Expand description
Write a pack data header at version with num_objects and return a buffer.
Examples found in repository?
src/data/input/entries_to_bytes.rs (line 71)
69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115
fn next_inner(&mut self, entry: input::Entry) -> Result<input::Entry, input::Error> {
if self.num_entries == 0 {
let header_bytes = crate::data::header::encode(self.data_version, 0);
self.output.write_all(&header_bytes[..])?;
}
self.num_entries += 1;
entry.header.write_to(entry.decompressed_size, &mut self.output)?;
std::io::copy(
&mut entry
.compressed
.as_deref()
.expect("caller must configure generator to keep compressed bytes"),
&mut self.output,
)?;
Ok(entry)
}
fn write_header_and_digest(&mut self, last_entry: Option<&mut input::Entry>) -> Result<(), input::Error> {
let header_bytes = crate::data::header::encode(self.data_version, self.num_entries);
let num_bytes_written = if last_entry.is_some() {
self.output.stream_position()?
} else {
header_bytes.len() as u64
};
self.output.seek(std::io::SeekFrom::Start(0))?;
self.output.write_all(&header_bytes[..])?;
self.output.flush()?;
self.output.seek(std::io::SeekFrom::Start(0))?;
let interrupt_never = std::sync::atomic::AtomicBool::new(false);
let digest = hash::bytes(
&mut self.output,
num_bytes_written as usize,
self.object_hash,
&mut git_features::progress::Discard,
&interrupt_never,
)?;
self.output.write_all(digest.as_slice())?;
self.output.flush()?;
self.is_done = true;
if let Some(last_entry) = last_entry {
last_entry.trailer = Some(digest);
}
self.trailer = Some(digest);
Ok(())
}More examples
src/data/output/bytes.rs (line 99)
96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132
fn next_inner(&mut self) -> Result<u64, Error<E>> {
let previous_written = self.written;
if let Some((version, num_entries)) = self.header_info.take() {
let header_bytes = crate::data::header::encode(version, num_entries);
self.output.write_all(&header_bytes[..])?;
self.written += header_bytes.len() as u64;
}
match self.input.next() {
Some(entries) => {
for entry in entries.map_err(Error::Input)? {
if entry.is_invalid() {
self.pack_offsets_and_validity.push((0, false));
continue;
};
self.pack_offsets_and_validity.push((self.written, true));
let header = entry.to_entry_header(self.entry_version, |index| {
let (base_offset, is_valid_object) = self.pack_offsets_and_validity[index];
if !is_valid_object {
unreachable!("if you see this the object database is correct as a delta refers to a non-existing object")
}
self.written - base_offset
});
self.written += header.write_to(entry.decompressed_size as u64, &mut self.output)? as u64;
self.written += std::io::copy(&mut &*entry.compressed_data, &mut self.output)?;
}
}
None => {
let digest = self.output.hash.clone().digest();
self.output.inner.write_all(&digest[..])?;
self.written += digest.len() as u64;
self.output.inner.flush()?;
self.is_done = true;
self.trailer = Some(git_hash::ObjectId::from(digest));
}
};
Ok(self.written - previous_written)
}src/index/write/mod.rs (line 194)
53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220
pub fn write_data_iter_to_stream<F, F2>(
version: crate::index::Version,
make_resolver: F,
entries: impl Iterator<Item = Result<crate::data::input::Entry, crate::data::input::Error>>,
thread_limit: Option<usize>,
mut root_progress: impl Progress,
out: impl io::Write,
should_interrupt: &AtomicBool,
object_hash: git_hash::Kind,
pack_version: crate::data::Version,
) -> Result<Outcome, Error>
where
F: FnOnce() -> io::Result<F2>,
F2: for<'r> Fn(crate::data::EntryRange, &'r mut Vec<u8>) -> Option<()> + Send + Clone,
{
if version != crate::index::Version::default() {
return Err(Error::Unsupported(version));
}
let mut num_objects: usize = 0;
let mut last_seen_trailer = None;
let anticipated_num_objects = entries.size_hint().1.unwrap_or_else(|| entries.size_hint().0);
let mut tree = Tree::with_capacity(anticipated_num_objects)?;
let indexing_start = std::time::Instant::now();
root_progress.init(Some(4), progress::steps());
let mut objects_progress = root_progress.add_child_with_id("indexing", *b"IWIO"); /* Index Write Index Objects */
objects_progress.init(entries.size_hint().1, progress::count("objects"));
let mut decompressed_progress = root_progress.add_child_with_id("decompressing", *b"IWDB"); /* Index Write Decompressed Bytes */
decompressed_progress.init(None, progress::bytes());
let mut pack_entries_end: u64 = 0;
for entry in entries {
let crate::data::input::Entry {
header,
pack_offset,
crc32,
header_size,
compressed: _,
compressed_size,
decompressed_size,
trailer,
} = entry?;
decompressed_progress.inc_by(decompressed_size as usize);
let entry_len = header_size as u64 + compressed_size;
pack_entries_end = pack_offset + entry_len;
let crc32 = crc32.expect("crc32 to be computed by the iterator. Caller assures correct configuration.");
use crate::data::entry::Header::*;
match header {
Tree | Blob | Commit | Tag => {
tree.add_root(
pack_offset,
TreeEntry {
id: object_hash.null(),
crc32,
},
)?;
}
RefDelta { .. } => return Err(Error::IteratorInvariantNoRefDelta),
OfsDelta { base_distance } => {
let base_pack_offset =
crate::data::entry::Header::verified_base_pack_offset(pack_offset, base_distance).ok_or(
Error::IteratorInvariantBaseOffset {
pack_offset,
distance: base_distance,
},
)?;
tree.add_child(
base_pack_offset,
pack_offset,
TreeEntry {
id: object_hash.null(),
crc32,
},
)?;
}
};
last_seen_trailer = trailer;
num_objects += 1;
objects_progress.inc();
}
if num_objects != anticipated_num_objects {
objects_progress.info(format!(
"{} objects were resolved into {} objects during thin-pack resolution",
anticipated_num_objects, num_objects
));
}
let num_objects: u32 = num_objects
.try_into()
.map_err(|_| Error::IteratorInvariantTooManyObjects(num_objects))?;
objects_progress.show_throughput(indexing_start);
decompressed_progress.show_throughput(indexing_start);
drop(objects_progress);
drop(decompressed_progress);
root_progress.inc();
let resolver = make_resolver()?;
let sorted_pack_offsets_by_oid = {
let traverse::Outcome { roots, children } = tree.traverse(
resolver,
pack_entries_end,
|| (),
|data,
_progress,
traverse::Context {
entry,
decompressed: bytes,
..
}| {
modify_base(data, entry, bytes, version.hash());
Ok::<_, Error>(())
},
traverse::Options {
object_progress: root_progress.add_child_with_id("Resolving", *b"IWRO"), /* Index Write Resolve Objects */
size_progress: root_progress.add_child_with_id("Decoding", *b"IWDB"), /* Index Write Decode Bytes */
thread_limit,
should_interrupt,
object_hash,
},
)?;
root_progress.inc();
let mut items = roots;
items.extend(children);
{
let _progress = root_progress.add_child_with_id("sorting by id", *b"info");
items.sort_by_key(|e| e.data.id);
}
root_progress.inc();
items
};
let pack_hash = match last_seen_trailer {
Some(ph) => ph,
None if num_objects == 0 => {
let header = crate::data::header::encode(pack_version, 0);
let mut hasher = git_features::hash::hasher(object_hash);
hasher.update(&header);
git_hash::ObjectId::from(hasher.digest())
}
None => return Err(Error::IteratorInvariantTrailer),
};
let index_hash = encode::write_to(
out,
sorted_pack_offsets_by_oid,
&pack_hash,
version,
root_progress.add_child_with_id("writing index file", *b"IWBW"), /* Index Write Bytes Written */
)?;
root_progress.show_throughput_with(
indexing_start,
num_objects as usize,
progress::count("objects").expect("unit always set"),
progress::MessageLevel::Success,
);
Ok(Outcome {
index_version: version,
index_hash,
data_hash: pack_hash,
num_objects,
})
}