker/Dockerfile-tests b/docker/Dockerfile-tests
index 9ed2061b..065f3857 100644
--- a/docker/Dockerfile-tests
+++ b/docker/Dockerfile-tests
@@ -35,6 +35,7 @@ RUN yum -y install \
     python-six \
     python-solv \
     python-sqlalchemy \
+    python2-pungi \
     # Test-only dependencies
     python-flake8 \
     python-mock \
diff --git a/module_build_service/builder/KojiContentGenerator.py b/module_build_service/builder/KojiContentGenerator.py
index 383f7446..2940abd9 100644
--- a/module_build_service/builder/KojiContentGenerator.py
+++ b/module_build_service/builder/KojiContentGenerator.py
@@ -38,6 +38,7 @@ import kobo.rpmlib
 
 from six import text_type
 import koji
+import pungi.arch
 
 from module_build_service import log, build_logs, Modulemd
 
@@ -49,6 +50,16 @@ def get_session(config, owner):
     return KojiModuleBuilder.get_session(config, owner)
 
 
+def koji_retrying_multicall_map(*args, **kwargs):
+    """
+    Wrapper around KojiModuleBuilder.koji_retrying_multicall_map, because
+    we cannot import that method normally because of import loop.
+    """
+    from module_build_service.builder.KojiModuleBuilder import \
+        koji_retrying_multicall_map as multicall
+    return multicall(*args, **kwargs)
+
+
 class KojiContentGenerator(object):
     """ Class for handling content generator imports of module builds into Koji """
 
@@ -204,12 +215,36 @@ class KojiContentGenerator(object):
             # If the tag doesn't exist.. then there are no rpms in that tag.
             return []
 
-        # Extract some srpm-level info from the build attach it to each rpm
+        # Get the exclusivearch and excludearch lists for each RPM.
+        # The exclusivearch and excludearch lists are set in source RPM from which the RPM
+        # was built.
+        # Create temporary dict with source RPMs in rpm_id:build_id format.
+        src_rpms_ids = {rpm["id"]: rpm["build_id"] for rpm in rpms if rpm["arch"] == "src"}
+        # Prepare the arguments for Koji multicall.
+        # We will call session.getRPMHeaders(...) for each SRC RPM to get exclusivearch and
+        # excludearch headers.
+        multicall_kwargs = [{"rpmID": rpm_id, "headers": ["exclusivearch", "excludearch"]}
+                            for rpm_id in src_rpms_ids.keys()]
+        src_rpms_headers = koji_retrying_multicall_map(
+            session, session.getRPMHeaders, list_of_kwargs=multicall_kwargs)
+
+        # Temporary dict with build_id as a key to find builds easily.
         builds = {build['build_id']: build for build in builds}
+
+        # Handle the multicall result. For each build associated with the source RPM,
+        # store the exclusivearch and excludearch lists.
+        for build_id, headers in zip(src_rpms_ids.values(), src_rpms_headers):
+            builds[build_id]["exclusivearch"] = headers["exclusivearch"]
+            builds[build_id]["excludearch"] = headers["excludearch"]
+
+        # Check each RPM and fill-in additional data from its build to get them
+        # easily in other methods.
         for rpm in rpms:
             idx = rpm['build_id']
             rpm['srpm_name'] = builds[idx]['name']
             rpm['srpm_nevra'] = builds[idx]['nvr']
+            rpm['exclusivearch'] = builds[idx]['exclusivearch']
+            rpm['excludearch'] = builds[idx]['excludearch']
 
         return rpms
 
@@ -333,7 +368,7 @@ class KojiContentGenerator(object):
                     self._koji_rpm_to_component_record(rpm))
         else:
             # Check the RPM artifacts built for this architecture in modulemd file,
-            # find the matchign RPM in the `rpms_dict` comming from Koji and use it
+            # find the matching RPM in the `rpms_dict` coming from Koji and use it
             # to generate list of components for this architecture.
             # We cannot simply use the data from MMD here without `rpms_dict`, because
             # RPM sigmd5 signature is not stored in MMD.
@@ -384,10 +419,109 @@ class KojiContentGenerator(object):
 
         return ret
 
+    def _fill_in_rpms_list(self, mmd, arch):
+        """
+        Fills in the list of built RPMs in architecture specific `mmd` for `arch`
+        using the data from `self.rpms_dict`.
+
+        :param Modulemd.Module mmd: MMD to add built RPMs to.
+        :param str arch: Architecture for which to add RPMs.
+        :rtype: Modulemd.Module
+        :return: MMD with built RPMs filled in.
+        """
+        # List of all architectures compatible with input architecture including
+        # the multilib architectures.
+        # Example input/output:
+        #   "x86_64" -> ['x86_64', 'athlon', 'i686', 'i586', 'i486', 'i386', 'noarch']
+        #   "i686" -> ['i686', 'i586', 'i486', 'i386', 'noarch']
+        compatible_arches = pungi.arch.get_compatible_arches(arch, multilib=True)
+        # List of only multilib architectures.
+        # For example:
+        #   "x86_64" -> ['athlon', 'i386', 'i586', 'i486', 'i686']
+        #   "i686" -> []
+        multilib_arches = set(compatible_arches) - set(
+            pungi.arch.get_compatible_arches(arch))
+
+        # Modulemd.SimpleSet into which we will add the RPMs.
+        rpm_artifacts = Modulemd.SimpleSet()
+
+        # Check each RPM in `self.rpms_dict` to find out if it can be included in mmd
+        # for this architecture.
+        for nevra, rpm in self.rpms_dict.items():
+            srpm = rpm["srpm_name"]
+
+            # Skip the RPM if it is excluded on this arch or exclusive
+            # for different arch.
+            if rpm["excludearch"] and set(rpm["excludearch"]) & set(compatible_arches):
+                continue
+            if rpm["exclusivearch"] and not set(rpm["exclusivearch"]) & set(compatible_arches):
+                continue
+
+            # Check the "whitelist" buildopts section of MMD.
+            # When "whitelist" is defined, it overrides component names in
+            # `mmd.get_rpm_components()`. The whitelist is used when module needs to build
+            # package with different SRPM name than the package name. This is case for example
+            # for software collections where SRPM name can be "httpd24-httpd", but package name
+            # is still "httpd". In this case, get_rpm_components() would contain "httpd", but the
+            # rpm["srpm_name"] would be "httpd24-httpd".
+            whitelist = None
+            buildopts = mmd.get_buildopts()
+            if buildopts:
+                whitelist = buildopts.get_rpm_whitelist()
+                if whitelist:
+                    if srpm not in whitelist:
+                        # Package is not in the whitelist, skip it.
+                        continue
+
+            # If there is no whitelist, just check that the SRPM name we have here
+            # exists in the list of components.
+            # In theory, there should never be situation where modular tag contains
+            # some RPM built from SRPM not included in get_rpm_components() or in whitelist,
+            # but the original Pungi code checked for this case.
+            if not whitelist and srpm not in mmd.get_rpm_components().keys():
+                continue
+
+            # Do not include this RPM if it is filtered.
+            if rpm["name"] in mmd.get_rpm_filter().get():
+                continue
+
+            # Skip the rpm if it's built for multilib arch, but
+            # multilib is not enabled for this srpm in MMD.
+            try:
+                mmd_component = mmd.get_rpm_components()[srpm]
+                multilib = mmd_component.get_multilib()
+                multilib = multilib.get() if multilib else set()
+                # The `multilib` set defines the list of architectures for which
+                # the multilib is enabled.
+                #
+                # Filter out RPMs from multilib architectures if multilib is not
+                # enabled for current arch. Keep the RPMs from non-multilib compatible
+                # architectures.
+                if arch not in multilib and rpm["arch"] in multilib_arches:
+                    continue
+            except KeyError:
+                # TODO: This exception is raised only when "whitelist" is used.
+                # Since components in whitelist have different names than ones in
+                # components list, we won't find them there.
+                # We would need to track the RPMs srpm_name from whitelist back to
+                # original package name used in MMD's components list. This is possible
+                # but original Pungi code is not doing that. This is TODO for future
+                # improvements.
+
+                # No such component, disable any multilib
+                if rpm["arch"] not in ("noarch", arch):
+                    continue
+
+            # Add RPM to packages.
+            rpm_artifacts.add(nevra)
+
+        mmd.set_rpm_artifacts(rpm_artifacts)
+        return mmd
+
     def _finalize_mmd(self, arch):
         """
         Finalizes the modulemd:
-            - TODO: Fills in the list of built RPMs respecting filters, whitelist and multilib.
+            - Fills in the list of built RPMs respecting filters, whitelist and multilib.
             - TODO: Fills in the list of licences.
 
         :param str arch: Name of arch to generate the final modulemd for.
@@ -395,7 +529,9 @@ class KojiContentGenerator(object):
         :return: Finalized modulemd string.
         """
         mmd = self.module.mmd()
-        # TODO: Fill in the list of built RPMs.
+        # Fill in the list of built RPMs.
+        mmd = self._fill_in_rpms_list(mmd, arch)
+
         # TODO: Fill in the licences.
         return unicode(mmd.dumps())
 
diff --git a/tests/test_content_generator.py b/tests/test_content_generator.py
index c8e2bee1..7d9c41a9 100644
--- a/tests/test_content_generator.py
+++ b/tests/test_content_generator.py
@@ -27,9 +27,10 @@ from os import path
 
 import module_build_service.messaging
 import module_build_service.scheduler.handlers.repos # noqa
-from module_build_service import models, conf, build_logs
+from module_build_service import models, conf, build_logs, Modulemd
 
 from mock import patch, Mock, MagicMock, call, mock_open
+import kobo.rpmlib
 
 from tests import init_data
 
@@ -301,3 +302,220 @@ class TestBuild:
             'filesize': 315,
             'type': 'file'
         }
+
+    @patch("module_build_service.builder.KojiContentGenerator.get_session")
+    def test_koji_rpms_in_tag(self, get_session):
+        koji_session = MagicMock()
+        koji_session.getUser.return_value = GET_USER_RV
+        koji_session.getTag.return_value = {"arches": "x86_64"}
+
+        rpms = [
+            {
+                'id': 1,
+                'arch': 'src',
+                'build_id': 875991,
+                'name': 'module-build-macros',
+                'release': '1.module_92011fe6',
+                'version': '0.1'
+            },
+            {
+                'id': 2,
+                'arch': 'noarch',
+                'build_id': 875991,
+                'name': 'module-build-macros',
+                'release': '1.module_92011fe6',
+                'version': '0.1'
+            },
+            {
+                'id': 3,
+                'arch': 'src',
+                'build_id': 875636,
+                'name': 'ed',
+                'release': '2.module_bd6e0eb1',
+                'version': '1.14.1'
+            },
+            {
+                'id': 4,
+                'arch': 'x86_64',
+                'build_id': 875636,
+                'name': 'ed',
+                'release': '2.module_bd6e0eb1',
+                'version': '1.14.1'
+            },
+        ]
+
+        builds = [
+            {
+                'build_id': 875636,
+                'name': 'ed',
+                'release': '2.module_bd6e0eb1',
+                'version': '1.14.1',
+                'nvr': 'ed-2.module_bd6e0eb1-1.14.1',
+            },
+            {
+                'build_id': 875991,
+                'name': 'module-build-macros',
+                'release': '1.module_92011fe6',
+                'version': '0.1',
+                'nvr': 'module-build-macros-0.1-1.module_92011fe6',
+            }
+        ]
+
+        koji_session.listTaggedRPMS.return_value = (rpms, builds)
+        koji_session.multiCall.side_effect = [
+            # getRPMHeaders response
+            [[{'excludearch': ["x86_64"], 'exclusivearch': []}],
+             [{'excludearch': [], 'exclusivearch': ["x86_64"]}]]
+        ]
+        get_session.return_value = koji_session
+
+        rpms = self.cg._koji_rpms_in_tag("tag")
+        for rpm in rpms:
+            # We want to mainly check the excludearch and exclusivearch code.
+            if rpm["name"] == "module-build-macros":
+                assert rpm["excludearch"] == ["x86_64"]
+            else:
+                assert rpm["exclusivearch"] == ["x86_64"]
+
+    def _add_test_rpm(self, nevra, srpm_name=None, multilib=None,
+                      koji_srpm_name=None, excludearch=None, exclusivearch=None):
+        """
+        Helper method to add test RPM to ModuleBuild used by KojiContentGenerator
+        and also to Koji tag used to generate the Content Generator build.
+
+        :param str nevra: NEVRA of the RPM to add.
+        :param str srpm_name: Name of SRPM the added RPM is built from.
+        :param list multilib: List of architecture for which the multilib should be turned on.
+        :param str koji_srpm_name: If set, overrides the `srpm_name` in Koji tag. This is
+            needed to test the case when the built "src" package has different name than
+            the package in Koji. This is for example case of software collections where
+            `srpm_name` is "httpd" but `koji_srpm_name` would be "httpd24-httpd".
+        :param list excludearch: List of architectures this package is excluded from.
+        :param list exclusivearch: List of architectures this package is exclusive for.
+        """
+        parsed_nevra = kobo.rpmlib.parse_nvra(nevra)
+        parsed_nevra["payloadhash"] = "hash"
+        if koji_srpm_name:
+            parsed_nevra["srpm_name"] = koji_srpm_name
+        else:
+            parsed_nevra["srpm_name"] = srpm_name
+        parsed_nevra["excludearch"] = excludearch or []
+        parsed_nevra["exclusivearch"] = exclusivearch or []
+        self.cg.rpms.append(parsed_nevra)
+        self.cg.rpms_dict[nevra] = parsed_nevra
+
+        mmd = self.cg.module.mmd()
+        if srpm_name not in mmd.get_rpm_components().keys():
+            component = Modulemd.ComponentRpm()
+            component.set_name(srpm_name)
+            component.set_rationale("foo")
+
+            if multilib:
+                multilib_set = Modulemd.SimpleSet()
+                for arch in multilib:
+                    multilib_set.add(arch)
+                component.set_multilib(multilib_set)
+
+            mmd.add_rpm_component(component)
+            self.cg.module.modulemd = mmd.dumps()
+            self.cg.modulemd = mmd.dumps()
+
+    def test_fill_in_rpms_list(self):
+        self._add_test_rpm("dhcp-libs-12:4.3.5-5.module_2118aef6.x86_64", "dhcp")
+        self._add_test_rpm("dhcp-libs-12:4.3.5-5.module_2118aef6.i686", "dhcp")
+        self._add_test_rpm("perl-Tangerine-12:4.3.5-5.module_2118aef6.x86_64", "perl-Tangerine")
+        self._add_test_rpm("perl-Tangerine-12:4.3.5-5.module_2118aef6.i686", "perl-Tangerine")
+
+        mmd = self.cg.module.mmd()
+        mmd = self.cg._fill_in_rpms_list(mmd, "x86_64")
+
+        # Only x86_64 packages should be filled in, because we requested x86_64 arch.
+        assert set(mmd.get_rpm_artifacts().get()) == set([
+            "dhcp-libs-12:4.3.5-5.module_2118aef6.x86_64",
+            "perl-Tangerine-12:4.3.5-5.module_2118aef6.x86_64"])
+
+    def test_fill_in_rpms_exclusivearch(self):
+        self._add_test_rpm("dhcp-libs-12:4.3.5-5.module_2118aef6.noarch", "dhcp",
+                           exclusivearch=["x86_64"])
+        self._add_test_rpm("perl-Tangerine-12:4.3.5-5.module_2118aef6.noarch", "perl-Tangerine",
+                           exclusivearch=["ppc64le"])
+
+        mmd = self.cg.module.mmd()
+        mmd = self.cg._fill_in_rpms_list(mmd, "x86_64")
+
+        # Only dhcp-libs should be filled in, because perl-Tangerine has different
+        # exclusivearch.
+        assert set(mmd.get_rpm_artifacts().get()) == set([
+            "dhcp-libs-12:4.3.5-5.module_2118aef6.noarch"])
+
+    def test_fill_in_rpms_excludearch(self):
+        self._add_test_rpm("dhcp-libs-12:4.3.5-5.module_2118aef6.noarch", "dhcp",
+                           excludearch=["x86_64"])
+        self._add_test_rpm("perl-Tangerine-12:4.3.5-5.module_2118aef6.noarch", "perl-Tangerine",
+                           excludearch=["ppc64le"])
+
+        mmd = self.cg.module.mmd()
+        mmd = self.cg._fill_in_rpms_list(mmd, "x86_64")
+
+        # Only perl-Tangerine should be filled in, because dhcp-libs is excluded from x86_64.
+        assert set(mmd.get_rpm_artifacts().get()) == set([
+            "perl-Tangerine-12:4.3.5-5.module_2118aef6.noarch"])
+
+    def test_fill_in_rpms_rpm_whitelist(self):
+        self._add_test_rpm("dhcp-libs-12:4.3.5-5.module_2118aef6.x86_64", "dhcp",
+                           koji_srpm_name="python27-dhcp")
+        self._add_test_rpm("dhcp-libs-12:4.3.5-5.module_2118aef6.i686", "dhcp",
+                           koji_srpm_name="python27-dhcp")
+        self._add_test_rpm("perl-Tangerine-12:4.3.5-5.module_2118aef6.x86_64", "perl-Tangerine",
+                           koji_srpm_name="foo-perl-Tangerine")
+        self._add_test_rpm("perl-Tangerine-12:4.3.5-5.module_2118aef6.i686", "perl-Tangerine",
+                           koji_srpm_name="foo-perl-Tangerine")
+
+        mmd = self.cg.module.mmd()
+        opts = mmd.get_buildopts()
+        opts.set_rpm_whitelist(["python27-dhcp"])
+        mmd.set_buildopts(opts)
+
+        mmd = self.cg._fill_in_rpms_list(mmd, "x86_64")
+
+        # Only x86_64 dhcp-libs should be filled in, because only python27-dhcp is whitelisted
+        # srpm name.
+        assert set(mmd.get_rpm_artifacts().get()) == set([
+            "dhcp-libs-12:4.3.5-5.module_2118aef6.x86_64"])
+
+    def test_fill_in_rpms_list_filters(self):
+        self._add_test_rpm("dhcp-libs-12:4.3.5-5.module_2118aef6.x86_64", "dhcp")
+        self._add_test_rpm("dhcp-libs-12:4.3.5-5.module_2118aef6.i686", "dhcp")
+        self._add_test_rpm("perl-Tangerine-12:4.3.5-5.module_2118aef6.x86_64", "perl-Tangerine")
+        self._add_test_rpm("perl-Tangerine-12:4.3.5-5.module_2118aef6.i686", "perl-Tangerine")
+
+        mmd = self.cg.module.mmd()
+        filter_list = Modulemd.SimpleSet()
+        filter_list.add("dhcp-libs")
+        mmd.set_rpm_filter(filter_list)
+
+        mmd = self.cg._fill_in_rpms_list(mmd, "x86_64")
+
+        # Only x86_64 perl-Tangerine should be filled in, because dhcp-libs is filtered out.
+        assert set(mmd.get_rpm_artifacts().get()) == set([
+            "perl-Tangerine-12:4.3.5-5.module_2118aef6.x86_64"])
+
+    def test_fill_in_rpms_list_multilib(self):
+        self._add_test_rpm("dhcp-libs-12:4.3.5-5.module_2118aef6.x86_64", "dhcp",
+                           multilib=["x86_64"])
+        self._add_test_rpm("dhcp-libs-12:4.3.5-5.module_2118aef6.i686", "dhcp",
+                           multilib=["x86_64"])
+        self._add_test_rpm("perl-Tangerine-12:4.3.5-5.module_2118aef6.x86_64", "perl-Tangerine",
+                           multilib=["ppc64le"])
+        self._add_test_rpm("perl-Tangerine-12:4.3.5-5.module_2118aef6.i686", "perl-Tangerine",
+                           multilib=["ppc64le"])
+
+        mmd = self.cg.module.mmd()
+        mmd = self.cg._fill_in_rpms_list(mmd, "x86_64")
+
+        # Only i686 package for dhcp-libs should be added, because perl-Tangerine does not have
+        # multilib set.
+        assert set(mmd.get_rpm_artifacts().get()) == set([
+            "dhcp-libs-12:4.3.5-5.module_2118aef6.x86_64",
+            "dhcp-libs-12:4.3.5-5.module_2118aef6.i686",
+            "perl-Tangerine-12:4.3.5-5.module_2118aef6.x86_64"])