diff --git a/swh/loader/svn/replay.py b/swh/loader/svn/replay.py index f7c4158..7a2f799 100644 --- a/swh/loader/svn/replay.py +++ b/swh/loader/svn/replay.py @@ -1,917 +1,942 @@ # Copyright (C) 2016-2022 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information """Remote Access client to svn server. """ from __future__ import annotations import codecs from collections import defaultdict from dataclasses import dataclass, field from itertools import chain import logging import os import shutil import tempfile from typing import ( TYPE_CHECKING, Any, BinaryIO, Callable, Dict, List, Optional, Set, Tuple, Union, cast, ) import click from subvertpy import SubversionException, delta, properties from subvertpy.ra import Auth, RemoteAccess, get_username_provider from swh.model import from_disk, hashutil from swh.model.from_disk import DiskBackedContent from swh.model.model import Content, Directory, SkippedContent if TYPE_CHECKING: from swh.loader.svn.svn import SvnRepo -from swh.loader.svn.utils import parse_external_definition, svn_urljoin +from swh.loader.svn.utils import ( + is_recursive_external, + parse_external_definition, + svn_urljoin, +) _eol_style = {"native": b"\n", "CRLF": b"\r\n", "LF": b"\n", "CR": b"\r"} logger = logging.getLogger(__name__) def _normalize_line_endings(lines: bytes, eol_style: str = "native") -> bytes: r"""Normalize line endings to unix (\\n), windows (\\r\\n) or mac (\\r). Args: lines: The lines to normalize eol_style: The line ending format as defined for svn:eol-style property. Acceptable values are 'native', 'CRLF', 'LF' and 'CR' Returns: Lines with endings normalized """ if eol_style in _eol_style: lines = lines.replace(_eol_style["CRLF"], _eol_style["LF"]).replace( _eol_style["CR"], _eol_style["LF"] ) if _eol_style[eol_style] != _eol_style["LF"]: lines = lines.replace(_eol_style["LF"], _eol_style[eol_style]) return lines def apply_txdelta_handler( sbuf: bytes, target_stream: BinaryIO ) -> Callable[[Any, bytes, BinaryIO], None]: """Return a function that can be called repeatedly with txdelta windows. When done, closes the target_stream. Adapted from subvertpy.delta.apply_txdelta_handler to close the stream when done. Args: sbuf: Source buffer target_stream: Target stream to write to. Returns: Function to be called to apply txdelta windows """ def apply_window( window: Any, sbuf: bytes = sbuf, target_stream: BinaryIO = target_stream ): if window is None: target_stream.close() return # Last call patch = delta.apply_txdelta_window(sbuf, window) target_stream.write(patch) return apply_window def read_svn_link(data: bytes) -> Tuple[bytes, bytes]: """Read the svn link's content. Args: data: svn link's raw content Returns: The tuple of (filetype, destination path) """ split_byte = b" " first_line = data.split(b"\n")[0] filetype, *src = first_line.split(split_byte) target = split_byte.join(src) return filetype, target def is_file_an_svnlink_p(fullpath: bytes) -> Tuple[bool, bytes]: """Determine if a filepath is an svnlink or something else. Args: fullpath: Full path to the potential symlink to check Returns: Tuple containing a boolean value to determine if it's indeed a symlink (as per svn) and the link target. """ if os.path.islink(fullpath): return False, b"" with open(fullpath, "rb") as f: filetype, src = read_svn_link(f.read()) return filetype == b"link", src def _ra_codecs_error_handler(e: UnicodeError) -> Tuple[Union[str, bytes], int]: """Subvertpy may fail to decode to utf-8 the user svn properties. As they are not used by the loader, return an empty string instead of the decoded content. Args: e: exception raised during the svn properties decoding. """ return "", cast(UnicodeDecodeError, e).end DEFAULT_FLAG = 0 EXEC_FLAG = 1 NOEXEC_FLAG = 2 SVN_PROPERTY_EOL = "svn:eol-style" @dataclass class FileState: """Persists some file states (eg. end of lines style) across revisions while replaying them.""" eol_style: Optional[str] = None """EOL state check mess""" svn_special_path_non_link_data: Optional[bytes] = None """keep track of non link file content with svn:special property set""" # default value: 0, 1: set the flag, 2: remove the exec flag executable: int = DEFAULT_FLAG """keep track if file is executable when setting svn:executable property""" link: bool = False """keep track if file is a svn link when setting svn:special property""" class FileEditor: """File Editor in charge of updating file on disk and memory objects. """ __slots__ = [ "directory", "path", "fullpath", "executable", "link", "state", "svnrepo", "editor", ] def __init__( self, directory: from_disk.Directory, rootpath: bytes, path: bytes, state: FileState, svnrepo: SvnRepo, ): self.directory = directory self.path = path self.fullpath = os.path.join(rootpath, path) self.state = state self.svnrepo = svnrepo self.editor = svnrepo.swhreplay.editor self.editor.modified_paths.add(path) def change_prop(self, key: str, value: str) -> None: if key == properties.PROP_EXECUTABLE: if value is None: # bit flip off self.state.executable = NOEXEC_FLAG else: self.state.executable = EXEC_FLAG elif key == properties.PROP_SPECIAL: # Possibly a symbolic link. We cannot check further at # that moment though, patch(s) not being applied yet self.state.link = value is not None elif key == SVN_PROPERTY_EOL: # backup end of line style for file self.state.eol_style = value def __make_symlink(self, src: bytes) -> None: """Convert the svnlink to a symlink on disk. This function expects self.fullpath to be a svn link. Args: src: Path to the link's source Return: tuple: The svnlink's data tuple: - type (should be only 'link') - """ os.remove(self.fullpath) os.symlink(src=src, dst=self.fullpath) def __make_svnlink(self) -> bytes: """Convert the symlink to a svnlink on disk. Return: The symlink's svnlink data (``b'type '``) """ # we replace the symlink by a svnlink # to be able to patch the file on future commits src = os.readlink(self.fullpath) os.remove(self.fullpath) sbuf = b"link " + src with open(self.fullpath, "wb") as f: f.write(sbuf) return sbuf def apply_textdelta(self, base_checksum) -> Callable[[Any, bytes, BinaryIO], None]: # if the filepath matches an external, do not apply local patch if self.path in self.editor.external_paths: return lambda *args: None if os.path.lexists(self.fullpath): if os.path.islink(self.fullpath): # svn does not deal with symlink so we transform into # real svn symlink for potential patching in later # commits sbuf = self.__make_svnlink() self.state.link = True else: with open(self.fullpath, "rb") as f: sbuf = f.read() else: sbuf = b"" t = open(self.fullpath, "wb") return apply_txdelta_handler(sbuf, target_stream=t) def close(self) -> None: """When done with the file, this is called. So the file exists and is updated, we can: - adapt accordingly its execution flag if any - compute the objects' checksums - replace the svnlink with a real symlink (for disk computation purposes) """ if self.state.link: # can only check now that the link is a real one # since patch has been applied is_link, src = is_file_an_svnlink_p(self.fullpath) if is_link: self.__make_symlink(src) elif not os.path.isdir(self.fullpath): # not a real link ... # when a file with the svn:special property set is not a svn link, # the svn export operation might extract a truncated version of it # if it is a binary file, so ensure to produce the same file as the # export operation. with open(self.fullpath, "rb") as f: content = f.read() self.svnrepo.client.export( os.path.join(self.svnrepo.remote_url.encode(), self.path), to=self.fullpath, peg_rev=self.editor.revnum, ignore_keywords=True, overwrite=True, ) with open(self.fullpath, "rb") as f: exported_data = f.read() if exported_data != content: # keep track of original file content in order to restore # it if the svn:special property gets unset in another revision self.state.svn_special_path_non_link_data = content elif os.path.islink(self.fullpath): # path was a symbolic link in previous revision but got the property # svn:special unset in current one, revert its content to svn link format self.__make_svnlink() elif self.state.svn_special_path_non_link_data is not None: # path was a non link file with the svn:special property previously set # and got truncated on export, restore its original content with open(self.fullpath, "wb") as f: f.write(self.state.svn_special_path_non_link_data) self.state.svn_special_path_non_link_data = None is_link = os.path.islink(self.fullpath) if not is_link: # if a link, do nothing regarding flag if self.state.executable == EXEC_FLAG: os.chmod(self.fullpath, 0o755) elif self.state.executable == NOEXEC_FLAG: os.chmod(self.fullpath, 0o644) # And now compute file's checksums if self.state.eol_style and not is_link: # ensure to normalize line endings as defined by svn:eol-style # property to get the same file checksum as after an export # or checkout operation with subversion with open(self.fullpath, "rb") as f: data = f.read() data = _normalize_line_endings(data, self.state.eol_style) mode = os.lstat(self.fullpath).st_mode self.directory[self.path] = from_disk.Content.from_bytes( mode=mode, data=data ) else: self.directory[self.path] = from_disk.Content.from_file(path=self.fullpath) @dataclass class DirState: """Persists some directory states (eg. externals) across revisions while replaying them.""" externals: Dict[str, Tuple[str, Optional[int]]] = field(default_factory=dict) class DirEditor: """Directory Editor in charge of updating directory hashes computation. This implementation includes empty folder in the hash computation. """ __slots__ = [ "directory", "rootpath", "path", "file_states", "dir_states", "svnrepo", "editor", "externals", ] def __init__( self, directory: from_disk.Directory, rootpath: bytes, path: bytes, file_states: Dict[bytes, FileState], dir_states: Dict[bytes, DirState], svnrepo: SvnRepo, ): self.directory = directory self.rootpath = rootpath self.path = path # build directory on init os.makedirs(rootpath, exist_ok=True) self.file_states = file_states self.dir_states = dir_states self.svnrepo = svnrepo self.editor = svnrepo.swhreplay.editor self.externals: Dict[str, Tuple[str, Optional[int], bool]] = {} # repository root dir has empty path if path: self.editor.modified_paths.add(path) def remove_child(self, path: bytes) -> None: """Remove a path from the current objects. The path can be resolved as link, file or directory. This function takes also care of removing the link between the child and the parent. Args: path: to remove from the current objects. """ try: entry_removed = self.directory[path] except KeyError: entry_removed = None else: del self.directory[path] fpath = os.path.join(self.rootpath, path) if isinstance(entry_removed, from_disk.Directory): shutil.rmtree(fpath) else: os.remove(fpath) # when deleting a directory ensure to remove any svn property for the # file it contains as they can be added again later in another revision # without the same property set fullpath = os.path.join(self.rootpath, path) for state_path in list(self.file_states): if state_path.startswith(fullpath + b"/"): del self.file_states[state_path] self.editor.modified_paths.discard(path) def open_directory(self, path: str, *args) -> DirEditor: """Updating existing directory. """ return DirEditor( self.directory, rootpath=self.rootpath, path=os.fsencode(path), file_states=self.file_states, dir_states=self.dir_states, svnrepo=self.svnrepo, ) def add_directory(self, path: str, *args) -> DirEditor: """Adding a new directory. """ path_bytes = os.fsencode(path) os.makedirs(os.path.join(self.rootpath, path_bytes), exist_ok=True) if path_bytes and path_bytes not in self.directory: self.dir_states[path_bytes] = DirState() self.directory[path_bytes] = from_disk.Directory() return DirEditor( self.directory, self.rootpath, path_bytes, self.file_states, self.dir_states, svnrepo=self.svnrepo, ) def open_file(self, *args) -> FileEditor: """Updating existing file. """ path = os.fsencode(args[0]) self.directory[path] = from_disk.Content() fullpath = os.path.join(self.rootpath, path) return FileEditor( self.directory, rootpath=self.rootpath, path=path, state=self.file_states[fullpath], svnrepo=self.svnrepo, ) def add_file(self, path: str, *args) -> FileEditor: """Creating a new file. """ path_bytes = os.fsencode(path) self.directory[path_bytes] = from_disk.Content() fullpath = os.path.join(self.rootpath, path_bytes) self.file_states[fullpath] = FileState() return FileEditor( self.directory, self.rootpath, path_bytes, state=self.file_states[fullpath], svnrepo=self.svnrepo, ) def change_prop(self, key: str, value: str) -> None: """Change property callback on directory. """ if key == properties.PROP_EXTERNALS: logger.debug( "Setting '%s' property with value '%s' on path %s", key, value, self.path, ) self.externals = {} if value is not None: # externals are set on that directory path, parse and store them # for later processing in the close method for external in value.split("\n"): external = external.rstrip("\r") # skip empty line or comment if not external or external.startswith("#"): continue ( path, external_url, revision, relative_url, ) = parse_external_definition( external, os.fsdecode(self.path), self.svnrepo.origin_url ) self.externals[path] = (external_url, revision, relative_url) if not self.externals: # externals might have been unset on that directory path, # remove associated paths from the reconstructed filesystem externals = self.dir_states[self.path].externals for path in externals.keys(): self.remove_external_path(os.fsencode(path)) self.dir_states[self.path].externals = {} def delete_entry(self, path: str, revision: int) -> None: """Remove a path. """ path_bytes = os.fsencode(path) if path_bytes not in self.editor.external_paths: fullpath = os.path.join(self.rootpath, path_bytes) self.file_states.pop(fullpath, None) self.remove_child(path_bytes) def close(self): """Function called when we finish processing a repository. SVN external definitions are processed by it. """ prev_externals = self.dir_states[self.path].externals if self.externals: # externals definition list might have changed in the current replayed # revision, we need to determine if some were removed and delete the # associated paths externals = self.externals prev_externals_set = { (path, url, rev) for path, (url, rev, _) in prev_externals.items() } externals_set = { (path, url, rev) for path, (url, rev, _) in externals.items() } old_externals = prev_externals_set - externals_set for path, _, _ in old_externals: self.remove_external_path(os.fsencode(path)) else: # some external paths might have been removed in the current replayed # revision by a delete operation on an overlapping versioned path so we # need to restore them externals = prev_externals # For each external, try to export it in reconstructed filesystem for path, (external_url, revision, relative_url) in externals.items(): external = (external_url, revision) dest_path = os.fsencode(path) dest_fullpath = os.path.join(self.path, dest_path) if ( path in prev_externals and prev_externals[path] == external and dest_fullpath in self.directory ): # external already exported, nothing to do continue + if is_recursive_external( + self.svnrepo.origin_url, os.fsdecode(self.path), path, external_url + ): + # recursive external, skip it + continue + if external not in self.editor.externals_cache: try: # try to export external in a temporary path, destination path could # be versioned and must be overridden only if the external URL is # still valid temp_dir = os.fsencode( tempfile.mkdtemp(dir=self.editor.externals_cache_dir) ) temp_path = os.path.join(temp_dir, dest_path) os.makedirs(b"/".join(temp_path.split(b"/")[:-1]), exist_ok=True) if external_url not in self.editor.dead_externals: logger.debug( "Exporting external %s to path %s", external_url, path ) self.svnrepo.client.export( external_url.rstrip("/"), to=temp_path, peg_rev=revision, ignore_keywords=True, ) self.editor.externals_cache[external] = temp_path self.editor.valid_externals[dest_fullpath] = ( external_url, relative_url, ) except SubversionException as se: # external no longer available (404) logger.debug(se) self.editor.dead_externals.add(external_url) else: temp_path = self.editor.externals_cache[external] # subversion export will always create the subdirectories of the external # path regardless the validity of the remote URL dest_path_split = dest_path.split(b"/") current_path = self.path self.add_directory(current_path) for subpath in dest_path_split[:-1]: current_path = os.path.join(current_path, subpath) self.add_directory(current_path) if os.path.exists(temp_path): # external successfully exported # remove previous path in from_disk model self.remove_child(dest_fullpath) # copy exported path to reconstructed filesystem fullpath = os.path.join(self.rootpath, dest_fullpath) self.editor.external_paths.add(dest_fullpath) self.editor.modified_paths.add(dest_fullpath) # update from_disk model and store external paths self.editor.external_paths.add(dest_fullpath) if os.path.isfile(temp_path): shutil.copy(temp_path, fullpath) self.directory[dest_fullpath] = from_disk.Content.from_file( path=fullpath ) else: shutil.copytree(temp_path, fullpath, symlinks=True) self.directory[dest_fullpath] = from_disk.Directory.from_disk( path=fullpath ) external_paths = set() for root, dirs, files in os.walk(fullpath): external_paths.update( [ os.path.join(root.replace(self.rootpath + b"/", b""), p) for p in chain(dirs, files) ] ) self.editor.external_paths.update(external_paths) self.editor.modified_paths.update(external_paths) # ensure hash update for the directory with externals set self.directory[self.path].update_hash(force=True) # backup externals in directory state if self.externals: self.dir_states[self.path].externals = self.externals self.svnrepo.has_relative_externals = any( [relative_url for (_, relative_url) in self.editor.valid_externals.values()] ) + self.svnrepo.has_recursive_externals = any( + is_recursive_external( + self.svnrepo.origin_url, os.fsdecode(path), external_path, external_url + ) + for path, dir_state in self.dir_states.items() + for external_path, (external_url, _, _) in dir_state.externals.items() + ) + if self.svnrepo.has_recursive_externals: + # If the repository has recursive externals, we stop processing externals + # and remove those already exported, + # We will then ignore externals when exporting the revision to check for + # divergence with the reconstructed filesystem + for external_path in list(self.editor.external_paths): + self.remove_external_path(external_path) + def remove_external_path(self, external_path: bytes) -> None: """Remove a previously exported SVN external path from the reconstruted filesystem. """ fullpath = os.path.join(self.path, external_path) self.remove_child(fullpath) self.editor.external_paths.discard(fullpath) self.editor.valid_externals.pop(fullpath, None) for path in list(self.editor.external_paths): if path.startswith(fullpath + b"/"): self.editor.external_paths.remove(path) subpath_split = external_path.split(b"/")[:-1] for i in reversed(range(1, len(subpath_split) + 1)): # delete external sub-directory only if it is not versioned subpath = os.path.join(self.path, b"/".join(subpath_split[0:i])) try: self.svnrepo.client.info( svn_urljoin(self.svnrepo.remote_url, os.fsdecode(subpath)), peg_revision=self.editor.revnum, ) except SubversionException: self.remove_child(subpath) else: break try: # externals can overlap with versioned files so we must restore # them after removing the path above dest_path = os.path.join(self.rootpath, fullpath) self.svnrepo.client.export( svn_urljoin(self.svnrepo.remote_url, os.fsdecode(fullpath)), to=dest_path, peg_rev=self.editor.revnum, ignore_keywords=True, ) if os.path.isfile(dest_path): self.directory[fullpath] = from_disk.Content.from_file(path=dest_path) else: self.directory[fullpath] = from_disk.Directory.from_disk(path=dest_path) except SubversionException: pass class Editor: """Editor in charge of replaying svn events and computing objects along. This implementation accounts for empty folder during hash computations. """ def __init__( self, rootpath: bytes, directory: from_disk.Directory, svnrepo: SvnRepo, temp_dir: str, ): self.rootpath = rootpath self.directory = directory self.file_states: Dict[bytes, FileState] = defaultdict(FileState) self.dir_states: Dict[bytes, DirState] = defaultdict(DirState) self.external_paths: Set[bytes] = set() self.valid_externals: Dict[bytes, Tuple[str, bool]] = {} self.dead_externals: Set[str] = set() self.externals_cache_dir = tempfile.mkdtemp(dir=temp_dir) self.externals_cache: Dict[Tuple[str, Optional[int]], str] = {} self.svnrepo = svnrepo self.revnum = None # to store the set of paths added or modified when replaying a revision self.modified_paths: Set[bytes] = set() def set_target_revision(self, revnum) -> None: self.revnum = revnum def abort(self) -> None: pass def close(self) -> None: pass def open_root(self, base_revnum: int) -> DirEditor: # a new revision is being replayed so clear the modified_paths set self.modified_paths.clear() return DirEditor( self.directory, rootpath=self.rootpath, path=b"", file_states=self.file_states, dir_states=self.dir_states, svnrepo=self.svnrepo, ) class Replay: """Replay class. """ def __init__( self, conn: RemoteAccess, rootpath: bytes, svnrepo: SvnRepo, temp_dir: str, directory: Optional[from_disk.Directory] = None, ): self.conn = conn self.rootpath = rootpath if directory is None: directory = from_disk.Directory() self.directory = directory self.editor = Editor( rootpath=rootpath, directory=directory, svnrepo=svnrepo, temp_dir=temp_dir ) def replay(self, rev: int) -> from_disk.Directory: """Replay svn actions between rev and rev+1. This method updates in place the self.editor.directory, as well as the filesystem. Returns: The updated root directory """ codecs.register_error("strict", _ra_codecs_error_handler) self.conn.replay(rev, rev + 1, self.editor) codecs.register_error("strict", codecs.strict_errors) return self.editor.directory def compute_objects( self, rev: int ) -> Tuple[List[Content], List[SkippedContent], List[Directory]]: """Compute objects added or modified at revisions rev. Expects the state to be at previous revision's objects. Args: rev: The revision to start the replay from. Returns: The updated objects between rev and rev+1. Beware that this mutates the filesystem at rootpath accordingly. """ self.replay(rev) contents: List[Content] = [] skipped_contents: List[SkippedContent] = [] directories: List[Directory] = [] directories.append(self.editor.directory.to_model()) for path in self.editor.modified_paths: obj = self.directory[path].to_model() obj_type = obj.object_type if obj_type in (Content.object_type, DiskBackedContent.object_type): contents.append(obj.with_data()) elif obj_type == SkippedContent.object_type: skipped_contents.append(obj) elif obj_type == Directory.object_type: directories.append(obj) return contents, skipped_contents, directories @click.command() @click.option("--local-url", default="/tmp", help="local svn working copy") @click.option( "--svn-url", default="file:///home/storage/svn/repos/pkg-fox", help="svn repository's url.", ) @click.option( "--revision-start", default=1, type=click.INT, help="svn repository's starting revision.", ) @click.option( "--revision-end", default=-1, type=click.INT, help="svn repository's ending revision.", ) @click.option( "--debug/--nodebug", default=True, help="Indicates if the server should run in debug mode.", ) @click.option( "--cleanup/--nocleanup", default=True, help="Indicates whether to cleanup disk when done or not.", ) def main(local_url, svn_url, revision_start, revision_end, debug, cleanup): """Script to present how to use Replay class. """ conn = RemoteAccess(svn_url.encode("utf-8"), auth=Auth([get_username_provider()])) os.makedirs(local_url, exist_ok=True) rootpath = tempfile.mkdtemp( prefix=local_url, suffix="-" + os.path.basename(svn_url) ) rootpath = os.fsencode(rootpath) # Do not go beyond the repository's latest revision revision_end_max = conn.get_latest_revnum() if revision_end == -1: revision_end = revision_end_max revision_end = min(revision_end, revision_end_max) try: replay = Replay(conn, rootpath) for rev in range(revision_start, revision_end + 1): contents, skipped_contents, directories = replay.compute_objects(rev) print( "r%s %s (%s new contents, %s new directories)" % ( rev, hashutil.hash_to_hex(replay.directory.hash), len(contents) + len(skipped_contents), len(directories), ) ) if debug: print("%s" % rootpath.decode("utf-8")) finally: if cleanup: if os.path.exists(rootpath): shutil.rmtree(rootpath) if __name__ == "__main__": main() diff --git a/swh/loader/svn/svn.py b/swh/loader/svn/svn.py index fdfc5da..d10f6bd 100644 --- a/swh/loader/svn/svn.py +++ b/swh/loader/svn/svn.py @@ -1,344 +1,365 @@ # Copyright (C) 2015-2022 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information """SVN client in charge of iterating over svn logs and yield commit representations including the hash tree/content computations per svn commit. """ import logging import os import shutil import tempfile from typing import Dict, Iterator, List, Optional, Tuple, Union from subvertpy import client, properties from subvertpy.ra import Auth, RemoteAccess, get_username_provider from swh.model.from_disk import Directory as DirectoryFromDisk from swh.model.model import ( Content, Directory, Person, SkippedContent, TimestampWithTimezone, ) from . import converters, replay -from .utils import parse_external_definition +from .utils import is_recursive_external, parse_external_definition # When log message contains empty data DEFAULT_AUTHOR_MESSAGE = "" logger = logging.getLogger(__name__) class SvnRepo: """Svn repository representation. Args: remote_url: Remove svn repository url origin_url: Associated origin identifier local_dirname: Path to write intermediary svn action results """ def __init__( self, remote_url: str, origin_url: str, local_dirname: str, max_content_length: int, ): self.remote_url = remote_url.rstrip("/") self.origin_url = origin_url auth = Auth([get_username_provider()]) # one connection for log iteration self.conn_log = RemoteAccess(self.remote_url, auth=auth) # another for replay self.conn = RemoteAccess(self.remote_url, auth=auth) # one client for update operation self.client = client.Client(auth=auth) self.local_dirname = local_dirname local_name = os.path.basename(self.remote_url) self.local_url = os.path.join(self.local_dirname, local_name).encode("utf-8") self.uuid = self.conn.get_uuid().encode("utf-8") self.swhreplay = replay.Replay( conn=self.conn, rootpath=self.local_url, svnrepo=self, temp_dir=local_dirname, ) self.max_content_length = max_content_length self.has_relative_externals = False + self.has_recursive_externals = False self.replay_started = False def __str__(self): return str( { "swh-origin": self.origin_url, "remote_url": self.remote_url, "local_url": self.local_url, "uuid": self.uuid, } ) def head_revision(self) -> int: """Retrieve current head revision. """ return self.conn.get_latest_revnum() def initial_revision(self) -> int: """Retrieve the initial revision from which the remote url appeared. """ return 1 def convert_commit_message(self, msg: Union[str, bytes]) -> bytes: """Simply encode the commit message. Args: msg: the commit message to convert. Returns: The transformed message as bytes. """ if isinstance(msg, bytes): return msg return msg.encode("utf-8") def convert_commit_date(self, date: bytes) -> TimestampWithTimezone: """Convert the message commit date into a timestamp in swh format. The precision is kept. Args: date: the commit date to convert. Returns: The transformed date. """ return converters.svn_date_to_swh_date(date) def convert_commit_author(self, author: Optional[bytes]) -> Person: """Convert the commit author into an swh person. Args: author: the commit author to convert. Returns: Person as model object """ return converters.svn_author_to_swh_person(author) def __to_entry(self, log_entry: Tuple) -> Dict: changed_paths, rev, revprops, has_children = log_entry author_date = self.convert_commit_date( revprops.get(properties.PROP_REVISION_DATE) ) author = self.convert_commit_author( revprops.get(properties.PROP_REVISION_AUTHOR) ) message = self.convert_commit_message( revprops.get(properties.PROP_REVISION_LOG, DEFAULT_AUTHOR_MESSAGE) ) return { "rev": rev, "author_date": author_date, "author_name": author, "message": message, } def logs(self, revision_start: int, revision_end: int) -> Iterator[Dict]: """Stream svn logs between revision_start and revision_end by chunks of block_size logs. Yields revision and associated revision information between the revision start and revision_end. Args: revision_start: the svn revision starting bound revision_end: the svn revision ending bound Yields: tuple: tuple of revisions and logs: - revisions: list of revisions in order - logs: Dictionary with key revision number and value the log entry. The log entry is a dictionary with the following keys: - author_date: date of the commit - author_name: name of the author - message: commit message """ for log_entry in self.conn_log.iter_log( paths=None, start=revision_start, end=revision_end, discover_changed_paths=False, ): yield self.__to_entry(log_entry) def export_temporary(self, revision: int) -> Tuple[str, bytes]: """Export the repository to a given revision in a temporary location. This is up to the caller of this function to clean up the temporary location when done (cf. self.clean_fs method) Args: revision: Revision to export at Returns: The tuple local_dirname the temporary location root folder, local_url where the repository was exported. """ local_dirname = tempfile.mkdtemp( dir=self.local_dirname, prefix=f"check-revision-{revision}." ) local_name = os.path.basename(self.remote_url) local_url = os.path.join(local_dirname, local_name) url = self.remote_url # if some paths have external URLs relative to the repository URL but targeting # paths oustide it, we need to export from the origin URL as the remote URL can # target a dump mounted on the local filesystem if self.replay_started and self.has_relative_externals: # externals detected while replaying revisions url = self.origin_url - elif not self.replay_started and self.remote_url.startswith("file://"): + elif not self.replay_started: # revisions replay has not started, we need to check if svn:externals # properties are set from a checkout of the revision and if some - # external URLs are relative to pick the right export URL + # external URLs are relative to pick the right export URL, + # recursive externals are also checked with tempfile.TemporaryDirectory( dir=self.local_dirname, prefix=f"checkout-revision-{revision}." ) as co_dirname: self.client.checkout( self.remote_url, co_dirname, revision, ignore_externals=True ) # get all svn:externals properties recursively externals = self.client.propget( "svn:externals", co_dirname, None, revision, True ) self.has_relative_externals = False + self.has_recursive_externals = False for path, external_defs in externals.items(): - if self.has_relative_externals: + if self.has_relative_externals or self.has_recursive_externals: break + path = path.replace(self.remote_url.rstrip("/") + "/", "") for external_def in os.fsdecode(external_defs).split("\n"): # skip empty line or comment if not external_def or external_def.startswith("#"): continue - _, _, _, relative_url = parse_external_definition( + ( + external_path, + external_url, + _, + relative_url, + ) = parse_external_definition( external_def.rstrip("\r"), path, self.origin_url ) + + if is_recursive_external( + self.origin_url, path, external_path, external_url, + ): + self.has_recursive_externals = True + url = self.remote_url + break + if relative_url: self.has_relative_externals = True url = self.origin_url break self.client.export( - url.rstrip("/"), to=local_url, rev=revision, ignore_keywords=True, + url.rstrip("/"), + to=local_url, + rev=revision, + ignore_keywords=True, + ignore_externals=self.has_recursive_externals, ) return local_dirname, os.fsencode(local_url) def swh_hash_data_per_revision( self, start_revision: int, end_revision: int ) -> Iterator[ Tuple[ int, Optional[int], Dict, Tuple[List[Content], List[SkippedContent], List[Directory]], DirectoryFromDisk, ], ]: """Compute swh hash data per each revision between start_revision and end_revision. Args: start_revision: starting revision end_revision: ending revision Yields: Tuple (rev, nextrev, commit, objects_per_path): - rev: current revision - nextrev: next revision or None if we reached end_revision. - commit: commit data (author, date, message) for such revision - objects_per_path: Tuple of list of objects between start_revision and end_revision - complete Directory representation """ # even in incremental loading mode, we need to replay the whole set of # path modifications from first revision to restore possible file states induced # by setting svn properties on those files (end of line style for instance) self.replay_started = True first_revision = 1 if start_revision else 0 # handle empty repository edge case for commit in self.logs(first_revision, end_revision): rev = commit["rev"] objects = self.swhreplay.compute_objects(rev) if rev == end_revision: nextrev = None else: nextrev = rev + 1 if rev >= start_revision: # start yielding new data to archive once we reached the revision to # resume the loading from yield rev, nextrev, commit, objects, self.swhreplay.directory def swh_hash_data_at_revision( self, revision: int ) -> Tuple[Dict, DirectoryFromDisk]: """Compute the information at a given svn revision. This is expected to be used for checks only. Yields: The tuple (commit dictionary, targeted directory object). """ # Update disk representation of the repository at revision id local_dirname, local_url = self.export_temporary(revision) # Compute the current hashes on disk directory = DirectoryFromDisk.from_disk( path=local_url, max_content_length=self.max_content_length ) # Retrieve the commit information for revision commit = list(self.logs(revision, revision))[0] # Clean export directory self.clean_fs(local_dirname) return commit, directory def clean_fs(self, local_dirname: Optional[str] = None) -> None: """Clean up the local working copy. Args: local_dirname: Path to remove recursively if provided. Otherwise, remove the temporary upper root tree used for svn repository loading. """ dirname = local_dirname or self.local_dirname if os.path.exists(dirname): logger.debug("cleanup %s", dirname) shutil.rmtree(dirname) diff --git a/swh/loader/svn/tests/test_loader.py b/swh/loader/svn/tests/test_loader.py index 87e18d3..be37832 100644 --- a/swh/loader/svn/tests/test_loader.py +++ b/swh/loader/svn/tests/test_loader.py @@ -1,2736 +1,2832 @@ # Copyright (C) 2016-2022 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information from enum import Enum from io import BytesIO import os import shutil import subprocess from typing import Any, Dict, List import pytest from subvertpy import SubversionException, delta, repos from subvertpy.ra import Auth, RemoteAccess, get_username_provider from typing_extensions import TypedDict from swh.loader.svn.loader import ( SvnLoader, SvnLoaderFromDumpArchive, SvnLoaderFromRemoteDump, ) from swh.loader.svn.svn import SvnRepo from swh.loader.svn.utils import init_svn_repo_from_dump, svn_urljoin from swh.loader.tests import ( assert_last_visit_matches, check_snapshot, get_stats, prepare_repository_from_archive, ) from swh.model.from_disk import DentryPerms from swh.model.hashutil import hash_to_bytes from swh.model.model import Snapshot, SnapshotBranch, TargetType GOURMET_SNAPSHOT = Snapshot( id=hash_to_bytes("889cacc2731e3312abfb2b1a0c18ade82a949e07"), branches={ b"HEAD": SnapshotBranch( target=hash_to_bytes("4876cb10aec6f708f7466dddf547567b65f6c39c"), target_type=TargetType.REVISION, ) }, ) GOURMET_UPDATES_SNAPSHOT = Snapshot( id=hash_to_bytes("11086d15317014e43d2438b7ffc712c44f1b8afe"), branches={ b"HEAD": SnapshotBranch( target=hash_to_bytes("171dc35522bfd17dda4e90a542a0377fb2fc707a"), target_type=TargetType.REVISION, ) }, ) def test_loader_svn_not_found_no_mock(swh_storage, tmp_path): """Given an unknown repository, the loader visit ends up in status not_found""" repo_url = "unknown-repository" loader = SvnLoader(swh_storage, repo_url, temp_directory=tmp_path) assert loader.load() == {"status": "uneventful"} assert_last_visit_matches( swh_storage, repo_url, status="not_found", type="svn", ) @pytest.mark.parametrize( "exception_msg", ["Unable to connect to a repository at URL", "Unknown URL type",] ) def test_loader_svn_not_found(swh_storage, tmp_path, exception_msg, mocker): """Given unknown repository issues, the loader visit ends up in status not_found""" mock = mocker.patch("swh.loader.svn.loader.SvnRepo") mock.side_effect = SubversionException(exception_msg, 0) unknown_repo_url = "unknown-repository" loader = SvnLoader(swh_storage, unknown_repo_url, temp_directory=tmp_path) assert loader.load() == {"status": "uneventful"} assert_last_visit_matches( swh_storage, unknown_repo_url, status="not_found", type="svn", ) @pytest.mark.parametrize( "exception", [ SubversionException("Irrelevant message, considered a failure", 10), SubversionException("Present but fails to read, considered a failure", 20), ValueError("considered a failure"), ], ) def test_loader_svn_failures(swh_storage, tmp_path, exception, mocker): """Given any errors raised, the loader visit ends up in status failed""" mock = mocker.patch("swh.loader.svn.loader.SvnRepo") mock.side_effect = exception existing_repo_url = "existing-repo-url" loader = SvnLoader(swh_storage, existing_repo_url, temp_directory=tmp_path) assert loader.load() == {"status": "failed"} assert_last_visit_matches( swh_storage, existing_repo_url, status="failed", type="svn", ) def test_loader_svnrdump_not_found(swh_storage, tmp_path, mocker): """Loading from remote dump which does not exist should end up as not_found visit""" unknown_repo_url = "file:///tmp/svn.code.sf.net/p/white-rats-studios/svn" loader = SvnLoaderFromRemoteDump( swh_storage, unknown_repo_url, temp_directory=tmp_path ) assert loader.load() == {"status": "uneventful"} assert_last_visit_matches( swh_storage, unknown_repo_url, status="not_found", type="svn", ) def test_loader_svnrdump_no_such_revision(swh_storage, tmp_path, datadir): """Visit multiple times an origin with the remote loader should not raise. It used to fail the ingestion on the second visit with a "No such revision x, 160006" message. """ archive_ori_dump = os.path.join(datadir, "penguinsdbtools2018.dump.gz") archive_dump_dir = os.path.join(tmp_path, "dump") os.mkdir(archive_dump_dir) archive_dump = os.path.join(archive_dump_dir, "penguinsdbtools2018.dump.gz") # loader now drops the dump as soon as it's mounted so we need to make a copy first shutil.copyfile(archive_ori_dump, archive_dump) loading_path = str(tmp_path / "loading") os.mkdir(loading_path) # Prepare the dump as a local svn repository for test purposes temp_dir, repo_path = init_svn_repo_from_dump( archive_dump, root_dir=tmp_path, gzip=True ) repo_url = f"file://{repo_path}" loader = SvnLoaderFromRemoteDump(swh_storage, repo_url, temp_directory=loading_path) assert loader.load() == {"status": "eventful"} actual_visit = assert_last_visit_matches( swh_storage, repo_url, status="full", type="svn", ) check_snapshot(loader.snapshot, loader.storage) loader2 = SvnLoaderFromRemoteDump( swh_storage, repo_url, temp_directory=loading_path ) # Visiting a second time the same repository should be uneventful... assert loader2.load() == {"status": "uneventful"} actual_visit2 = assert_last_visit_matches( swh_storage, repo_url, status="full", type="svn", ) assert actual_visit.snapshot is not None # ... with the same snapshot as the first visit assert actual_visit2.snapshot == actual_visit.snapshot def test_loader_svn_new_visit(swh_storage, datadir, tmp_path): """Eventful visit should yield 1 snapshot""" archive_name = "pkg-gourmet" archive_path = os.path.join(datadir, f"{archive_name}.tgz") repo_url = prepare_repository_from_archive(archive_path, archive_name, tmp_path) loader = SvnLoader(swh_storage, repo_url, temp_directory=tmp_path) assert loader.load() == {"status": "eventful"} assert_last_visit_matches( loader.storage, repo_url, status="full", type="svn", snapshot=GOURMET_SNAPSHOT.id, ) check_snapshot(loader.snapshot, loader.storage) stats = get_stats(loader.storage) assert stats == { "content": 19, "directory": 17, "origin": 1, "origin_visit": 1, "release": 0, "revision": 6, "skipped_content": 0, "snapshot": 1, } check_snapshot(GOURMET_SNAPSHOT, loader.storage) def test_loader_svn_2_visits_no_change(swh_storage, datadir, tmp_path): """Visit multiple times a repository with no change should yield the same snapshot """ archive_name = "pkg-gourmet" archive_path = os.path.join(datadir, f"{archive_name}.tgz") repo_url = prepare_repository_from_archive(archive_path, archive_name, tmp_path) loader = SvnLoader(swh_storage, repo_url, temp_directory=tmp_path) assert loader.load() == {"status": "eventful"} visit_status1 = assert_last_visit_matches( loader.storage, repo_url, status="full", type="svn", snapshot=GOURMET_SNAPSHOT.id, ) check_snapshot(loader.snapshot, loader.storage) assert loader.load() == {"status": "uneventful"} visit_status2 = assert_last_visit_matches( loader.storage, repo_url, status="full", type="svn", snapshot=GOURMET_SNAPSHOT.id, ) assert visit_status1.date < visit_status2.date assert visit_status1.snapshot == visit_status2.snapshot stats = get_stats(loader.storage) assert stats["origin_visit"] == 1 + 1 # computed twice the same snapshot assert stats["snapshot"] == 1 # even starting from previous revision... start_revision = loader.storage.revision_get( [hash_to_bytes("95edacc8848369d6fb1608e887d6d2474fd5224f")] )[0] assert start_revision is not None loader = SvnLoader(swh_storage, repo_url, temp_directory=tmp_path) assert loader.load() == {"status": "uneventful"} stats = get_stats(loader.storage) assert stats["origin_visit"] == 2 + 1 # ... with no change in repository, this yields the same snapshot assert stats["snapshot"] == 1 assert_last_visit_matches( loader.storage, repo_url, status="full", type="svn", snapshot=GOURMET_SNAPSHOT.id, ) def test_loader_tampered_repository(swh_storage, datadir, tmp_path): """In this scenario, the dump has been tampered with to modify the commit log [1]. This results in a hash divergence which is detected at startup after a new run for the same origin. In effect, this will perform a complete reloading of the repository. [1] Tampering with revision 6 log message following: ``` tar xvf pkg-gourmet.tgz # initial repository ingested cd pkg-gourmet/ echo "Tampering with commit log message for fun and profit" > log.txt svnadmin setlog . -r 6 log.txt --bypass-hooks tar cvf pkg-gourmet-tampered-rev6-log.tgz pkg-gourmet/ ``` """ archive_name = "pkg-gourmet" archive_path = os.path.join(datadir, f"{archive_name}.tgz") repo_url = prepare_repository_from_archive(archive_path, archive_name, tmp_path) loader = SvnLoader(swh_storage, repo_url, temp_directory=tmp_path) assert loader.load() == {"status": "eventful"} check_snapshot(GOURMET_SNAPSHOT, loader.storage) assert_last_visit_matches( loader.storage, repo_url, status="full", type="svn", snapshot=GOURMET_SNAPSHOT.id, ) check_snapshot(loader.snapshot, loader.storage) archive_path2 = os.path.join(datadir, "pkg-gourmet-tampered-rev6-log.tgz") repo_tampered_url = prepare_repository_from_archive( archive_path2, archive_name, tmp_path ) loader2 = SvnLoader( swh_storage, repo_tampered_url, origin_url=repo_url, temp_directory=tmp_path ) assert loader2.load() == {"status": "eventful"} assert_last_visit_matches( loader2.storage, repo_url, status="full", type="svn", snapshot=hash_to_bytes("5aa61959e788e281fd6e187053d0f46c68e8d8bb"), ) check_snapshot(loader.snapshot, loader.storage) stats = get_stats(loader.storage) assert stats["origin"] == 1 assert stats["origin_visit"] == 2 assert stats["snapshot"] == 2 def test_loader_svn_visit_with_changes(swh_storage, datadir, tmp_path): """In this scenario, the repository has been updated with new changes. The loading visit should result in new objects stored and 1 new snapshot. """ archive_name = "pkg-gourmet" archive_path = os.path.join(datadir, f"{archive_name}.tgz") repo_initial_url = prepare_repository_from_archive( archive_path, archive_name, tmp_path ) # repo_initial_url becomes the origin_url we want to visit some more below loader = SvnLoader(swh_storage, repo_initial_url, temp_directory=tmp_path) assert loader.load() == {"status": "eventful"} visit_status1 = assert_last_visit_matches( loader.storage, repo_initial_url, status="full", type="svn", snapshot=GOURMET_SNAPSHOT.id, ) check_snapshot(GOURMET_SNAPSHOT, loader.storage) archive_path = os.path.join(datadir, "pkg-gourmet-with-updates.tgz") repo_updated_url = prepare_repository_from_archive( archive_path, "pkg-gourmet", tmp_path ) loader = SvnLoader( swh_storage, repo_updated_url, origin_url=repo_initial_url, temp_directory=tmp_path, ) assert loader.load() == {"status": "eventful"} visit_status2 = assert_last_visit_matches( loader.storage, repo_updated_url, status="full", type="svn", snapshot=GOURMET_UPDATES_SNAPSHOT.id, ) assert visit_status1.date < visit_status2.date assert visit_status1.snapshot != visit_status2.snapshot stats = get_stats(loader.storage) assert stats == { "content": 22, "directory": 28, "origin": 1, "origin_visit": 2, "release": 0, "revision": 11, "skipped_content": 0, "snapshot": 2, } check_snapshot(GOURMET_UPDATES_SNAPSHOT, loader.storage) # Let's start the ingestion from the start, this should yield the same result loader = SvnLoader( swh_storage, repo_updated_url, origin_url=repo_initial_url, incremental=False, temp_directory=tmp_path, ) assert loader.load() == {"status": "eventful"} visit_status3 = assert_last_visit_matches( loader.storage, repo_updated_url, status="full", type="svn", snapshot=GOURMET_UPDATES_SNAPSHOT.id, ) assert visit_status2.date < visit_status3.date assert visit_status3.snapshot == visit_status2.snapshot check_snapshot(GOURMET_UPDATES_SNAPSHOT, loader.storage) stats = get_stats(loader.storage) assert stats["origin"] == 1 # always the same visit assert stats["origin_visit"] == 2 + 1 # 1 more visit assert stats["snapshot"] == 2 # no new snapshot def test_loader_svn_visit_start_from_revision(swh_storage, datadir, tmp_path): """Starting from existing revision, next visit on changed repo should yield 1 new snapshot. """ archive_name = "pkg-gourmet" archive_path = os.path.join(datadir, f"{archive_name}.tgz") repo_initial_url = prepare_repository_from_archive( archive_path, archive_name, tmp_path ) # repo_initial_url becomes the origin_url we want to visit some more below loader = SvnLoader(swh_storage, repo_initial_url, temp_directory=tmp_path) assert loader.load() == {"status": "eventful"} visit_status1 = assert_last_visit_matches( loader.storage, repo_initial_url, status="full", type="svn", snapshot=GOURMET_SNAPSHOT.id, ) check_snapshot(GOURMET_SNAPSHOT, loader.storage) start_revision = loader.storage.revision_get( [hash_to_bytes("95edacc8848369d6fb1608e887d6d2474fd5224f")] )[0] assert start_revision is not None archive_path = os.path.join(datadir, "pkg-gourmet-with-updates.tgz") repo_updated_url = prepare_repository_from_archive( archive_path, "pkg-gourmet", tmp_path ) # we'll start from start_revision loader = SvnLoader( swh_storage, repo_updated_url, origin_url=repo_initial_url, temp_directory=tmp_path, ) assert loader.load() == {"status": "eventful"} # nonetheless, we obtain the same snapshot (as previous tests on that repository) visit_status2 = assert_last_visit_matches( loader.storage, repo_updated_url, status="full", type="svn", snapshot=GOURMET_UPDATES_SNAPSHOT.id, ) assert visit_status1.date < visit_status2.date assert visit_status1.snapshot != visit_status2.snapshot stats = get_stats(loader.storage) assert stats == { "content": 22, "directory": 28, "origin": 1, "origin_visit": 2, "release": 0, "revision": 11, "skipped_content": 0, "snapshot": 2, } check_snapshot(GOURMET_UPDATES_SNAPSHOT, loader.storage) def test_loader_svn_visit_with_eol_style(swh_storage, datadir, tmp_path): """Check that a svn repo containing a versioned file with CRLF line endings with svn:eol-style property set to 'native' (this is a violation of svn specification as the file should have been stored with LF line endings) can be loaded anyway. """ archive_name = "mediawiki-repo-r407-eol-native-crlf" archive_path = os.path.join(datadir, f"{archive_name}.tgz") repo_url = prepare_repository_from_archive(archive_path, archive_name, tmp_path) loader = SvnLoader(swh_storage, repo_url, temp_directory=tmp_path) assert loader.load() == {"status": "eventful"} mediawiki_snapshot = Snapshot( id=hash_to_bytes("d6d6e9703f157c5702d9a4a5dec878926ed4ab76"), branches={ b"HEAD": SnapshotBranch( target=hash_to_bytes("7da4975c363101b819756d33459f30a866d01b1b"), target_type=TargetType.REVISION, ) }, ) check_snapshot(mediawiki_snapshot, loader.storage) assert_last_visit_matches( loader.storage, repo_url, status="full", type="svn", snapshot=mediawiki_snapshot.id, ) stats = get_stats(loader.storage) assert stats["origin"] == 1 assert stats["origin_visit"] == 1 assert stats["snapshot"] == 1 def test_loader_svn_visit_with_mixed_crlf_lf(swh_storage, datadir, tmp_path): """Check that a svn repo containing a versioned file with mixed CRLF/LF line endings with svn:eol-style property set to 'native' (this is a violation of svn specification as mixed line endings for textual content should not be stored when the svn:eol-style property is set) can be loaded anyway. """ archive_name = "pyang-repo-r343-eol-native-mixed-lf-crlf" archive_path = os.path.join(datadir, f"{archive_name}.tgz") repo_url = prepare_repository_from_archive(archive_path, archive_name, tmp_path) loader = SvnLoader(swh_storage, repo_url, temp_directory=tmp_path) assert loader.load() == {"status": "eventful"} pyang_snapshot = Snapshot( id=hash_to_bytes("6d9590de11b00a5801de0ff3297c5b44bbbf7d24"), branches={ b"HEAD": SnapshotBranch( target=hash_to_bytes("9c6962eeb9164a636c374be700672355e34a98a7"), target_type=TargetType.REVISION, ) }, ) check_snapshot(pyang_snapshot, loader.storage) assert_last_visit_matches( loader.storage, repo_url, status="full", type="svn", snapshot=pyang_snapshot.id, ) stats = get_stats(loader.storage) assert stats["origin"] == 1 assert stats["origin_visit"] == 1 assert stats["snapshot"] == 1 def test_loader_svn_with_symlink(swh_storage, datadir, tmp_path): """Repository with symlinks should be ingested ok Edge case: - first create a file and commit it. Remove it, then add folder holding the same name, commit. - do the same scenario with symbolic link (instead of file) """ archive_name = "pkg-gourmet" archive_path = os.path.join( datadir, "pkg-gourmet-with-edge-case-links-and-files.tgz" ) repo_url = prepare_repository_from_archive(archive_path, archive_name, tmp_path) loader = SvnLoader(swh_storage, repo_url, temp_directory=tmp_path) assert loader.load() == {"status": "eventful"} gourmet_edge_cases_snapshot = Snapshot( id=hash_to_bytes("18e60982fe521a2546ab8c3c73a535d80462d9d0"), branches={ b"HEAD": SnapshotBranch( target=hash_to_bytes("3f43af2578fccf18b0d4198e48563da7929dc608"), target_type=TargetType.REVISION, ) }, ) check_snapshot(gourmet_edge_cases_snapshot, loader.storage) assert_last_visit_matches( loader.storage, repo_url, status="full", type="svn", snapshot=gourmet_edge_cases_snapshot.id, ) stats = get_stats(loader.storage) assert stats["origin"] == 1 assert stats["origin_visit"] == 1 assert stats["snapshot"] == 1 assert stats["revision"] == 19 def test_loader_svn_with_wrong_symlinks(swh_storage, datadir, tmp_path): """Repository with wrong symlinks should be ingested ok nonetheless Edge case: - wrong symbolic link - wrong symbolic link with empty space names """ archive_name = "pkg-gourmet" archive_path = os.path.join(datadir, "pkg-gourmet-with-wrong-link-cases.tgz") repo_url = prepare_repository_from_archive(archive_path, archive_name, tmp_path) loader = SvnLoader(swh_storage, repo_url, temp_directory=tmp_path) assert loader.load() == {"status": "eventful"} gourmet_wrong_links_snapshot = Snapshot( id=hash_to_bytes("b17f38acabb90f066dedd30c29f01a02af88a5c4"), branches={ b"HEAD": SnapshotBranch( target=hash_to_bytes("cf30d3bb9d5967d0a2bbeacc405f10a5dd9b138a"), target_type=TargetType.REVISION, ) }, ) check_snapshot(gourmet_wrong_links_snapshot, loader.storage) assert_last_visit_matches( loader.storage, repo_url, status="full", type="svn", snapshot=gourmet_wrong_links_snapshot.id, ) stats = get_stats(loader.storage) assert stats["origin"] == 1 assert stats["origin_visit"] == 1 assert stats["snapshot"] == 1 assert stats["revision"] == 21 def test_loader_svn_cleanup_loader(swh_storage, datadir, tmp_path): """Loader should clean up its working directory after the load """ archive_name = "pkg-gourmet" archive_path = os.path.join(datadir, f"{archive_name}.tgz") repo_url = prepare_repository_from_archive(archive_path, archive_name, tmp_path) loading_temp_directory = str(tmp_path / "loading") os.mkdir(loading_temp_directory) loader = SvnLoader(swh_storage, repo_url, temp_directory=loading_temp_directory) assert loader.load() == {"status": "eventful"} # the root temporary directory still exists assert os.path.exists(loader.temp_directory) # but it should be empty assert os.listdir(loader.temp_directory) == [] def test_loader_svn_cleanup_loader_from_remote_dump(swh_storage, datadir, tmp_path): """Loader should clean up its working directory after the load """ archive_name = "pkg-gourmet" archive_path = os.path.join(datadir, f"{archive_name}.tgz") repo_url = prepare_repository_from_archive(archive_path, archive_name, tmp_path) loading_temp_directory = str(tmp_path / "loading") os.mkdir(loading_temp_directory) loader = SvnLoaderFromRemoteDump( swh_storage, repo_url, temp_directory=loading_temp_directory ) assert loader.load() == {"status": "eventful"} # the root temporary directory still exists assert os.path.exists(loader.temp_directory) # but it should be empty assert os.listdir(loader.temp_directory) == [] # the internal temp_dir should be cleaned up though assert not os.path.exists(loader.temp_dir) def test_loader_svn_cleanup_loader_from_dump_archive(swh_storage, datadir, tmp_path): """Loader should clean up its working directory after the load """ archive_ori_dump = os.path.join(datadir, "penguinsdbtools2018.dump.gz") archive_dump_dir = os.path.join(tmp_path, "dump") os.mkdir(archive_dump_dir) archive_dump = os.path.join(archive_dump_dir, "penguinsdbtools2018.dump.gz") # loader now drops the dump as soon as it's mounted so we need to make a copy first shutil.copyfile(archive_ori_dump, archive_dump) loading_path = str(tmp_path / "loading") os.mkdir(loading_path) # Prepare the dump as a local svn repository for test purposes temp_dir, repo_path = init_svn_repo_from_dump( archive_dump, root_dir=tmp_path, gzip=True ) repo_url = f"file://{repo_path}" loader = SvnLoaderFromRemoteDump(swh_storage, repo_url, temp_directory=loading_path) assert loader.load() == {"status": "eventful"} # the root temporary directory still exists assert os.path.exists(loader.temp_directory) # but it should be empty assert os.listdir(loader.temp_directory) == [] # the internal temp_dir should be cleaned up though assert not os.path.exists(loader.temp_dir) def test_svn_loader_from_remote_dump(swh_storage, datadir, tmp_path): archive_name = "pkg-gourmet" archive_path = os.path.join(datadir, f"{archive_name}.tgz") repo_url = prepare_repository_from_archive(archive_path, archive_name, tmp_path) loaderFromDump = SvnLoaderFromRemoteDump( swh_storage, repo_url, temp_directory=tmp_path ) assert loaderFromDump.load() == {"status": "eventful"} assert_last_visit_matches( loaderFromDump.storage, repo_url, status="full", type="svn", snapshot=GOURMET_SNAPSHOT.id, ) origin_url = repo_url + "2" # rename to another origin loader = SvnLoader( swh_storage, repo_url, origin_url=origin_url, temp_directory=tmp_path ) assert loader.load() == {"status": "eventful"} # because are working on new origin assert_last_visit_matches( loader.storage, origin_url, status="full", type="svn", snapshot=GOURMET_SNAPSHOT.id, ) check_snapshot(GOURMET_SNAPSHOT, loader.storage) stats = get_stats(loader.storage) assert stats["origin"] == 2 # created one more origin assert stats["origin_visit"] == 2 assert stats["snapshot"] == 1 loader = SvnLoader( swh_storage, repo_url, temp_directory=tmp_path ) # no change on the origin-url assert loader.load() == {"status": "uneventful"} assert_last_visit_matches( loader.storage, origin_url, status="full", type="svn", snapshot=GOURMET_SNAPSHOT.id, ) stats = get_stats(loader.storage) assert stats["origin"] == 2 assert stats["origin_visit"] == 3 assert stats["snapshot"] == 1 # second visit from the dump should be uneventful loaderFromDump = SvnLoaderFromRemoteDump( swh_storage, repo_url, temp_directory=tmp_path ) assert loaderFromDump.load() == {"status": "uneventful"} def test_svn_loader_from_remote_dump_incremental_load_on_stale_repo( swh_storage, datadir, tmp_path, mocker ): archive_name = "pkg-gourmet" archive_path = os.path.join(datadir, f"{archive_name}.tgz") repo_url = prepare_repository_from_archive(archive_path, archive_name, tmp_path) # first load: a dump file will be created, mounted to a local repository # and the latter will be loaded into the archive loaderFromDump = SvnLoaderFromRemoteDump( swh_storage, repo_url, temp_directory=tmp_path ) assert loaderFromDump.load() == {"status": "eventful"} assert_last_visit_matches( loaderFromDump.storage, repo_url, status="full", type="svn", snapshot=GOURMET_SNAPSHOT.id, ) # second load on same repository: the loader will detect there is no changes # since last load and will skip the dump, mount and load phases loaderFromDump = SvnLoaderFromRemoteDump( swh_storage, repo_url, temp_directory=tmp_path ) loaderFromDump.dump_svn_revisions = mocker.MagicMock() init_svn_repo_from_dump = mocker.patch( "swh.loader.svn.loader.init_svn_repo_from_dump" ) loaderFromDump.process_svn_revisions = mocker.MagicMock() loaderFromDump._check_revision_divergence = mocker.MagicMock() assert loaderFromDump.load() == {"status": "uneventful"} assert_last_visit_matches( loaderFromDump.storage, repo_url, status="full", type="svn", snapshot=GOURMET_SNAPSHOT.id, ) # no dump loaderFromDump.dump_svn_revisions.assert_not_called() # no mount init_svn_repo_from_dump.assert_not_called() # no loading loaderFromDump.process_svn_revisions.assert_not_called() # no redundant post_load processing loaderFromDump._check_revision_divergence.assert_not_called() def test_svn_loader_from_remote_dump_incremental_load_on_non_stale_repo( swh_storage, datadir, tmp_path, mocker ): archive_name = "pkg-gourmet" archive_path = os.path.join(datadir, f"{archive_name}.tgz") repo_url = prepare_repository_from_archive(archive_path, archive_name, tmp_path) # first load loader = SvnLoaderFromRemoteDump(swh_storage, repo_url, temp_directory=tmp_path) loader.load() archive_path = os.path.join(datadir, "pkg-gourmet-with-updates.tgz") repo_updated_url = prepare_repository_from_archive( archive_path, archive_name, tmp_path ) # second load loader = SvnLoaderFromRemoteDump( swh_storage, repo_updated_url, temp_directory=tmp_path ) dump_svn_revisions = mocker.spy(loader, "dump_svn_revisions") process_svn_revisions = mocker.spy(loader, "process_svn_revisions") loader.load() dump_svn_revisions.assert_called() process_svn_revisions.assert_called() def test_loader_user_defined_svn_properties(swh_storage, datadir, tmp_path): """Edge cases: The repository held some user defined svn-properties with special encodings, this prevented the repository from being loaded even though we do not ingest those information. """ archive_name = "httthttt" archive_path = os.path.join(datadir, f"{archive_name}.tgz") repo_url = prepare_repository_from_archive(archive_path, archive_name, tmp_path) loader = SvnLoader(swh_storage, repo_url) assert loader.load() == {"status": "eventful"} expected_snapshot = Snapshot( id=hash_to_bytes("70487267f682c07e52a2371061369b6cf5bffa47"), branches={ b"HEAD": SnapshotBranch( target=hash_to_bytes("604a17dbb15e8d7ecb3e9f3768d09bf493667a93"), target_type=TargetType.REVISION, ) }, ) check_snapshot(expected_snapshot, loader.storage) assert_last_visit_matches( loader.storage, repo_url, status="full", type="svn", snapshot=expected_snapshot.id, ) stats = get_stats(loader.storage) assert stats["origin"] == 1 assert stats["origin_visit"] == 1 assert stats["snapshot"] == 1 assert stats["revision"] == 7 def test_loader_svn_dir_added_then_removed(swh_storage, datadir, tmp_path): """Loader should handle directory removal when processing a commit""" archive_name = "pkg-gourmet" archive_path = os.path.join(datadir, f"{archive_name}-add-remove-dir.tgz") repo_url = prepare_repository_from_archive(archive_path, archive_name, tmp_path) loader = SvnLoader(swh_storage, repo_url, temp_directory=tmp_path) assert loader.load() == {"status": "eventful"} assert_last_visit_matches( loader.storage, repo_url, status="full", type="svn", ) check_snapshot(loader.snapshot, loader.storage) def test_loader_svn_loader_from_dump_archive(swh_storage, datadir, tmp_path): archive_name = "pkg-gourmet" archive_path = os.path.join(datadir, f"{archive_name}.tgz") repo_url = prepare_repository_from_archive(archive_path, archive_name, tmp_path) origin_url = f"svn://{archive_name}" dump_filename = f"{archive_name}.dump" with open(os.path.join(tmp_path, dump_filename), "wb") as dump_file: # create compressed dump file of pkg-gourmet repo subprocess.run(["svnrdump", "dump", repo_url], stdout=dump_file) subprocess.run(["gzip", dump_filename], cwd=tmp_path) # load svn repo from that compressed dump file loader = SvnLoaderFromDumpArchive( swh_storage, url=origin_url, archive_path=os.path.join(tmp_path, f"{dump_filename}.gz"), temp_directory=tmp_path, ) assert loader.load() == {"status": "eventful"} assert_last_visit_matches( loader.storage, origin_url, status="full", type="svn", snapshot=GOURMET_SNAPSHOT.id, ) check_snapshot(GOURMET_SNAPSHOT, loader.storage) assert get_stats(loader.storage) == { "content": 19, "directory": 17, "origin": 1, "origin_visit": 1, "release": 0, "revision": 6, "skipped_content": 0, "snapshot": 1, } class CommitChangeType(Enum): AddOrUpdate = 1 Delete = 2 class CommitChange(TypedDict, total=False): change_type: CommitChangeType path: str properties: Dict[str, str] data: bytes def add_commit(repo_url: str, message: str, changes: List[CommitChange]) -> None: conn = RemoteAccess(repo_url, auth=Auth([get_username_provider()])) editor = conn.get_commit_editor({"svn:log": message}) root = editor.open_root() for change in changes: if change["change_type"] == CommitChangeType.Delete: root.delete_entry(change["path"].rstrip("/")) else: dir_change = change["path"].endswith("/") split_path = change["path"].rstrip("/").split("/") for i in range(len(split_path)): path = "/".join(split_path[0 : i + 1]) if i < len(split_path) - 1: try: root.add_directory(path).close() except SubversionException: pass else: if dir_change: try: dir = root.add_directory(path) except SubversionException: dir = root.open_directory(path) if "properties" in change: for prop, value in change["properties"].items(): dir.change_prop(prop, value) dir.close() else: try: file = root.add_file(path) except SubversionException: file = root.open_file(path) if "properties" in change: for prop, value in change["properties"].items(): file.change_prop(prop, value) if "data" in change: txdelta = file.apply_textdelta() delta.send_stream(BytesIO(change["data"]), txdelta) file.close() root.close() editor.close() def create_repo(tmp_path): repo_path = os.path.join(tmp_path, "tmprepo") repos.create(repo_path) repo_url = f"file://{repo_path}" return repo_url @pytest.fixture def repo_url(tmpdir_factory): # create a repository return create_repo(tmpdir_factory.mktemp("repos")) def test_loader_eol_style_file_property_handling_edge_case( swh_storage, repo_url, tmp_path ): # # first commit add_commit( repo_url, ( "Add a directory containing a file with CRLF end of line " "and set svn:eol-style property to native so CRLF will be " "replaced by LF in the file when exporting the revision" ), [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="directory/file_with_crlf_eol.txt", properties={"svn:eol-style": "native"}, data=b"Hello world!\r\n", ) ], ) # second commit add_commit( repo_url, "Remove previously added directory and file", [CommitChange(change_type=CommitChangeType.Delete, path="directory/",)], ) # third commit add_commit( repo_url, ( "Add again same directory containing same file with CRLF end of line " "but do not set svn:eol-style property value so CRLF will not be " "replaced by LF when exporting the revision" ), [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="directory/file_with_crlf_eol.txt", data=b"Hello world!\r\n", ) ], ) # instantiate a svn loader checking after each processed revision that # the repository filesystem it reconstructed does not differ from a subversion # export of that revision loader = SvnLoader(swh_storage, repo_url, temp_directory=tmp_path, check_revision=1) assert loader.load() == {"status": "eventful"} assert_last_visit_matches( loader.storage, repo_url, status="full", type="svn", ) check_snapshot(loader.snapshot, loader.storage) assert get_stats(loader.storage) == { "content": 2, "directory": 5, "origin": 1, "origin_visit": 1, "release": 0, "revision": 3, "skipped_content": 0, "snapshot": 1, } def get_head_revision_paths_info(loader: SvnLoader) -> Dict[bytes, Dict[str, Any]]: assert loader.snapshot is not None root_dir = loader.snapshot.branches[b"HEAD"].target revision = loader.storage.revision_get([root_dir])[0] assert revision is not None paths = {} for entry in loader.storage.directory_ls(revision.directory, recursive=True): paths[entry["name"]] = entry return paths def test_loader_eol_style_on_svn_link_handling(swh_storage, repo_url, tmp_path): # first commit add_commit( repo_url, ( "Add a regular file, a directory and a link to the regular file " "in the directory. Set svn:eol-style property for the regular " "file and the link. Set svn:special property for the link." ), [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="file_with_crlf_eol.txt", properties={"svn:eol-style": "native"}, data=b"Hello world!\r\n", ), CommitChange( change_type=CommitChangeType.AddOrUpdate, path="directory/file_with_crlf_eol.txt", properties={"svn:eol-style": "native", "svn:special": "*"}, data=b"link ../file_with_crlf_eol.txt", ), ], ) # instantiate a svn loader checking after each processed revision that # the repository filesystem it reconstructed does not differ from a subversion # export of that revision loader = SvnLoader(swh_storage, repo_url, temp_directory=tmp_path, check_revision=1) assert loader.load() == {"status": "eventful"} assert_last_visit_matches( loader.storage, repo_url, status="full", type="svn", ) check_snapshot(loader.snapshot, loader.storage) # check loaded objects are those expected assert get_stats(loader.storage) == { "content": 2, "directory": 2, "origin": 1, "origin_visit": 1, "release": 0, "revision": 1, "skipped_content": 0, "snapshot": 1, } paths = get_head_revision_paths_info(loader) assert ( loader.storage.content_get_data(paths[b"file_with_crlf_eol.txt"]["sha1"]) == b"Hello world!\n" ) assert paths[b"directory/file_with_crlf_eol.txt"]["perms"] == DentryPerms.symlink assert ( loader.storage.content_get_data( paths[b"directory/file_with_crlf_eol.txt"]["sha1"] ) == b"../file_with_crlf_eol.txt" ) def test_loader_svn_special_property_unset(swh_storage, repo_url, tmp_path): # first commit add_commit( repo_url, ( "Create a regular file, a link to a file and a link to an " "external file. Set the svn:special property on the links." ), [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="file.txt", data=b"Hello world!\n", ), CommitChange( change_type=CommitChangeType.AddOrUpdate, path="link.txt", properties={"svn:special": "*"}, data=b"link ./file.txt", ), CommitChange( change_type=CommitChangeType.AddOrUpdate, path="external_link.txt", properties={"svn:special": "*"}, data=b"link /home/user/data.txt", ), ], ) # second commit add_commit( repo_url, "Unset the svn:special property on the links.", [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="link.txt", properties={"svn:special": None}, ), CommitChange( change_type=CommitChangeType.AddOrUpdate, path="external_link.txt", properties={"svn:special": None}, ), ], ) # instantiate a svn loader checking after each processed revision that # the repository filesystem it reconstructed does not differ from a subversion # export of that revision loader = SvnLoader(swh_storage, repo_url, temp_directory=tmp_path, check_revision=1) assert loader.load() == {"status": "eventful"} assert_last_visit_matches( loader.storage, repo_url, status="full", type="svn", ) check_snapshot(loader.snapshot, loader.storage) # check loaded objects are those expected assert get_stats(loader.storage) == { "content": 5, "directory": 2, "origin": 1, "origin_visit": 1, "release": 0, "revision": 2, "skipped_content": 0, "snapshot": 1, } paths = get_head_revision_paths_info(loader) assert paths[b"link.txt"]["perms"] == DentryPerms.content assert ( loader.storage.content_get_data(paths[b"link.txt"]["sha1"]) == b"link ./file.txt" ) assert paths[b"external_link.txt"]["perms"] == DentryPerms.content assert ( loader.storage.content_get_data(paths[b"external_link.txt"]["sha1"]) == b"link /home/user/data.txt" ) def test_loader_invalid_svn_eol_style_property_value(swh_storage, repo_url, tmp_path): filename = "file_with_crlf_eol.txt" file_content = b"Hello world!\r\n" # # first commit add_commit( repo_url, ( "Add a file with CRLF end of line and set svn:eol-style property " "to an invalid value." ), [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path=filename, properties={"svn:eol-style": "foo"}, data=file_content, ) ], ) # instantiate a svn loader checking after each processed revision that # the repository filesystem it reconstructed does not differ from a subversion # export of that revision loader = SvnLoader(swh_storage, repo_url, temp_directory=tmp_path, check_revision=1) assert loader.load() == {"status": "eventful"} assert_last_visit_matches( loader.storage, repo_url, status="full", type="svn", ) check_snapshot(loader.snapshot, loader.storage) paths = get_head_revision_paths_info(loader) # end of lines should not have been processed assert ( loader.storage.content_get_data(paths[filename.encode()]["sha1"]) == file_content ) def test_loader_first_revision_is_not_number_one( swh_storage, mocker, repo_url, tmp_path ): class SvnRepoSkipFirstRevision(SvnRepo): def logs(self, revision_start, revision_end): """Overrides logs method to skip revision number one in yielded revisions""" yield from super().logs(revision_start + 1, revision_end) from swh.loader.svn import loader mocker.patch.object(loader, "SvnRepo", SvnRepoSkipFirstRevision) for filename in ("foo", "bar", "baz"): add_commit( repo_url, f"Add {filename} file", [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path=filename, data=f"{filename}\n".encode(), ) ], ) loader = SvnLoader(swh_storage, repo_url, temp_directory=tmp_path) # post loading will detect an issue and make a partial visit with a snapshot assert loader.load() == {"status": "failed"} assert_last_visit_matches( loader.storage, repo_url, status="partial", type="svn", ) check_snapshot(loader.snapshot, loader.storage) assert get_stats(loader.storage) == { "content": 2, "directory": 2, "origin": 1, "origin_visit": 1, "release": 0, "revision": 2, "skipped_content": 0, "snapshot": 1, } def test_loader_svn_special_property_on_binary_file(swh_storage, repo_url, tmp_path): """When a file has the svn:special property set but is not a svn link, it might be truncated under certain conditions when performing an export operation.""" data = ( b"!\xff\xfea\x00p\x00t\x00-\x00c\x00y\x00g\x00.\x00s\x00h\x00\x00\x00" ) # first commit add_commit( repo_url, ( "Add a non svn link binary file and set the svn:special property on it." "That file will be truncated when exporting it." ), [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="binary_file", properties={"svn:special": "*"}, data=data, ), ], ) # second commit add_commit( repo_url, ( "Add a non svn link binary file and set the svn:special and " "svn:mime-type properties on it." "That file will not be truncated when exporting it." ), [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="another_binary_file", properties={ "svn:special": "*", "svn:mime-type": "application/octet-stream", }, data=data, ), ], ) # third commit add_commit( repo_url, "Remove the svn:special property on the previously added files", [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="binary_file", properties={"svn:special": None}, ), CommitChange( change_type=CommitChangeType.AddOrUpdate, path="another_binary_file", properties={"svn:special": None}, ), ], ) # instantiate a svn loader checking after each processed revision that # the repository filesystem it reconstructed does not differ from a subversion # export of that revision loader = SvnLoader(swh_storage, repo_url, temp_directory=tmp_path, check_revision=1) assert loader.load() == {"status": "eventful"} assert_last_visit_matches( loader.storage, repo_url, status="full", type="svn", ) check_snapshot(loader.snapshot, loader.storage) def test_loader_last_revision_divergence(swh_storage, datadir, tmp_path): archive_name = "pkg-gourmet" archive_path = os.path.join(datadir, f"{archive_name}.tgz") repo_url = prepare_repository_from_archive(archive_path, archive_name, tmp_path) class SvnLoaderRevisionDivergence(SvnLoader): def _check_revision_divergence(self, count, rev, dir_id): raise ValueError("revision divergence detected") loader = SvnLoaderRevisionDivergence(swh_storage, repo_url, temp_directory=tmp_path) assert loader.load() == {"status": "failed"} assert_last_visit_matches( loader.storage, repo_url, status="partial", type="svn", snapshot=GOURMET_SNAPSHOT.id, ) check_snapshot(GOURMET_SNAPSHOT, loader.storage) def test_loader_delete_directory_while_file_has_same_prefix( swh_storage, repo_url, tmp_path ): # first commit add_commit( repo_url, "Add a file and a directory with same prefix", [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="foo/bar.c", data=b'#include "../foo.c"', ), CommitChange( change_type=CommitChangeType.AddOrUpdate, path="foo.c", data=b"int foo() {return 0;}", ), ], ) # second commit add_commit( repo_url, "Delete previously added directory and update file content", [ CommitChange(change_type=CommitChangeType.Delete, path="foo"), CommitChange( change_type=CommitChangeType.AddOrUpdate, path="foo.c", data=b"int foo() {return 1;}", ), ], ) # instantiate a svn loader checking after each processed revision that # the repository filesystem it reconstructed does not differ from a subversion # export of that revision loader = SvnLoader(swh_storage, repo_url, temp_directory=tmp_path, check_revision=1) assert loader.load() == {"status": "eventful"} assert_last_visit_matches( loader.storage, repo_url, status="full", type="svn", ) check_snapshot(loader.snapshot, loader.storage) def test_svn_loader_incremental(swh_storage, repo_url, tmp_path): # first commit add_commit( repo_url, ( "Add a directory containing a file with CRLF end of line " "and set svn:eol-style property to native so CRLF will be " "replaced by LF in the file when exporting the revision" ), [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="file_with_crlf_eol.txt", properties={"svn:eol-style": "native"}, data=b"Hello world!\r\n", ) ], ) # first load loader = SvnLoader(swh_storage, repo_url, temp_directory=tmp_path, check_revision=1) assert loader.load() == {"status": "eventful"} assert_last_visit_matches( loader.storage, repo_url, status="full", type="svn", ) check_snapshot(loader.snapshot, loader.storage) # second commit add_commit( repo_url, "Modify previously added file", [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="file_with_crlf_eol.txt", data=b"Hello World!\r\n", ) ], ) # second load, incremental loader = SvnLoader(swh_storage, repo_url, temp_directory=tmp_path, check_revision=1) assert loader.load() == {"status": "eventful"} assert_last_visit_matches( loader.storage, repo_url, status="full", type="svn", ) check_snapshot(loader.snapshot, loader.storage) # third commit add_commit( repo_url, "Unset svn:eol-style property on file", [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="file_with_crlf_eol.txt", properties={"svn:eol-style": None}, ) ], ) # third load, incremental loader = SvnLoader(swh_storage, repo_url, temp_directory=tmp_path, check_revision=1) assert loader.load() == {"status": "eventful"} assert_last_visit_matches( loader.storage, repo_url, status="full", type="svn", ) check_snapshot(loader.snapshot, loader.storage) def test_svn_loader_incremental_replay_start_with_empty_directory( swh_storage, mocker, repo_url, tmp_path ): # first commit add_commit( repo_url, ("Add a file"), [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="foo.txt", data=b"foo\n", ) ], ) # first load loader = SvnLoader(swh_storage, repo_url, temp_directory=tmp_path) assert loader.load() == {"status": "eventful"} assert_last_visit_matches( loader.storage, repo_url, status="full", type="svn", ) check_snapshot(loader.snapshot, loader.storage) # second commit add_commit( repo_url, "Modify previously added file", [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="foo.txt", data=b"bar\n", ) ], ) class SvnRepoCheckReplayStartWithEmptyDirectory(SvnRepo): def swh_hash_data_per_revision(self, start_revision: int, end_revision: int): """Overrides swh_hash_data_per_revision method to grab the content of the directory where the svn revisions will be replayed before that process starts.""" self.replay_dir_content_before_start = [ os.path.join(root, name) for root, _, files in os.walk(self.local_url) for name in files ] yield from super().swh_hash_data_per_revision(start_revision, end_revision) from swh.loader.svn import loader mocker.patch.object(loader, "SvnRepo", SvnRepoCheckReplayStartWithEmptyDirectory) # second load, incremental loader = SvnLoader(swh_storage, repo_url, temp_directory=tmp_path) loader.load() # check work directory was empty before replaying revisions assert loader.svnrepo.replay_dir_content_before_start == [] def test_loader_svn_executable_property_on_svn_link_handling( swh_storage, repo_url, tmp_path ): # first commit add_commit( repo_url, ( "Add an executable file and a svn link to it." "Set svn:executable property for both paths." ), [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="hello-world", properties={"svn:executable": "*"}, data=b"#!/bin/bash\necho Hello World !", ), CommitChange( change_type=CommitChangeType.AddOrUpdate, path="hello", properties={"svn:executable": "*", "svn:special": "*"}, data=b"link hello-world", ), ], ) # second commit add_commit( repo_url, ( "Remove executable file, unset link and replace it with executable content." "As the link was previously marked as executable, execution rights should" "be set after turning it to a regular file." ), [ CommitChange(change_type=CommitChangeType.Delete, path="hello-world"), CommitChange( change_type=CommitChangeType.AddOrUpdate, path="hello", properties={"svn:special": None}, data=b"#!/bin/bash\necho Hello World !", ), ], ) # instantiate a svn loader checking after each processed revision that # the repository filesystem it reconstructed does not differ from a subversion # export of that revision loader = SvnLoader(swh_storage, repo_url, temp_directory=tmp_path, check_revision=1) assert loader.load() == {"status": "eventful"} assert_last_visit_matches( loader.storage, repo_url, status="full", type="svn", ) check_snapshot(loader.snapshot, loader.storage) def test_loader_svn_add_property_on_link(swh_storage, repo_url, tmp_path): # first commit add_commit( repo_url, "Add an executable file and a svn link to it.", [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="hello-world", properties={"svn:executable": "*"}, data=b"#!/bin/bash\necho Hello World !", ), CommitChange( change_type=CommitChangeType.AddOrUpdate, path="hello", properties={"svn:special": "*"}, data=b"link hello-world", ), ], ) # second commit add_commit( repo_url, "Set svn:eol-style property on link", [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="hello", properties={"svn:eol-style": "native"}, ), ], ) # instantiate a svn loader checking after each processed revision that # the repository filesystem it reconstructed does not differ from a subversion # export of that revision loader = SvnLoader(swh_storage, repo_url, temp_directory=tmp_path, check_revision=1) assert loader.load() == {"status": "eventful"} assert_last_visit_matches( loader.storage, repo_url, status="full", type="svn", ) check_snapshot(loader.snapshot, loader.storage) def test_loader_svn_link_parsing(swh_storage, repo_url, tmp_path): # first commit add_commit( repo_url, "Add an executable file and a svn link to it.", [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="hello-world", properties={"svn:executable": "*"}, data=b"#!/bin/bash\necho Hello World !", ), CommitChange( change_type=CommitChangeType.AddOrUpdate, path="hello", properties={"svn:special": "*"}, data=b"link hello-world", ), ], ) # second commit add_commit( repo_url, "Update svn link content", [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="hello", data=b"link hello-world\r\n", ), ], ) # instantiate a svn loader checking after each processed revision that # the repository filesystem it reconstructed does not differ from a subversion # export of that revision loader = SvnLoader(swh_storage, repo_url, temp_directory=tmp_path, check_revision=1) assert loader.load() == {"status": "eventful"} assert_last_visit_matches( loader.storage, repo_url, status="full", type="svn", ) check_snapshot(loader.snapshot, loader.storage) def test_loader_svn_empty_local_dir_before_post_load(swh_storage, datadir, tmp_path): archive_name = "pkg-gourmet" archive_path = os.path.join(datadir, f"{archive_name}.tgz") repo_url = prepare_repository_from_archive(archive_path, archive_name, tmp_path) class SvnLoaderPostLoadLocalDirIsEmpty(SvnLoader): def post_load(self, success=True): if success: self.local_dirname_content = [ os.path.join(root, name) for root, _, files in os.walk(self.svnrepo.local_dirname) for name in files ] return super().post_load(success) loader = SvnLoaderPostLoadLocalDirIsEmpty( swh_storage, repo_url, temp_directory=tmp_path ) assert loader.load() == {"status": "eventful"} assert loader.local_dirname_content == [] assert_last_visit_matches( loader.storage, repo_url, status="full", type="svn", snapshot=GOURMET_SNAPSHOT.id, ) check_snapshot(GOURMET_SNAPSHOT, loader.storage) def test_loader_svn_add_property_on_directory_link(swh_storage, repo_url, tmp_path): # first commit add_commit( repo_url, "Add an executable file in a directory and a svn link to the directory.", [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="code/hello-world", properties={"svn:executable": "*"}, data=b"#!/bin/bash\necho Hello World !", ), CommitChange( change_type=CommitChangeType.AddOrUpdate, path="hello", properties={"svn:special": "*"}, data=b"link code", ), ], ) # second commit add_commit( repo_url, "Set svn:eol-style property on link", [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="hello", properties={"svn:eol-style": "native"}, ), ], ) # instantiate a svn loader checking after each processed revision that # the repository filesystem it reconstructed does not differ from a subversion # export of that revision loader = SvnLoader(swh_storage, repo_url, temp_directory=tmp_path, check_revision=1) assert loader.load() == {"status": "eventful"} assert_last_visit_matches( loader.storage, repo_url, status="full", type="svn", ) check_snapshot(loader.snapshot, loader.storage) @pytest.fixture def external_repo_url(tmpdir_factory): # create a repository return create_repo(tmpdir_factory.mktemp("external")) def test_loader_with_valid_svn_externals( swh_storage, repo_url, external_repo_url, tmp_path ): # first commit on external add_commit( external_repo_url, "Create some directories and files in an external repository", [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="code/hello/hello-world", properties={"svn:executable": "*"}, data=b"#!/bin/bash\necho Hello World !", ), CommitChange( change_type=CommitChangeType.AddOrUpdate, path="foo.sh", properties={"svn:executable": "*"}, data=b"#!/bin/bash\necho foo", ), ], ) # first commit add_commit( repo_url, "Create repository structure.", [ CommitChange(change_type=CommitChangeType.AddOrUpdate, path="branches/",), CommitChange(change_type=CommitChangeType.AddOrUpdate, path="tags/",), CommitChange(change_type=CommitChangeType.AddOrUpdate, path="trunk/",), ], ) # second commit add_commit( repo_url, ( "Set svn:externals property on trunk/externals path of repository to load." "One external targets a remote directory and another one a remote file." ), [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="trunk/externals/", properties={ "svn:externals": ( f"{svn_urljoin(external_repo_url, 'code/hello')} hello\n" f"{svn_urljoin(external_repo_url, 'foo.sh')} foo.sh" ) }, ), ], ) # first load loader = SvnLoader(swh_storage, repo_url, temp_directory=tmp_path, check_revision=1) assert loader.load() == {"status": "eventful"} assert_last_visit_matches( loader.storage, repo_url, status="full", type="svn", ) check_snapshot(loader.snapshot, loader.storage) # third commit add_commit( repo_url, "Unset svn:externals property on trunk/externals path", [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="trunk/externals/", properties={"svn:externals": None}, ), ], ) # second load loader = SvnLoader(swh_storage, repo_url, temp_directory=tmp_path, check_revision=1) assert loader.load() == {"status": "eventful"} assert_last_visit_matches( loader.storage, repo_url, status="full", type="svn", ) check_snapshot(loader.snapshot, loader.storage) def test_loader_with_invalid_svn_externals(swh_storage, repo_url, tmp_path): # first commit add_commit( repo_url, "Create repository structure.", [ CommitChange(change_type=CommitChangeType.AddOrUpdate, path="branches/",), CommitChange(change_type=CommitChangeType.AddOrUpdate, path="tags/",), CommitChange(change_type=CommitChangeType.AddOrUpdate, path="trunk/",), ], ) # second commit add_commit( repo_url, ( "Set svn:externals property on trunk/externals path of repository to load." "The externals URLs are not valid." ), [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="trunk/externals/", properties={ "svn:externals": ( "file:///tmp/invalid/svn/repo/hello hello\n" "file:///tmp/invalid/svn/repo/foo.sh foo.sh" ) }, ), ], ) loader = SvnLoader(swh_storage, repo_url, temp_directory=tmp_path, check_revision=1) assert loader.load() == {"status": "eventful"} assert_last_visit_matches( loader.storage, repo_url, status="full", type="svn", ) check_snapshot(loader.snapshot, loader.storage) def test_loader_with_valid_externals_modification( swh_storage, repo_url, external_repo_url, tmp_path ): # first commit on external add_commit( external_repo_url, "Create some directories and files in an external repository", [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="code/hello/hello-world", properties={"svn:executable": "*"}, data=b"#!/bin/bash\necho Hello World !", ), CommitChange( change_type=CommitChangeType.AddOrUpdate, path="code/bar/bar.sh", properties={"svn:executable": "*"}, data=b"#!/bin/bash\necho bar", ), CommitChange( change_type=CommitChangeType.AddOrUpdate, path="foo.sh", properties={"svn:executable": "*"}, data=b"#!/bin/bash\necho foo", ), ], ) # first commit add_commit( repo_url, ("Set svn:externals property on trunk/externals path of repository to load."), [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="trunk/externals/", properties={ "svn:externals": ( f"{svn_urljoin(external_repo_url, 'code/hello')} src/code/hello\n" # noqa f"{svn_urljoin(external_repo_url, 'foo.sh')} src/foo.sh\n" ) }, ), ], ) # second commit add_commit( repo_url, ( "Modify svn:externals property on trunk/externals path of repository to load." # noqa ), [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="trunk/externals/", properties={ "svn:externals": ( f"{svn_urljoin(external_repo_url, 'code/bar')} src/code/bar\n" # noqa f"{svn_urljoin(external_repo_url, 'foo.sh')} src/foo.sh\n" ) }, ), ], ) loader = SvnLoader(swh_storage, repo_url, temp_directory=tmp_path, check_revision=1) assert loader.load() == {"status": "eventful"} assert_last_visit_matches( loader.storage, repo_url, status="full", type="svn", ) check_snapshot(loader.snapshot, loader.storage) def test_loader_with_valid_externals_and_versioned_path( swh_storage, repo_url, external_repo_url, tmp_path ): # first commit on external add_commit( external_repo_url, "Create a file in an external repository", [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="code/script.sh", data=b"#!/bin/bash\necho Hello World !", ), ], ) # first commit add_commit( repo_url, "Add file with same name but different content in main repository", [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="trunk/script.sh", data=b"#!/bin/bash\necho foo", ), ], ) # second commit add_commit( repo_url, "Add externals targeting the versioned file", [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="trunk/", properties={ "svn:externals": ( f"{svn_urljoin(external_repo_url, 'code/script.sh')} script.sh" # noqa ) }, ), ], ) # third commit add_commit( repo_url, "Modify the versioned file", [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="trunk/script.sh", data=b"#!/bin/bash\necho bar", ), ], ) loader = SvnLoader(swh_storage, repo_url, temp_directory=tmp_path, check_revision=1) assert loader.load() == {"status": "eventful"} assert_last_visit_matches( loader.storage, repo_url, status="full", type="svn", ) check_snapshot(loader.snapshot, loader.storage) def test_loader_with_invalid_externals_and_versioned_path( swh_storage, repo_url, tmp_path ): # first commit add_commit( repo_url, "Add file in main repository", [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="trunk/script.sh", data=b"#!/bin/bash\necho foo", ), ], ) # second commit add_commit( repo_url, "Add invalid externals targeting the versioned file", [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="trunk/", properties={ "svn:externals": ( "file:///tmp/invalid/svn/repo/code/script.sh script.sh" ) }, ), ], ) loader = SvnLoader(swh_storage, repo_url, temp_directory=tmp_path, check_revision=1) assert loader.load() == {"status": "eventful"} assert_last_visit_matches( loader.storage, repo_url, status="full", type="svn", ) check_snapshot(loader.snapshot, loader.storage) def test_loader_set_externals_then_remove_and_add_as_local( swh_storage, repo_url, external_repo_url, tmp_path ): # first commit on external add_commit( external_repo_url, "Create a file in an external repository", [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="code/script.sh", data=b"#!/bin/bash\necho Hello World !", ), ], ) # first commit add_commit( repo_url, "Add trunk directory and set externals", [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="trunk/", properties={ "svn:externals": (f"{svn_urljoin(external_repo_url, 'code')} code") }, ), ], ) # second commit add_commit( repo_url, "Unset externals on trunk and add remote path as local path", [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="trunk/", properties={"svn:externals": None}, ), CommitChange( change_type=CommitChangeType.AddOrUpdate, path="trunk/code/script.sh", data=b"#!/bin/bash\necho Hello World !", ), ], ) loader = SvnLoader(swh_storage, repo_url, temp_directory=tmp_path, check_revision=1) assert loader.load() == {"status": "eventful"} assert_last_visit_matches( loader.storage, repo_url, status="full", type="svn", ) check_snapshot(loader.snapshot, loader.storage) def test_loader_set_invalid_externals_then_remove(swh_storage, repo_url, tmp_path): # first commit add_commit( repo_url, "Add trunk directory and set invalid external", [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="trunk/", properties={ "svn:externals": "file:///tmp/invalid/svn/repo/code external/code" }, ), ], ) # second commit add_commit( repo_url, "Unset externals on trunk", [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="trunk/", properties={"svn:externals": None}, ), ], ) loader = SvnLoader(swh_storage, repo_url, temp_directory=tmp_path, check_revision=1) assert loader.load() == {"status": "eventful"} assert_last_visit_matches( loader.storage, repo_url, status="full", type="svn", ) check_snapshot(loader.snapshot, loader.storage) def test_loader_set_externals_with_versioned_file_overlap( swh_storage, repo_url, external_repo_url, tmp_path ): # first commit on external add_commit( external_repo_url, "Create a file in an external repository", [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="code/script.sh", data=b"#!/bin/bash\necho Hello World !", ), ], ) # first commit add_commit( repo_url, "Add file with same name as in the external repository", [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="trunk/script.sh", data=b"#!/bin/bash\necho foo", ), ], ) # second commit add_commit( repo_url, "Set external on trunk overlapping versioned file", [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="trunk/", properties={ "svn:externals": ( f"{svn_urljoin(external_repo_url, 'code/script.sh')} script.sh" ) }, ), ], ) # third commit add_commit( repo_url, "Unset externals on trunk", [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="trunk/", properties={"svn:externals": None}, ), ], ) loader = SvnLoader(swh_storage, repo_url, temp_directory=tmp_path, check_revision=1) assert loader.load() == {"status": "eventful"} assert_last_visit_matches( loader.storage, repo_url, status="full", type="svn", ) check_snapshot(loader.snapshot, loader.storage) def test_dump_loader_relative_externals_detection( swh_storage, repo_url, external_repo_url, tmp_path ): add_commit( external_repo_url, "Create a file in external repository.", [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="project1/foo.sh", data=b"#!/bin/bash\necho foo", ), ], ) add_commit( external_repo_url, "Create another file in repository to load.", [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="project2/bar.sh", data=b"#!/bin/bash\necho bar", ), ], ) external_url = f"{external_repo_url.replace('file://', '//')}/project2/bar.sh" add_commit( repo_url, "Set external relative to URL scheme in repository to load", [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="project1/", properties={"svn:externals": (f"{external_url} bar.sh")}, ), ], ) loader = SvnLoaderFromRemoteDump( swh_storage, repo_url, temp_directory=tmp_path, check_revision=1 ) assert loader.load() == {"status": "eventful"} assert_last_visit_matches( loader.storage, repo_url, status="full", type="svn", ) check_snapshot(loader.snapshot, loader.storage) assert loader.svnrepo.has_relative_externals add_commit( repo_url, "Unset external in repository to load", [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="project1/", properties={"svn:externals": None}, ), ], ) loader = SvnLoaderFromRemoteDump( swh_storage, repo_url, temp_directory=tmp_path, check_revision=1 ) assert loader.load() == {"status": "eventful"} assert_last_visit_matches( loader.storage, repo_url, status="full", type="svn", ) check_snapshot(loader.snapshot, loader.storage) assert not loader.svnrepo.has_relative_externals def test_loader_externals_cache(swh_storage, repo_url, external_repo_url, tmp_path): # first commit on external add_commit( external_repo_url, "Create some directories and files in an external repository", [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="code/hello/hello-world", properties={"svn:executable": "*"}, data=b"#!/bin/bash\necho Hello World !", ), ], ) # first commit add_commit( repo_url, "Create repository structure.", [ CommitChange(change_type=CommitChangeType.AddOrUpdate, path="project1/",), CommitChange(change_type=CommitChangeType.AddOrUpdate, path="project2/",), ], ) external_url = svn_urljoin(external_repo_url, "code/hello") # second commit add_commit( repo_url, ( "Set svn:externals property on trunk/externals path of repository to load." "One external targets a remote directory and another one a remote file." ), [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="project1/externals/", properties={"svn:externals": (f"{external_url} hello\n")}, ), CommitChange( change_type=CommitChangeType.AddOrUpdate, path="project2/externals/", properties={"svn:externals": (f"{external_url} hello\n")}, ), ], ) loader = SvnLoader(swh_storage, repo_url, temp_directory=tmp_path, check_revision=1) assert loader.load() == {"status": "eventful"} assert_last_visit_matches( loader.storage, repo_url, status="full", type="svn", ) check_snapshot(loader.snapshot, loader.storage) assert (external_url, None) in loader.svnrepo.swhreplay.editor.externals_cache def test_loader_remove_versioned_path_with_external_overlap( swh_storage, repo_url, external_repo_url, tmp_path ): # first commit on external add_commit( external_repo_url, "Create a file in an external repository", [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="code/hello.sh", data=b"#!/bin/bash\necho Hello World !", ), ], ) # first commit add_commit( repo_url, "Add a file", [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="trunk/project/script.sh", data=b"#!/bin/bash\necho foo", ), ], ) # second commit add_commit( repo_url, "Set external on trunk overlapping versioned path", [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="trunk/", properties={ "svn:externals": ( f"{svn_urljoin(external_repo_url, 'code')} project/code" ) }, ), ], ) # third commit add_commit( repo_url, "Remove trunk/project/ versioned path", [CommitChange(change_type=CommitChangeType.Delete, path="trunk/project/",),], ) loader = SvnLoader( swh_storage, repo_url, temp_directory=tmp_path, check_revision=1, ) assert loader.load() == {"status": "eventful"} assert_last_visit_matches( loader.storage, repo_url, status="full", type="svn", ) check_snapshot(loader.snapshot, loader.storage) def test_loader_export_external_path_using_peg_rev( swh_storage, repo_url, external_repo_url, tmp_path ): # first commit on external add_commit( external_repo_url, "Create a file in an external repository", [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="code/foo.sh", data=b"#!/bin/bash\necho foo", ), ], ) # second commit on external add_commit( external_repo_url, "Remove previously added file", [CommitChange(change_type=CommitChangeType.Delete, path="code/foo.sh",),], ) # third commit on external add_commit( external_repo_url, "Add file again but with different content", [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="code/foo.sh", data=b"#!/bin/bash\necho bar", ), ], ) # first commit add_commit( repo_url, "Add trunk dir", [CommitChange(change_type=CommitChangeType.AddOrUpdate, path="trunk/",),], ) # second commit add_commit( repo_url, "Set external on trunk targeting first revision of external repo", [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="trunk/", properties={ "svn:externals": ( f"{svn_urljoin(external_repo_url, 'code/foo.sh')}@1 foo.sh" ) }, ), ], ) # third commit add_commit( repo_url, "Modify external on trunk to target third revision of external repo", [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="trunk/", properties={ "svn:externals": ( f"{svn_urljoin(external_repo_url, 'code/foo.sh')}@3 foo.sh" ) }, ), ], ) loader = SvnLoader( swh_storage, repo_url, temp_directory=tmp_path, check_revision=1, ) assert loader.load() == {"status": "eventful"} assert_last_visit_matches( loader.storage, repo_url, status="full", type="svn", ) check_snapshot(loader.snapshot, loader.storage) def test_loader_remove_external_overlapping_versioned_path( swh_storage, repo_url, external_repo_url, tmp_path ): # first commit on external add_commit( external_repo_url, "Create a file in an external repository", [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="code/foo.sh", data=b"#!/bin/bash\necho foo", ), ], ) # first commit add_commit( repo_url, "Add trunk dir", [CommitChange(change_type=CommitChangeType.AddOrUpdate, path="trunk/")], ) # second commit add_commit( repo_url, "Set external on root dir overlapping versioned trunk path", [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="", # repo root dir properties={ "svn:externals": ( f"{svn_urljoin(external_repo_url, 'code/foo.sh')} trunk/code/foo.sh" # noqa ) }, ), ], ) # third commit add_commit( repo_url, "Remove external on root dir", [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="", properties={"svn:externals": None}, ), ], ) loader = SvnLoader( swh_storage, repo_url, temp_directory=tmp_path, check_revision=1, ) assert loader.load() == {"status": "eventful"} assert_last_visit_matches( loader.storage, repo_url, status="full", type="svn", ) check_snapshot(loader.snapshot, loader.storage) def test_loader_modify_external_same_path( swh_storage, repo_url, external_repo_url, tmp_path ): # first commit on external add_commit( external_repo_url, "Create a file in an external repository", [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="code/foo.sh", data=b"#!/bin/bash\necho foo", ), ], ) # first commit add_commit( repo_url, "Add trunk dir", [CommitChange(change_type=CommitChangeType.AddOrUpdate, path="trunk/")], ) # second commit add_commit( repo_url, "Set external code on trunk dir", [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="trunk/", properties={ "svn:externals": (f"{svn_urljoin(external_repo_url, 'code')} code") }, ), ], ) # third commit add_commit( repo_url, "Change code external on trunk targeting an invalid URL", [ CommitChange( change_type=CommitChangeType.AddOrUpdate, path="trunk/", properties={"svn:externals": "file:///tmp/invalid/svn/repo/path code"}, ), ], ) loader = SvnLoader( swh_storage, repo_url, temp_directory=tmp_path, check_revision=1, ) assert loader.load() == {"status": "eventful"} assert_last_visit_matches( loader.storage, repo_url, status="full", type="svn", ) check_snapshot(loader.snapshot, loader.storage) + + +def test_loader_with_recursive_external( + swh_storage, repo_url, external_repo_url, tmp_path +): + # first commit on external + add_commit( + external_repo_url, + "Create a file in an external repository", + [ + CommitChange( + change_type=CommitChangeType.AddOrUpdate, + path="code/foo.sh", + data=b"#!/bin/bash\necho foo", + ), + ], + ) + + # first commit + add_commit( + repo_url, + "Add trunk dir and a file", + [ + CommitChange( + change_type=CommitChangeType.AddOrUpdate, + path="trunk/bar.sh", + data=b"#!/bin/bash\necho bar", + ) + ], + ) + + # second commit + add_commit( + repo_url, + "Set externals code on trunk/externals dir, one being recursive", + [ + CommitChange( + change_type=CommitChangeType.AddOrUpdate, + path="trunk/externals/", + properties={ + "svn:externals": ( + f"{svn_urljoin(external_repo_url, 'code')} code\n" + f"{repo_url} recursive" + ) + }, + ), + ], + ) + + # first load + loader = SvnLoader( + swh_storage, repo_url, temp_directory=tmp_path, check_revision=1, + ) + assert loader.load() == {"status": "eventful"} + assert_last_visit_matches( + loader.storage, repo_url, status="full", type="svn", + ) + check_snapshot(loader.snapshot, loader.storage) + assert loader.svnrepo.has_recursive_externals + + # second load on stale repo + loader = SvnLoader( + swh_storage, repo_url, temp_directory=tmp_path, check_revision=1, + ) + assert loader.load() == {"status": "uneventful"} + assert_last_visit_matches( + loader.storage, repo_url, status="full", type="svn", + ) + check_snapshot(loader.snapshot, loader.storage) + assert loader.svnrepo.has_recursive_externals + + # third commit + add_commit( + repo_url, + "Remove recursive external on trunk/externals dir", + [ + CommitChange( + change_type=CommitChangeType.AddOrUpdate, + path="trunk/externals/", + properties={ + "svn:externals": (f"{svn_urljoin(external_repo_url, 'code')} code") + }, + ), + ], + ) + + # third load + loader = SvnLoader( + swh_storage, repo_url, temp_directory=tmp_path, check_revision=1, + ) + assert loader.load() == {"status": "eventful"} + assert_last_visit_matches( + loader.storage, repo_url, status="full", type="svn", + ) + check_snapshot(loader.snapshot, loader.storage) + assert not loader.svnrepo.has_recursive_externals diff --git a/swh/loader/svn/utils.py b/swh/loader/svn/utils.py index 4311536..7761f18 100644 --- a/swh/loader/svn/utils.py +++ b/swh/loader/svn/utils.py @@ -1,294 +1,320 @@ # Copyright (C) 2016-2022 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information import errno import logging import os import re import shutil from subprocess import PIPE, Popen, call import tempfile from typing import Optional, Tuple -from urllib.parse import urlparse +from urllib.parse import quote, urlparse, urlunparse logger = logging.getLogger(__name__) class OutputStream: """Helper class to read lines from a program output while it is running Args: fileno (int): File descriptor of a program output stream opened in text mode """ def __init__(self, fileno): self._fileno = fileno self._buffer = "" def read_lines(self): """ Read available lines from the output stream and return them. Returns: Tuple[List[str], bool]: A tuple whose first member is the read lines and second member a boolean indicating if there are still some other lines available to read. """ try: output = os.read(self._fileno, 1000).decode() except OSError as e: if e.errno != errno.EIO: raise output = "" output = output.replace("\r\n", "\n") lines = output.split("\n") lines[0] = self._buffer + lines[0] if output: self._buffer = lines[-1] return (lines[:-1], True) else: self._buffer = "" if len(lines) == 1 and not lines[0]: lines = [] return (lines, False) def init_svn_repo_from_dump( dump_path: str, prefix: Optional[str] = None, suffix: Optional[str] = None, root_dir: str = "/tmp", gzip: bool = False, cleanup_dump: bool = True, ) -> Tuple[str, str]: """Given a path to a svn dump, initialize an svn repository with the content of said dump. Args: dump_path: The dump to the path prefix: optional prefix file name for the working directory suffix: optional suffix file name for the working directory root_dir: the root directory where the working directory is created gzip: Boolean to determine whether we treat the dump as compressed or not. cleanup_dump: Whether we want this function call to clean up the dump at the end of the repository initialization. Raises: ValueError in case of failure to run the command to uncompress and load the dump. Returns: A tuple: - temporary folder: containing the mounted repository - repo_path: path to the mounted repository inside the temporary folder """ project_name = os.path.basename(os.path.dirname(dump_path)) temp_dir = tempfile.mkdtemp(prefix=prefix, suffix=suffix, dir=root_dir) try: repo_path = os.path.join(temp_dir, project_name) # create the repository that will be loaded with the dump cmd = ["svnadmin", "create", repo_path] r = call(cmd) if r != 0: raise ValueError( "Failed to initialize empty svn repo for %s" % project_name ) read_dump_cmd = ["cat", dump_path] if gzip: read_dump_cmd = ["gzip", "-dc", dump_path] with Popen(read_dump_cmd, stdout=PIPE) as dump: # load dump and bypass properties validation as Unicode decoding errors # are already handled in loader implementation (see _ra_codecs_error_handler # in ra.py) cmd = ["svnadmin", "load", "-q", "--bypass-prop-validation", repo_path] r = call(cmd, stdin=dump.stdout) if r != 0: raise ValueError( "Failed to mount the svn dump for project %s" % project_name ) return temp_dir, repo_path except Exception as e: shutil.rmtree(temp_dir) raise e finally: if cleanup_dump: try: # At this time, the temporary svn repository is mounted from the dump or # the svn repository failed to mount. Either way, we can drop the dump. os.remove(dump_path) assert not os.path.exists(dump_path) except OSError as e: logger.warn("Failure to remove the dump %s: %s", dump_path, e) def init_svn_repo_from_archive_dump( archive_path: str, prefix: Optional[str] = None, suffix: Optional[str] = None, root_dir: str = "/tmp", cleanup_dump: bool = True, ) -> Tuple[str, str]: """Given a path to an archive containing an svn dump, initializes an svn repository with the content of the uncompressed dump. Args: archive_path: The archive svn dump path prefix: optional prefix file name for the working directory suffix: optional suffix file name for the working directory root_dir: the root directory where the working directory is created gzip: Boolean to determine whether we treat the dump as compressed or not. cleanup_dump: Whether we want this function call to clean up the dump at the end of the repository initialization. Raises: ValueError in case of failure to run the command to uncompress and load the dump. Returns: A tuple: - temporary folder: containing the mounted repository - repo_path: path to the mounted repository inside the temporary folder """ return init_svn_repo_from_dump( archive_path, prefix=prefix, suffix=suffix, root_dir=root_dir, gzip=True, cleanup_dump=cleanup_dump, ) def svn_urljoin(base_url: str, *args) -> str: """Join a base URL and a list of paths in a SVN way. For instance: - svn_urljoin("http://example.org", "foo", "bar") will return "https://example.org/foo/bar - svn_urljoin("http://example.org/foo", "../bar") will return "https://example.org/bar Args: base_url: Base URL to join paths with args: path components Returns: The joined URL """ parsed_url = urlparse(base_url) path = os.path.abspath( os.path.join(parsed_url.path or "/", *[arg.strip("/") for arg in args]) ) return f"{parsed_url.scheme}://{parsed_url.netloc}{path}" def parse_external_definition( external: str, dir_path: str, repo_url: str ) -> Tuple[str, str, Optional[int], bool]: """Parse a subversion external definition. Args: external: an external definition, extracted from the lines split of a svn:externals property value dir_path: The path of the directory in the subversion repository where the svn:externals property was set repo_url: URL of the subversion repository Returns: A tuple with the following members: - path relative to dir_path where the external should be exported - URL of the external to export - optional revision of the external to export - boolean indicating if the external URL is relative to the repository URL and targets a path not in the repository """ path = "" external_url = "" revision = None relative_url = False prev_part = None # turn multiple spaces into a single one and split on space for external_part in external.split(): if prev_part == "-r": # parse revision in the form "-r XXX" revision = int(external_part) elif external_part.startswith("-r") and external_part != "-r": # parse revision in the form "-rXXX" revision = int(external_part[2:]) elif external_part.startswith("^/"): # URL relative to the root of the repository in which the svn:externals # property is versioned external_url = svn_urljoin(repo_url, external_part[2:]) elif external_part.startswith("//"): # URL relative to the scheme of the URL of the directory on which the # svn:externals property is set scheme = urlparse(repo_url).scheme external_url = f"{scheme}:{external_part}" relative_url = not external_url.startswith(repo_url) elif external_part.startswith("/"): # URL relative to the root URL of the server on which the svn:externals # property is versioned parsed_url = urlparse(repo_url) root_url = f"{parsed_url.scheme}://{parsed_url.netloc}" external_url = svn_urljoin(root_url, external_part) relative_url = not external_url.startswith(repo_url) elif external_part.startswith("../"): # URL relative to the URL of the directory on which the svn:externals # property is set external_url = svn_urljoin(repo_url, dir_path, external_part) relative_url = not external_url.startswith(repo_url) elif re.match(r"^.*:*//.*", external_part): # absolute external URL external_url = external_part # subversion >= 1.6 added a quoting and escape mechanism to the syntax so # that the path of the external working copy may contain whitespace. elif external_part.startswith('\\"'): external_split = external.split('\\"') path = [ e.replace("\\ ", " ") for e in external_split if e.startswith(external_part[2:]) ][0] path = f'"{path}"' elif external_part.endswith('\\"'): continue elif external_part.startswith('"'): external_split = external.split('"') path_prefix = external_part.strip('"') path = next(iter([e for e in external_split if e.startswith(path_prefix)])) elif external_part.endswith('"'): continue elif not external_part.startswith("\\") and external_part != "-r": # path of the external relative to dir_path path = external_part.replace("\\\\", "\\") if path == external_part: path = external_part.replace("\\", "") path = path.lstrip("./") prev_part = external_part if "@" in external_url: # try to extract revision number if external URL is in the form # http://svn.example.org/repos/test/path@XXX url, revision_s = external_url.rsplit("@", maxsplit=1) try: revision = int(revision_s) external_url = url except ValueError: # handle URL like http://user@svn.example.org/ pass return (path.rstrip("/"), external_url, revision, relative_url) + + +def is_recursive_external( + origin_url: str, dir_path: str, external_path: str, external_url: str +) -> bool: + """ + Check if an external definition can lead to a recursive subversion export + operation (https://issues.apache.org/jira/browse/SVN-1703). + + Args: + origin_url: repository URL + dir_path: path of the directory where external is defined + external_path: path of the external relative to the directory + external_url: external URL + + Returns: + Whether the external definition is recursive + """ + parsed_origin_url = urlparse(origin_url) + parsed_external_url = urlparse(external_url) + external_url = urlunparse( + parsed_external_url._replace(scheme=parsed_origin_url.scheme) + ) + return svn_urljoin(origin_url, quote(dir_path), quote(external_path)).startswith( + external_url + )