def content_handler(self, path):
"""Serve config.repo and repomd.xml.key."""
if path == self.repository_config_file_name:
repository, publication = self.get_repository_and_publication()
if not publication:
return
base_url = "{}/".format(
urlpath_sanitize(settings.CONTENT_ORIGIN,
settings.CONTENT_PATH_PREFIX, self.base_path))
val = textwrap.dedent(f"""\
[{self.name}]
name={self.name}
enabled=1
baseurl={base_url}
gpgcheck={publication.gpgcheck}
repo_gpgcheck={publication.repo_gpgcheck}
""")
signing_service = repository.metadata_signing_service
if signing_service:
gpgkey_path = urlpath_sanitize(
base_url,
"/repodata/repomd.xml.key",
)
val += f"gpgkey={gpgkey_path}\n"
return Response(body=val)
def parse(self):
"""Parse repository metadata."""
required_metadata_found = set()
for record in self.data.repomd.records:
self.checksum_types[record.type] = getattr(
CHECKSUM_TYPES, record.checksum_type.upper())
record.checksum_type = getattr(CHECKSUM_TYPES,
record.checksum_type.upper())
if record.type in PACKAGE_REPODATA:
required_metadata_found.add(record.type)
self.data.package_repodata_urls[
record.type] = urlpath_sanitize(self.data.remote_url,
record.location_href)
elif record.type in UPDATE_REPODATA:
self.data.updateinfo_url = urlpath_sanitize(
self.data.remote_url, record.location_href)
elif record.type in COMPS_REPODATA:
self.data.comps_url = urlpath_sanitize(self.data.remote_url,
record.location_href)
elif record.type in MODULAR_REPODATA:
self.data.modules_url = urlpath_sanitize(
self.data.remote_url, record.location_href)
elif record.type in SKIP_REPODATA:
pass
else:
self._set_repomd_file(record)
missing_types = set(PACKAGE_REPODATA) - required_metadata_found
if missing_types:
raise FileNotFoundError(
_("XML file(s): {filenames} not found").format(
filenames=", ".join(missing_types)))
async def parse_distribution_tree(self):
"""Parse content from the file treeinfo if present."""
if self.treeinfo:
d_artifacts = [
DeclarativeArtifact(
artifact=Artifact(),
url=urlpath_sanitize(self.data.remote_url,
self.treeinfo["filename"]),
relative_path=".treeinfo",
remote=self.remote,
deferred_download=False,
)
]
for path, checksum in self.treeinfo["download"]["images"].items():
artifact = Artifact(**checksum)
da = DeclarativeArtifact(
artifact=artifact,
url=urlpath_sanitize(self.data.remote_url, path),
relative_path=path,
remote=self.remote,
deferred_download=self.deferred_download,
)
d_artifacts.append(da)
distribution_tree = DistributionTree(
**self.treeinfo["distribution_tree"])
dc = DeclarativeContent(content=distribution_tree,
d_artifacts=d_artifacts)
dc.extra_data = self.treeinfo
await self.put(dc)
def content_handler(self, path):
"""Serve config.repo and public.key."""
if path == self.repository_config_file_name:
base_url = f"{settings.CONTENT_ORIGIN}{settings.CONTENT_PATH_PREFIX}{self.base_path}/"
publication = self.publication.cast()
val = textwrap.dedent(
f"""\
[{self.name}]
enabled=1
baseurl={base_url}
gpgcheck={publication.gpgcheck}
repo_gpgcheck={publication.repo_gpgcheck}
"""
)
repository_pk = self.publication.repository.pk
repository = RpmRepository.objects.get(pk=repository_pk)
signing_service = repository.metadata_signing_service
if signing_service:
gpgkey_path = urlpath_sanitize(
settings.CONTENT_ORIGIN, settings.CONTENT_PATH_PREFIX
)
gpgkey_path = urllib.parse.urljoin(gpgkey_path, self.base_path, True)
gpgkey_path += "/repodata/public.key"
val += f"gpgkey={gpgkey_path}\n"
return Response(body=val)
def get_treeinfo_data(remote, remote_url):
"""
Get Treeinfo data from remote.
"""
treeinfo_serialized = {}
namespaces = [".treeinfo", "treeinfo"]
for namespace in namespaces:
downloader = remote.get_downloader(
url=urlpath_sanitize(remote_url, namespace),
silence_errors_for_response_status_codes={403, 404},
)
try:
result = downloader.fetch()
except FileNotFoundError:
continue
treeinfo = PulpTreeInfo()
treeinfo.load(f=result.path)
treeinfo_parsed = treeinfo.parsed_sections()
sha256 = result.artifact_attributes["sha256"]
treeinfo_serialized = TreeinfoData(treeinfo_parsed).to_dict(
hash=sha256, filename=namespace)
break
return treeinfo_serialized
async def _run(self, extra_data=None):
"""
Download, validate, and compute digests on the `url`. This is a coroutine.
This method provides the same return object type and documented in
:meth:`~pulpcore.plugin.download.BaseDownloader._run`.
"""
if self.sles_auth_token:
auth_param = f"?{self.sles_auth_token}"
url = urlpath_sanitize(self.url) + auth_param
else:
url = self.url
async with self.session.get(url,
proxy=self.proxy,
proxy_auth=self.proxy_auth,
auth=self.auth) as response:
self.raise_for_status(response)
to_return = await self._handle_response(response)
await response.release()
self.response_headers = response.headers
if self._close_session_on_finalize:
self.session.close()
return to_return
async def run(self):
"""Build `DeclarativeContent` from the repodata."""
progress_data = dict(message="Downloading Metadata Files",
code="sync.downloading.metadata")
with ProgressReport(**progress_data) as metadata_pb:
self.data.metadata_pb = metadata_pb
downloader = self.remote.get_downloader(url=urlpath_sanitize(
self.data.remote_url, "repodata/repomd.xml"))
result = await downloader.run()
metadata_pb.increment()
repomd_path = result.path
self.data.repomd = cr.Repomd(repomd_path)
self.repository.last_sync_revision_number = self.data.repomd.revision
self.repository.last_sync_repomd_checksum = get_sha256(repomd_path)
await self.parse_distribution_tree()
await self.parse_repository_metadata()
await self.parse_modules_metadata()
await self.parse_packages_components()
await self.parse_content()
# now send modules down the pipeline since all relations have been set up
for modulemd in self.data.modulemd_list:
await self.put(modulemd)
for dc_group in self.data.dc_groups:
await self.put(dc_group)
async def _parse_packages(self, packages):
progress_data = {
"message": "Parsed Packages",
"code": "sync.parsing.packages",
"total": len(packages),
}
with ProgressReport(**progress_data) as packages_pb:
while True:
try:
(_, pkg) = packages.popitem(last=False)
except KeyError:
break
package = Package(**Package.createrepo_to_dict(pkg))
del pkg
artifact = Artifact(size=package.size_package)
checksum_type = getattr(CHECKSUM_TYPES,
package.checksum_type.upper())
setattr(artifact, checksum_type, package.pkgId)
url = urlpath_sanitize(self.data.remote_url,
package.location_href)
filename = os.path.basename(package.location_href)
da = DeclarativeArtifact(
artifact=artifact,
url=url,
relative_path=filename,
remote=self.remote,
deferred_download=self.deferred_download,
)
dc = DeclarativeContent(content=package, d_artifacts=[da])
dc.extra_data = defaultdict(list)
# find if a package relates to a modulemd
if dc.content.nevra in self.data.nevra_to_module.keys():
dc.content.is_modular = True
for dc_modulemd in self.data.nevra_to_module[
dc.content.nevra]:
dc.extra_data["modulemd_relation"].append(dc_modulemd)
dc_modulemd.extra_data["package_relation"].append(dc)
if dc.content.name in self.data.pkgname_to_groups.keys():
for dc_group in self.data.pkgname_to_groups[
dc.content.name]:
dc.extra_data["group_relations"].append(dc_group)
dc_group.extra_data["related_packages"].append(dc)
packages_pb.increment()
await self.put(dc)
def __init__(
self,
*args,
silence_errors_for_response_status_codes=None,
sles_auth_token=None,
urlencode=True,
**kwargs,
):
"""
Initialize the downloader.
"""
self.sles_auth_token = sles_auth_token
if silence_errors_for_response_status_codes is None:
silence_errors_for_response_status_codes = set()
self.silence_errors_for_response_status_codes = silence_errors_for_response_status_codes
super().__init__(*args, **kwargs)
new_url = self.url
if urlencode:
# Some upstream-repos (eg, Amazon) require url-encoded paths for things like "libc++"
# Let's make them happy.
# We can't urlencode the whole url, because BasicAuth is still A Thing and we would
# break username/passwords in the url.
# So we need to urlencode **only the path** and **nothing else** .
# We can't use _replace() and urlunparse(), because urlunparse() "helpfully" undoes
# the urlencode we just did in the path.
# We can't use urljoin(), because urljoin() "helpfully" treats a number of schemes
# (like, say, uln:) as "can't take relative paths", and throws away everything
# **except** the path-portion
# So, we have a pretty ugly workaround.
parsed = urlparse(self.url)
# two pieces of the URL: pre- and post-path
(before_path, after_path) = self.url.split(parsed.path)
new_path = quote(unquote(parsed.path), safe=":/") # fix the path
new_url = "{}{}{}".format(before_path, new_path,
after_path) # rebuild
if self.sles_auth_token:
auth_param = f"?{self.sles_auth_token}"
self.url = urlpath_sanitize(new_url) + auth_param
else:
self.url = new_url
def _set_repomd_file(self, record):
if "_zck" not in record.type and record.type not in PACKAGE_DB_REPODATA:
file_data = {
record.checksum_type: record.checksum,
"size": record.size
}
da = DeclarativeArtifact(
artifact=Artifact(**file_data),
url=urlpath_sanitize(self.data.remote_url,
record.location_href),
relative_path=record.location_href,
remote=self.remote,
deferred_download=False,
)
repo_metadata_file = RepoMetadataFile(
data_type=record.type,
checksum_type=record.checksum_type,
checksum=record.checksum,
relative_path=record.location_href,
)
dc = DeclarativeContent(content=repo_metadata_file,
d_artifacts=[da])
self.repomd_dcs.append(dc)
def get_repomd_file(remote, url):
"""
Check if repodata exists.
Args:
remote(RpmRemote or UlnRemote): An RpmRemote or UlnRemote to download with.
url(str): A remote repository URL
Returns:
pulpcore.plugin.download.DownloadResult: downloaded repomd.xml
"""
downloader = remote.get_downloader(
url=urlpath_sanitize(url, "repodata/repomd.xml"))
try:
result = downloader.fetch()
except ClientResponseError as exc:
if 404 == exc.status:
return
except FileNotFoundError:
return
return result
def test_urlpath_sanitize(self):
"""Test urljoin-replacement."""
# arbitrary number of args become one single-slash-separated string
a_expected = "a"
ab_expected = "a/b"
abc_expected = "a/b/c"
# a /a a/ /a/
self.assertEqual(a_expected, urlpath_sanitize("a"))
self.assertEqual(a_expected, urlpath_sanitize("/a"))
self.assertEqual(a_expected, urlpath_sanitize("a/"))
self.assertEqual(a_expected, urlpath_sanitize("/a/"))
# a b : a/ b : /a b : a b/ : a /b : a /b/ : a/ /b
self.assertEqual(ab_expected, urlpath_sanitize("a", "b"))
self.assertEqual(ab_expected, urlpath_sanitize("a/", "b"))
self.assertEqual(ab_expected, urlpath_sanitize("/a", "b"))
self.assertEqual(ab_expected, urlpath_sanitize("a", "b/"))
self.assertEqual(ab_expected, urlpath_sanitize("a", "/b"))
self.assertEqual(ab_expected, urlpath_sanitize("a", "/b/"))
self.assertEqual(ab_expected, urlpath_sanitize("a/", "/b"))
self.assertEqual(ab_expected, urlpath_sanitize("a/", "", "/b"))
self.assertEqual(ab_expected, urlpath_sanitize("a/", "/", "/b"))
# a b c : a /b/ /c : /a/ /b/ /c/
self.assertEqual(abc_expected, urlpath_sanitize("a", "b", "c"))
self.assertEqual(abc_expected, urlpath_sanitize("a", "/b/", "/c"))
self.assertEqual(abc_expected, urlpath_sanitize("/a/", "/b/", "/c/"))
def synchronize(remote_pk, repository_pk, mirror, skip_types, optimize):
"""
Sync content from the remote repository.
Create a new version of the repository that is synchronized with the remote.
Args:
remote_pk (str): The remote PK.
repository_pk (str): The repository PK.
mirror (bool): Mirror mode.
skip_types (list): List of content to skip.
optimize(bool): Optimize mode.
Raises:
ValueError: If the remote does not specify a url to sync.
"""
try:
remote = RpmRemote.objects.get(pk=remote_pk)
except ObjectDoesNotExist:
remote = UlnRemote.objects.get(pk=remote_pk)
repository = RpmRepository.objects.get(pk=repository_pk)
if not remote.url:
raise ValueError(
_("A remote must have a url specified to synchronize."))
log.info(
_("Synchronizing: repository={r} remote={p}").format(r=repository.name,
p=remote.name))
deferred_download = remote.policy != Remote.IMMEDIATE # Interpret download policy
with tempfile.TemporaryDirectory("."):
remote_url = fetch_remote_url(remote)
if optimize and is_optimized_sync(repository, remote, remote_url):
return
with tempfile.TemporaryDirectory("."):
treeinfo = get_treeinfo_data(remote, remote_url)
if treeinfo:
treeinfo["repositories"] = {}
for repodata in set(treeinfo["download"]["repodatas"]):
if repodata == DIST_TREE_MAIN_REPO_PATH:
treeinfo["repositories"].update({repodata: None})
continue
name = f"{repodata}-{treeinfo['hash']}"
sub_repo, created = RpmRepository.objects.get_or_create(
name=name, sub_repo=True)
if created:
sub_repo.save()
directory = treeinfo["repo_map"][repodata]
treeinfo["repositories"].update({directory: str(sub_repo.pk)})
path = f"{repodata}/"
new_url = urlpath_sanitize(remote_url, path)
with tempfile.TemporaryDirectory("."):
repodata_exists = get_repomd_file(remote, new_url)
if repodata_exists:
if optimize and is_optimized_sync(sub_repo, remote, new_url):
continue
stage = RpmFirstStage(
remote,
sub_repo,
deferred_download,
skip_types=skip_types,
new_url=new_url,
)
dv = RpmDeclarativeVersion(first_stage=stage,
repository=sub_repo)
dv.create()
sub_repo.last_sync_remote = remote
sub_repo.last_sync_repo_version = sub_repo.latest_version(
).number
sub_repo.save()
first_stage = RpmFirstStage(
remote,
repository,
deferred_download,
skip_types=skip_types,
treeinfo=treeinfo,
new_url=remote_url,
)
dv = RpmDeclarativeVersion(first_stage=first_stage,
repository=repository,
mirror=mirror)
version = dv.create()
if version:
repository.last_sync_remote = remote
repository.last_sync_repo_version = version.number
repository.save()
return version
