Compute checksums in ImageChecksumPhase

The phase goes through all images declared in image manifest, computes
their checksums, stores them in appropriate files and updates the
manifest so that it includes the actual checksums.

The documentation contains details about new configuration options.

The test suite now needs Python's mock package.

Signed-off-by: Lubomír Sedlář <lsedlar@redhat.com>

doc/configuration.rst

@@ -522,6 +522,19 @@ Example
        }),
     ]
 
+
+Media Checksums Settings
+========================
+
+**media_checksums**
+    (*list*) -- list of checksum types to compute, allowed values are ``md5``,
+    ``sha1`` and ``sha256``
+
+**media_checksum_one_file**
+    (*bool*) -- when ``True``, only one ``CHECKSUM`` file will be created per
+    directory; this option requires ``media_checksums`` to only specify one
+    type
+
 Translate Paths Settings
 ========================
 

pungi/phases/createiso.py

@@ -154,9 +154,6 @@ class CreateisoPhase(PhaseBase):
                     isomd5sum_cmd = " ".join([pipes.quote(i) for i in isomd5sum_cmd])
                     cmd["cmd"].append(isomd5sum_cmd)
 
-                    # compute md5sum, sha1sum, sha256sum
-                    cmd["cmd"].extend(iso.get_checksum_cmds(iso_name))
-
                     # create iso manifest
                     cmd["cmd"].append(iso.get_manifest_cmd(iso_name))
 

pungi/phases/image_checksum.py

@@ -1,30 +1,114 @@
 # -*- coding: utf-8 -*-
 
 import os
+from kobo import shortcuts
 
 from .base import PhaseBase
 
 
+MULTIPLE_CHECKSUMS_ERROR = (
+    'Config option "media_checksum_one_file" requires only one checksum'
+    ' to be configured in "media_checksums".'
+)
+
+
 class ImageChecksumPhase(PhaseBase):
-    """Go through images generated in ImageBuild phase and generate their
+    """Go through images specified in image manifest and generate their
-    checksums.
+    checksums. The manifest will be updated with the checksums.
     """
 
     name = 'image_checksum'
 
+    config_options = (
+        {
+            "name": "media_checksums",
+            "expected_types": [list],
+            "optional": True,
+        },
+        {
+            "name": "media_checksum_one_file",
+            "expected_types": [bool],
+            "optional": True,
+        }
+    )
+
+    def __init__(self, compose):
+        super(ImageChecksumPhase, self).__init__(compose)
+        self.checksums = self.compose.conf.get('media_checksums', ['md5', 'sha1', 'sha256'])
+        self.one_file = self.compose.conf.get('media_checksum_one_file', False)
+
+    def validate(self):
+        errors = []
+        try:
+            super(ImageChecksumPhase, self).validate()
+        except ValueError as exc:
+            errors = exc.message.split('\n')
+
+        if self.one_file and len(self.checksums) != 1:
+            errors.append(MULTIPLE_CHECKSUMS_ERROR)
+
+        if errors:
+            raise ValueError('\n'.join(errors))
+
+    def _get_images(self):
+        """Returns a mapping from directories to sets of ``Image``s.
+
+        The paths to dirs are absolute.
+        """
+        top_dir = self.compose.paths.compose.topdir()
+        images = {}
+        for variant in self.compose.im.images:
+            for arch in self.compose.im.images[variant]:
+                for image in self.compose.im.images[variant][arch]:
+                    path = os.path.dirname(os.path.join(top_dir, image.path))
+                    images.setdefault(path, set()).add(image)
+        return images
+
     def run(self):
-        compose = self.compose
+        for path, images in self._get_images().iteritems():
-        # merge checksum files
+            checksums = {}
-        for variant in compose.get_variants(types=["variant", "layered-product"]):
+            for image in images:
-            for arch in variant.arches + ["src"]:
+                filename = os.path.basename(image.path)
-                iso_dir = compose.paths.compose.iso_dir(arch, variant, create_dir=False)
+                full_path = os.path.join(path, filename)
-                if not iso_dir or not os.path.exists(iso_dir):
+                if not os.path.exists(full_path):
                     continue
-                for checksum_type in ("md5", "sha1", "sha256"):
+
-                    checksum_upper = "%sSUM" % checksum_type.upper()
+                digests = shortcuts.compute_file_checksums(full_path, self.checksums)
-                    checksums = sorted([i for i in os.listdir(iso_dir) if i.endswith(".%s" % checksum_upper)])
+                for checksum, digest in digests.iteritems():
-                    fo = open(os.path.join(iso_dir, checksum_upper), "w")
+                    checksums.setdefault(checksum, {})[filename] = digest
-                    for i in checksums:
+                    image.add_checksum(None, checksum, digest)
-                        data = open(os.path.join(iso_dir, i), "r").read()
+                    if not self.one_file:
-                        fo.write(data)
+                        dump_individual(full_path, digest, checksum)
-                    fo.close()
+
+            if not checksums:
+                continue
+
+            if self.one_file:
+                dump_checksums(path, checksums[self.checksums[0]])
+            else:
+                for checksum in self.checksums:
+                    dump_checksums(path, checksums[checksum], '%sSUM' % checksum.upper())
+
+
+def dump_checksums(dir, checksums, filename='CHECKSUM'):
+    """Create file with checksums.
+
+    :param dir: where to put the file
+    :param checksums: mapping from filenames to checksums
+    :param filename: what to call the file
+    """
+    with open(os.path.join(dir, filename), 'w') as f:
+        for file, checksum in checksums.iteritems():
+            f.write('{} *{}\n'.format(checksum, file))
+
+
+def dump_individual(path, checksum, ext):
+    """Create a file with a single checksum, saved into a file with an extra
+    extension.
+
+    :param path: path to the checksummed file
+    :param checksum: the actual digest value
+    :param ext: what extension to add to the checksum file
+    """
+    with open('%s.%sSUM' % (path, ext.upper()), 'w') as f:
+        f.write('{} *{}\n'.format(checksum, os.path.basename(path)))

tests/test_imagechecksumphase.py

@@ -0,0 +1,126 @@
+#!/usr/bin/python
+# -*- coding: utf-8 -*-
+
+
+import unittest
+import mock
+
+import os
+import sys
+import tempfile
+import shutil
+
+sys.path.insert(0, os.path.join(os.path.dirname(__file__), ".."))
+
+from pungi.phases.image_checksum import (ImageChecksumPhase,
+                                         dump_checksums,
+                                         dump_individual)
+
+
+class _DummyCompose(object):
+    def __init__(self, config):
+        self.conf = config
+        self.paths = mock.Mock(
+            compose=mock.Mock(
+                topdir=mock.Mock(return_value='/a/b')
+            )
+        )
+        self.image = mock.Mock(
+            path='Client/i386/iso/image.iso',
+        )
+        self.im = mock.Mock(images={'Client': {'i386': [self.image]}})
+
+
+class TestImageChecksumPhase(unittest.TestCase):
+
+    def test_config_skip_individual_with_multiple_algorithms(self):
+        compose = _DummyCompose({
+            'media_checksums': ['md5', 'sha1'],
+            'media_checksum_one_file': True
+        })
+        phase = ImageChecksumPhase(compose)
+        with self.assertRaises(ValueError) as err:
+            phase.validate()
+            self.assertIn('media_checksum_one_file', err.message)
+
+    @mock.patch('os.path.exists')
+    @mock.patch('kobo.shortcuts.compute_file_checksums')
+    @mock.patch('pungi.phases.image_checksum.dump_checksums')
+    def test_checksum_one_file(self, dump, cc, exists):
+        compose = _DummyCompose({
+            'media_checksums': ['sha256'],
+            'media_checksum_one_file': True,
+        })
+
+        phase = ImageChecksumPhase(compose)
+
+        exists.return_value = True
+        cc.return_value = {'sha256': 'cafebabe'}
+
+        phase.run()
+
+        dump.assert_called_once_with('/a/b/Client/i386/iso', {'image.iso': 'cafebabe'})
+        cc.assert_called_once_with('/a/b/Client/i386/iso/image.iso', ['sha256'])
+        compose.image.add_checksum.assert_called_once_with(None, 'sha256', 'cafebabe')
+
+    @mock.patch('os.path.exists')
+    @mock.patch('kobo.shortcuts.compute_file_checksums')
+    @mock.patch('pungi.phases.image_checksum.dump_checksums')
+    @mock.patch('pungi.phases.image_checksum.dump_individual')
+    def test_checksum_save_individuals(self, indiv_dump, dump, cc, exists):
+        compose = _DummyCompose({
+            'media_checksums': ['md5', 'sha256'],
+        })
+
+        phase = ImageChecksumPhase(compose)
+
+        exists.return_value = True
+        cc.return_value = {'md5': 'cafebabe', 'sha256': 'deadbeef'}
+
+        phase.run()
+
+        indiv_dump.assert_has_calls(
+            [mock.call('/a/b/Client/i386/iso/image.iso', 'cafebabe', 'md5'),
+             mock.call('/a/b/Client/i386/iso/image.iso', 'deadbeef', 'sha256')],
+            any_order=True
+        )
+        dump.assert_has_calls(
+            [mock.call('/a/b/Client/i386/iso', {'image.iso': 'cafebabe'}, 'MD5SUM'),
+             mock.call('/a/b/Client/i386/iso', {'image.iso': 'deadbeef'}, 'SHA256SUM')],
+            any_order=True
+        )
+        cc.assert_called_once_with('/a/b/Client/i386/iso/image.iso', ['md5', 'sha256'])
+        compose.image.add_checksum.assert_has_calls([mock.call(None, 'sha256', 'deadbeef'),
+                                                     mock.call(None, 'md5', 'cafebabe')],
+                                                    any_order=True)
+
+
+class TestChecksums(unittest.TestCase):
+    def setUp(self):
+        _, name = tempfile.mkstemp()
+        self.tmp_dir = tempfile.mkdtemp()
+
+    def tearDown(self):
+        shutil.rmtree(self.tmp_dir)
+
+    def test_dump_checksums(self):
+        dump_checksums(self.tmp_dir, {'file1.iso': 'abcdef', 'file2.iso': 'cafebabe'})
+
+        with open(os.path.join(self.tmp_dir, 'CHECKSUM'), 'r') as f:
+            data = f.read().rstrip().split('\n')
+            expected = [
+                'abcdef *file1.iso',
+                'cafebabe *file2.iso',
+            ]
+            self.assertItemsEqual(expected, data)
+
+    def test_dump_individual(self):
+        base_path = os.path.join(self.tmp_dir, 'file.iso')
+        dump_individual(base_path, 'cafebabe', 'md5')
+
+        with open(base_path + '.MD5SUM', 'r') as f:
+            data = f.read()
+            self.assertEqual('cafebabe *file.iso\n', data)
+
+if __name__ == "__main__":
+    unittest.main()