Expand description
A bundle of pack data and the corresponding pack index
Fields§
§pack: File
The pack file corresponding to index
index: File
The index file corresponding to pack
Implementations§
source§impl Bundle
impl Bundle
Initialization
sourcepub fn at(path: impl AsRef<Path>, object_hash: Kind) -> Result<Self, Error>
pub fn at(path: impl AsRef<Path>, object_hash: Kind) -> Result<Self, Error>
Create a Bundle
from path
, which is either a pack file (*.pack) or an index file (*.idx).
The corresponding complementary file is expected to be present.
The object_hash
is a way to read (and write) the same file format with different hashes, as the hash kind
isn’t stored within the file format itself.
Examples found in repository?
More examples
117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312
fn verify_integrity_inner<C, P, F>(
&self,
mut progress: P,
should_interrupt: &AtomicBool,
deep_check: bool,
options: index::verify::integrity::Options<F>,
) -> Result<integrity::Outcome<P>, index::traverse::Error<integrity::Error>>
where
P: Progress,
C: crate::cache::DecodeEntry,
F: Fn() -> C + Send + Clone,
{
let parent = self.path.parent().expect("must be in a directory");
let actual_index_checksum = self
.verify_checksum(
progress.add_child_with_id(format!("{}: checksum", self.path.display()), *b"MVCK"), /* Multiindex Verify ChecKsum */
should_interrupt,
)
.map_err(integrity::Error::from)
.map_err(index::traverse::Error::Processor)?;
if let Some(first_invalid) = crate::verify::fan(&self.fan) {
return Err(index::traverse::Error::Processor(integrity::Error::Fan {
index: first_invalid,
}));
}
if self.num_objects == 0 {
return Err(index::traverse::Error::Processor(integrity::Error::Empty));
}
let mut pack_traverse_statistics = Vec::new();
let operation_start = Instant::now();
let mut total_objects_checked = 0;
let mut pack_ids_and_offsets = Vec::with_capacity(self.num_objects as usize);
{
let order_start = Instant::now();
let mut progress = progress.add_child_with_id("checking oid order", *b"MVOR"); /* Multiindex Verify Oid oRder */
progress.init(
Some(self.num_objects as usize),
git_features::progress::count("objects"),
);
for entry_index in 0..(self.num_objects - 1) {
let lhs = self.oid_at_index(entry_index);
let rhs = self.oid_at_index(entry_index + 1);
if rhs.cmp(lhs) != Ordering::Greater {
return Err(index::traverse::Error::Processor(integrity::Error::OutOfOrder {
index: entry_index,
}));
}
let (pack_id, _) = self.pack_id_and_pack_offset_at_index(entry_index);
pack_ids_and_offsets.push((pack_id, entry_index));
progress.inc();
}
{
let entry_index = self.num_objects - 1;
let (pack_id, _) = self.pack_id_and_pack_offset_at_index(entry_index);
pack_ids_and_offsets.push((pack_id, entry_index));
}
// sort by pack-id to allow handling all indices matching a pack while its open.
pack_ids_and_offsets.sort_by(|l, r| l.0.cmp(&r.0));
progress.show_throughput(order_start);
};
progress.init(
Some(self.num_indices as usize),
git_features::progress::count("indices"),
);
let mut pack_ids_slice = pack_ids_and_offsets.as_slice();
for (pack_id, index_file_name) in self.index_names.iter().enumerate() {
progress.set_name(index_file_name.display().to_string());
progress.inc();
let mut bundle = None;
let index;
let index_path = parent.join(index_file_name);
let index = if deep_check {
bundle = crate::Bundle::at(index_path, self.object_hash)
.map_err(integrity::Error::from)
.map_err(index::traverse::Error::Processor)?
.into();
bundle.as_ref().map(|b| &b.index).expect("just set")
} else {
index = Some(
index::File::at(index_path, self.object_hash)
.map_err(|err| integrity::Error::BundleInit(crate::bundle::init::Error::Index(err)))
.map_err(index::traverse::Error::Processor)?,
);
index.as_ref().expect("just set")
};
let slice_end = pack_ids_slice.partition_point(|e| e.0 == pack_id as crate::data::Id);
let multi_index_entries_to_check = &pack_ids_slice[..slice_end];
{
let offset_start = Instant::now();
let mut offsets_progress = progress.add_child_with_id("verify object offsets", *b"MVOF"); /* Multiindex Verify Object Offsets */
offsets_progress.init(
Some(pack_ids_and_offsets.len()),
git_features::progress::count("objects"),
);
pack_ids_slice = &pack_ids_slice[slice_end..];
for entry_id in multi_index_entries_to_check.iter().map(|e| e.1) {
let oid = self.oid_at_index(entry_id);
let (_, expected_pack_offset) = self.pack_id_and_pack_offset_at_index(entry_id);
let entry_in_bundle_index = index.lookup(oid).ok_or_else(|| {
index::traverse::Error::Processor(integrity::Error::OidNotFound { id: oid.to_owned() })
})?;
let actual_pack_offset = index.pack_offset_at_index(entry_in_bundle_index);
if actual_pack_offset != expected_pack_offset {
return Err(index::traverse::Error::Processor(
integrity::Error::PackOffsetMismatch {
id: oid.to_owned(),
expected_pack_offset,
actual_pack_offset,
},
));
}
offsets_progress.inc();
}
if should_interrupt.load(std::sync::atomic::Ordering::Relaxed) {
return Err(index::traverse::Error::Processor(integrity::Error::Interrupted));
}
offsets_progress.show_throughput(offset_start);
}
total_objects_checked += multi_index_entries_to_check.len();
if let Some(bundle) = bundle {
progress.set_name(format!("Validating {}", index_file_name.display()));
let crate::bundle::verify::integrity::Outcome {
actual_index_checksum: _,
pack_traverse_outcome,
progress: returned_progress,
} = bundle
.verify_integrity(progress, should_interrupt, options.clone())
.map_err(|err| {
use index::traverse::Error::*;
match err {
Processor(err) => Processor(integrity::Error::IndexIntegrity(err)),
VerifyChecksum(err) => VerifyChecksum(err),
Tree(err) => Tree(err),
TreeTraversal(err) => TreeTraversal(err),
PackDecode { id, offset, source } => PackDecode { id, offset, source },
PackMismatch { expected, actual } => PackMismatch { expected, actual },
PackObjectMismatch {
expected,
actual,
offset,
kind,
} => PackObjectMismatch {
expected,
actual,
offset,
kind,
},
Crc32Mismatch {
expected,
actual,
offset,
kind,
} => Crc32Mismatch {
expected,
actual,
offset,
kind,
},
Interrupted => Interrupted,
}
})?;
progress = returned_progress;
pack_traverse_statistics.push(pack_traverse_outcome);
}
}
assert_eq!(
self.num_objects as usize, total_objects_checked,
"BUG: our slicing should allow to visit all objects"
);
progress.set_name("Validating multi-pack");
progress.show_throughput(operation_start);
Ok(integrity::Outcome {
actual_index_checksum,
pack_traverse_statistics,
progress,
})
}
source§impl Bundle
impl Bundle
sourcepub fn find<'a>(
&self,
id: impl AsRef<oid>,
out: &'a mut Vec<u8>,
cache: &mut impl DecodeEntry
) -> Result<Option<(Data<'a>, Location)>, Error>
pub fn find<'a>(
&self,
id: impl AsRef<oid>,
out: &'a mut Vec<u8>,
cache: &mut impl DecodeEntry
) -> Result<Option<(Data<'a>, Location)>, Error>
sourcepub fn get_object_by_index<'a>(
&self,
idx: u32,
out: &'a mut Vec<u8>,
cache: &mut impl DecodeEntry
) -> Result<(Data<'a>, Location), Error>
pub fn get_object_by_index<'a>(
&self,
idx: u32,
out: &'a mut Vec<u8>,
cache: &mut impl DecodeEntry
) -> Result<(Data<'a>, Location), Error>
Special-use function to get an object given an index previously returned from internal_find_pack_index.
Panics
If index
is out of bounds.
Examples found in repository?
8 9 10 11 12 13 14 15 16 17 18 19
pub fn find<'a>(
&self,
id: impl AsRef<git_hash::oid>,
out: &'a mut Vec<u8>,
cache: &mut impl crate::cache::DecodeEntry,
) -> Result<Option<(git_object::Data<'a>, crate::data::entry::Location)>, crate::data::decode::Error> {
let idx = match self.index.lookup(id) {
Some(idx) => idx,
None => return Ok(None),
};
self.get_object_by_index(idx, out, cache).map(Some)
}
source§impl Bundle
impl Bundle
sourcepub fn write_to_directory<P>(
pack: impl BufRead,
directory: Option<impl AsRef<Path>>,
progress: P,
should_interrupt: &AtomicBool,
thin_pack_base_object_lookup_fn: Option<Box<dyn for<'a> FnMut(ObjectId, &'a mut Vec<u8>) -> Option<Data<'a>>>>,
options: Options
) -> Result<Outcome, Error>where
P: Progress,
pub fn write_to_directory<P>(
pack: impl BufRead,
directory: Option<impl AsRef<Path>>,
progress: P,
should_interrupt: &AtomicBool,
thin_pack_base_object_lookup_fn: Option<Box<dyn for<'a> FnMut(ObjectId, &'a mut Vec<u8>) -> Option<Data<'a>>>>,
options: Options
) -> Result<Outcome, Error>where
P: Progress,
Given a pack
data stream, write it along with a generated index into the directory
if Some
or discard all output if None
.
In the latter case, the functionality provided here is more akind of pack data stream validation.
progress
provides detailed progress information which can be discarded with [git_features::progress::Discard
].should_interrupt
is checked regularly and when true, the whole operation will stop.thin_pack_base_object_lookup_fn
If set, we expect to see a thin-pack with objects that reference their base object by object id which is expected to exist in the object database the bundle is contained within.options
further configure how the task is performed.
Note
- the resulting pack may be empty, that is, contains zero objects in some situations. This is a valid reply by a server and should
be accounted for.
- Empty packs always have the same name and not handling this case will result in at most one superfluous pack.
sourcepub fn write_to_directory_eagerly<P>(
pack: impl Read + Send + 'static,
pack_size: Option<u64>,
directory: Option<impl AsRef<Path>>,
progress: P,
should_interrupt: &'static AtomicBool,
thin_pack_base_object_lookup_fn: Option<Box<dyn for<'a> FnMut(ObjectId, &'a mut Vec<u8>) -> Option<Data<'a>> + Send + 'static>>,
options: Options
) -> Result<Outcome, Error>where
P: Progress,
P::SubProgress: 'static,
pub fn write_to_directory_eagerly<P>(
pack: impl Read + Send + 'static,
pack_size: Option<u64>,
directory: Option<impl AsRef<Path>>,
progress: P,
should_interrupt: &'static AtomicBool,
thin_pack_base_object_lookup_fn: Option<Box<dyn for<'a> FnMut(ObjectId, &'a mut Vec<u8>) -> Option<Data<'a>> + Send + 'static>>,
options: Options
) -> Result<Outcome, Error>where
P: Progress,
P::SubProgress: 'static,
Equivalent to write_to_directory()
but offloads reading of the pack into its own thread, hence the Send + 'static'
bounds.
Note
As it sends portions of the input to a thread it requires the ’static lifetime for the interrupt flags. This can only be satisfied by a static AtomicBool which is only suitable for programs that only run one of these operations at a time or don’t mind that all of them abort when the flag is set.
source§impl Bundle
impl Bundle
sourcepub fn verify_integrity<C, P, F>(
&self,
progress: P,
should_interrupt: &AtomicBool,
options: Options<F>
) -> Result<Outcome<P>, Error<Error>>where
P: Progress,
C: DecodeEntry,
F: Fn() -> C + Send + Clone,
pub fn verify_integrity<C, P, F>(
&self,
progress: P,
should_interrupt: &AtomicBool,
options: Options<F>
) -> Result<Outcome<P>, Error<Error>>where
P: Progress,
C: DecodeEntry,
F: Fn() -> C + Send + Clone,
Similar to crate::index::File::verify_integrity()
but more convenient to call as the presence of the
pack file is a given.
Examples found in repository?
117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312
fn verify_integrity_inner<C, P, F>(
&self,
mut progress: P,
should_interrupt: &AtomicBool,
deep_check: bool,
options: index::verify::integrity::Options<F>,
) -> Result<integrity::Outcome<P>, index::traverse::Error<integrity::Error>>
where
P: Progress,
C: crate::cache::DecodeEntry,
F: Fn() -> C + Send + Clone,
{
let parent = self.path.parent().expect("must be in a directory");
let actual_index_checksum = self
.verify_checksum(
progress.add_child_with_id(format!("{}: checksum", self.path.display()), *b"MVCK"), /* Multiindex Verify ChecKsum */
should_interrupt,
)
.map_err(integrity::Error::from)
.map_err(index::traverse::Error::Processor)?;
if let Some(first_invalid) = crate::verify::fan(&self.fan) {
return Err(index::traverse::Error::Processor(integrity::Error::Fan {
index: first_invalid,
}));
}
if self.num_objects == 0 {
return Err(index::traverse::Error::Processor(integrity::Error::Empty));
}
let mut pack_traverse_statistics = Vec::new();
let operation_start = Instant::now();
let mut total_objects_checked = 0;
let mut pack_ids_and_offsets = Vec::with_capacity(self.num_objects as usize);
{
let order_start = Instant::now();
let mut progress = progress.add_child_with_id("checking oid order", *b"MVOR"); /* Multiindex Verify Oid oRder */
progress.init(
Some(self.num_objects as usize),
git_features::progress::count("objects"),
);
for entry_index in 0..(self.num_objects - 1) {
let lhs = self.oid_at_index(entry_index);
let rhs = self.oid_at_index(entry_index + 1);
if rhs.cmp(lhs) != Ordering::Greater {
return Err(index::traverse::Error::Processor(integrity::Error::OutOfOrder {
index: entry_index,
}));
}
let (pack_id, _) = self.pack_id_and_pack_offset_at_index(entry_index);
pack_ids_and_offsets.push((pack_id, entry_index));
progress.inc();
}
{
let entry_index = self.num_objects - 1;
let (pack_id, _) = self.pack_id_and_pack_offset_at_index(entry_index);
pack_ids_and_offsets.push((pack_id, entry_index));
}
// sort by pack-id to allow handling all indices matching a pack while its open.
pack_ids_and_offsets.sort_by(|l, r| l.0.cmp(&r.0));
progress.show_throughput(order_start);
};
progress.init(
Some(self.num_indices as usize),
git_features::progress::count("indices"),
);
let mut pack_ids_slice = pack_ids_and_offsets.as_slice();
for (pack_id, index_file_name) in self.index_names.iter().enumerate() {
progress.set_name(index_file_name.display().to_string());
progress.inc();
let mut bundle = None;
let index;
let index_path = parent.join(index_file_name);
let index = if deep_check {
bundle = crate::Bundle::at(index_path, self.object_hash)
.map_err(integrity::Error::from)
.map_err(index::traverse::Error::Processor)?
.into();
bundle.as_ref().map(|b| &b.index).expect("just set")
} else {
index = Some(
index::File::at(index_path, self.object_hash)
.map_err(|err| integrity::Error::BundleInit(crate::bundle::init::Error::Index(err)))
.map_err(index::traverse::Error::Processor)?,
);
index.as_ref().expect("just set")
};
let slice_end = pack_ids_slice.partition_point(|e| e.0 == pack_id as crate::data::Id);
let multi_index_entries_to_check = &pack_ids_slice[..slice_end];
{
let offset_start = Instant::now();
let mut offsets_progress = progress.add_child_with_id("verify object offsets", *b"MVOF"); /* Multiindex Verify Object Offsets */
offsets_progress.init(
Some(pack_ids_and_offsets.len()),
git_features::progress::count("objects"),
);
pack_ids_slice = &pack_ids_slice[slice_end..];
for entry_id in multi_index_entries_to_check.iter().map(|e| e.1) {
let oid = self.oid_at_index(entry_id);
let (_, expected_pack_offset) = self.pack_id_and_pack_offset_at_index(entry_id);
let entry_in_bundle_index = index.lookup(oid).ok_or_else(|| {
index::traverse::Error::Processor(integrity::Error::OidNotFound { id: oid.to_owned() })
})?;
let actual_pack_offset = index.pack_offset_at_index(entry_in_bundle_index);
if actual_pack_offset != expected_pack_offset {
return Err(index::traverse::Error::Processor(
integrity::Error::PackOffsetMismatch {
id: oid.to_owned(),
expected_pack_offset,
actual_pack_offset,
},
));
}
offsets_progress.inc();
}
if should_interrupt.load(std::sync::atomic::Ordering::Relaxed) {
return Err(index::traverse::Error::Processor(integrity::Error::Interrupted));
}
offsets_progress.show_throughput(offset_start);
}
total_objects_checked += multi_index_entries_to_check.len();
if let Some(bundle) = bundle {
progress.set_name(format!("Validating {}", index_file_name.display()));
let crate::bundle::verify::integrity::Outcome {
actual_index_checksum: _,
pack_traverse_outcome,
progress: returned_progress,
} = bundle
.verify_integrity(progress, should_interrupt, options.clone())
.map_err(|err| {
use index::traverse::Error::*;
match err {
Processor(err) => Processor(integrity::Error::IndexIntegrity(err)),
VerifyChecksum(err) => VerifyChecksum(err),
Tree(err) => Tree(err),
TreeTraversal(err) => TreeTraversal(err),
PackDecode { id, offset, source } => PackDecode { id, offset, source },
PackMismatch { expected, actual } => PackMismatch { expected, actual },
PackObjectMismatch {
expected,
actual,
offset,
kind,
} => PackObjectMismatch {
expected,
actual,
offset,
kind,
},
Crc32Mismatch {
expected,
actual,
offset,
kind,
} => Crc32Mismatch {
expected,
actual,
offset,
kind,
},
Interrupted => Interrupted,
}
})?;
progress = returned_progress;
pack_traverse_statistics.push(pack_traverse_outcome);
}
}
assert_eq!(
self.num_objects as usize, total_objects_checked,
"BUG: our slicing should allow to visit all objects"
);
progress.set_name("Validating multi-pack");
progress.show_throughput(operation_start);
Ok(integrity::Outcome {
actual_index_checksum,
pack_traverse_statistics,
progress,
})
}