This is an archive of the discontinued Mercurial Phabricator instance.

Differential D11545

dirstate-v2: Move data file info in the docket closer together
ClosedPublic

Authored by SimonSapin on Oct 1 2021, 9:41 AM.

Details

Reviewers
Alphare
Group Reviewers
hg-reviewers
Commits
rHGd467e44f71d7: dirstate-v2: Move data file info in the docket closer together
Summary

Having data_size next to uuid_size (and the UUID itself) makes more sense.

Diff Detail

Repository
rHG Mercurial
Lint
Automatic diff as part of commit; lint not applicable.
Unit
Automatic diff as part of commit; unit tests not applicable.

Revision Contents
Changeset List

Diff 30667

mercurial/dirstateutils/docket.py

# Must match the constant of the same name in # Must match the constant of the same name in
# `rust/hg-core/src/dirstate_tree/on_disk.rs` # `rust/hg-core/src/dirstate_tree/on_disk.rs`
TREE_METADATA_SIZE = 44 TREE_METADATA_SIZE = 44
# * 12 bytes: format marker # * 12 bytes: format marker
# * 32 bytes: node ID of the working directory's first parent # * 32 bytes: node ID of the working directory's first parent
# * 32 bytes: node ID of the working directory's second parent # * 32 bytes: node ID of the working directory's second parent
# * 4 bytes: big-endian used size of the data file
# * {TREE_METADATA_SIZE} bytes: tree metadata, parsed separately # * {TREE_METADATA_SIZE} bytes: tree metadata, parsed separately
# * 4 bytes: big-endian used size of the data file
# * 1 byte: length of the data file's UUID # * 1 byte: length of the data file's UUID
# * variable: data file's UUID # * variable: data file's UUID
# #
# Node IDs are null-padded if shorter than 32 bytes. # Node IDs are null-padded if shorter than 32 bytes.
# A data file shorter than the specified used size is corrupted (truncated) # A data file shorter than the specified used size is corrupted (truncated)
HEADER = struct.Struct( HEADER = struct.Struct(
">{}s32s32sL{}sB".format(len(V2_FORMAT_MARKER), TREE_METADATA_SIZE) ">{}s32s32s{}sLB".format(len(V2_FORMAT_MARKER), TREE_METADATA_SIZE)
) )
class DirstateDocket(object): class DirstateDocket(object):
data_filename_pattern = b'dirstate.%s' data_filename_pattern = b'dirstate.%s'
def __init__(self, parents, data_size, tree_metadata, uuid): def __init__(self, parents, data_size, tree_metadata, uuid):
self.parents = parents self.parents = parents
self.data_size = data_size self.data_size = data_size
self.tree_metadata = tree_metadata self.tree_metadata = tree_metadata
self.uuid = uuid self.uuid = uuid
@classmethod @classmethod
def with_new_uuid(cls, parents, data_size, tree_metadata): def with_new_uuid(cls, parents, data_size, tree_metadata):
return cls(parents, data_size, tree_metadata, docket_mod.make_uid()) return cls(parents, data_size, tree_metadata, docket_mod.make_uid())
@classmethod @classmethod
def parse(cls, data, nodeconstants): def parse(cls, data, nodeconstants):
if not data: if not data:
parents = (nodeconstants.nullid, nodeconstants.nullid) parents = (nodeconstants.nullid, nodeconstants.nullid)
return cls(parents, 0, b'', None) return cls(parents, 0, b'', None)
marker, p1, p2, data_size, meta, uuid_size = HEADER.unpack_from(data) marker, p1, p2, meta, data_size, uuid_size = HEADER.unpack_from(data)
if marker != V2_FORMAT_MARKER: if marker != V2_FORMAT_MARKER:
raise ValueError("expected dirstate-v2 marker") raise ValueError("expected dirstate-v2 marker")
uuid = data[HEADER.size : HEADER.size + uuid_size] uuid = data[HEADER.size : HEADER.size + uuid_size]
p1 = p1[: nodeconstants.nodelen] p1 = p1[: nodeconstants.nodelen]
p2 = p2[: nodeconstants.nodelen] p2 = p2[: nodeconstants.nodelen]
return cls((p1, p2), data_size, meta, uuid) return cls((p1, p2), data_size, meta, uuid)
def serialize(self): def serialize(self):
p1, p2 = self.parents p1, p2 = self.parents
header = HEADER.pack( header = HEADER.pack(
V2_FORMAT_MARKER, V2_FORMAT_MARKER,
p1, p1,
p2, p2,
self.data_size,
self.tree_metadata, self.tree_metadata,
self.data_size,
len(self.uuid), len(self.uuid),
) )
return header + self.uuid return header + self.uuid
def data_filename(self): def data_filename(self):
return self.data_filename_pattern % self.uuid return self.data_filename_pattern % self.uuid

rust/hg-core/src/dirstate_tree/on_disk.rs

// Must match `HEADER` in `mercurial/dirstateutils/docket.py` // Must match `HEADER` in `mercurial/dirstateutils/docket.py`
#[derive(BytesCast)] #[derive(BytesCast)]
#[repr(C)] #[repr(C)]
struct DocketHeader { struct DocketHeader {
marker: [u8; V2_FORMAT_MARKER.len()], marker: [u8; V2_FORMAT_MARKER.len()],
parent_1: [u8; STORED_NODE_ID_BYTES], parent_1: [u8; STORED_NODE_ID_BYTES],
parent_2: [u8; STORED_NODE_ID_BYTES], parent_2: [u8; STORED_NODE_ID_BYTES],
metadata: TreeMetadata,
/// Counted in bytes /// Counted in bytes
data_size: Size, data_size: Size,
metadata: TreeMetadata,
uuid_size: u8, uuid_size: u8,
} }
pub struct Docket<'on_disk> { pub struct Docket<'on_disk> {
header: &'on_disk DocketHeader, header: &'on_disk DocketHeader,
uuid: &'on_disk [u8], uuid: &'on_disk [u8],
} }