It's been more than 6 months since we added support for dynamically
selecting the working copy implementation. This patch drops support
for selecting the default implementation of that and other stores.
This helps to eliminate higher-ranked trait bounds from RevWalkRevset and
RevWalk combinators to be added. Since &CompositeIndex is now a real reference,
it can be passed to functions as index: &T.
The shortest change id prefix will become a few digits longer, but I think
that's acceptable. Entries included in the "revsets.short-prefixes" set are
unaffected.
The reachable set is calculated eagerly, but this is still faster as we no
longer need to sort the reachable entries by change id. The lazy version will
save another ~100ms in mid-size repos.
"jj log" without working copy snapshot:
```
% hyperfine --sort command --warmup 3 --runs 20 -L bin jj-0,jj-1,jj-2 \
-s "target/release-with-debug/{bin} -R ~/mirrors/linux debug reindex" \
"target/release-with-debug/{bin} -R ~/mirrors/linux \
--ignore-working-copy log -r.. -l100 --config-toml='revsets.short-prefixes=\"\"'"
Benchmark 1: target/release-with-debug/jj-0 -R ~/mirrors/linux --ignore-working-copy log -r.. -l100 --config-toml='revsets.short-prefixes=""'
Time (mean ± σ): 353.6 ms ± 11.9 ms [User: 266.7 ms, System: 87.0 ms]
Range (min … max): 329.0 ms … 365.6 ms 20 runs
Benchmark 2: target/release-with-debug/jj-1 -R ~/mirrors/linux --ignore-working-copy log -r.. -l100 --config-toml='revsets.short-prefixes=""'
Time (mean ± σ): 271.3 ms ± 9.9 ms [User: 183.8 ms, System: 87.7 ms]
Range (min … max): 250.5 ms … 282.7 ms 20 runs
Relative speed comparison
1.99 ± 0.16 target/release-with-debug/jj-0 -R ~/mirrors/linux --ignore-working-copy log -r.. -l100 --config-toml='revsets.short-prefixes=""'
1.53 ± 0.12 target/release-with-debug/jj-1 -R ~/mirrors/linux --ignore-working-copy log -r.. -l100 --config-toml='revsets.short-prefixes=""'
```
"jj status" with working copy snapshot (watchman enabled):
```
% hyperfine --sort command --warmup 3 --runs 20 -L bin jj-0,jj-1,jj-2 \
-s "target/release-with-debug/{bin} -R ~/mirrors/linux debug reindex" \
"target/release-with-debug/{bin} -R ~/mirrors/linux \
status --config-toml='revsets.short-prefixes=\"\"'"
Benchmark 1: target/release-with-debug/jj-0 -R ~/mirrors/linux status --config-toml='revsets.short-prefixes=""'
Time (mean ± σ): 396.6 ms ± 10.1 ms [User: 300.7 ms, System: 94.0 ms]
Range (min … max): 373.6 ms … 408.0 ms 20 runs
Benchmark 2: target/release-with-debug/jj-1 -R ~/mirrors/linux status --config-toml='revsets.short-prefixes=""'
Time (mean ± σ): 318.6 ms ± 12.6 ms [User: 219.1 ms, System: 94.1 ms]
Range (min … max): 294.2 ms … 333.0 ms 20 runs
Relative speed comparison
1.85 ± 0.14 target/release-with-debug/jj-0 -R ~/mirrors/linux status --config-toml='revsets.short-prefixes=""'
1.48 ± 0.12 target/release-with-debug/jj-1 -R ~/mirrors/linux status --config-toml='revsets.short-prefixes=""'
```
This basically means that the change ids are interned. We'll implement binary
search over the sorted change ids table. The table could be sorted differently
for better cache locality, but it is in lexicographical order for simplicity.
With my testing, the cost of the id lookup isn't dominant.
Unlike the parent entries, the size of the per-id overflow items isn't saved.
That's s because the number of the same-change-id commits is either 1 or many.
It doesn't make sense to allocate 8 bytes for each change id. Instead, we'll
pay extra indirection cost to determine the size.
Since IdIndex sorts the entries by using .sort_unstable_by_key(), the order of
the same-key elements is undefined. Perhaps, it's stable for short arrays, and
the test passes because of that.
I'm going to introduce breaking changes in index format. Some of them will
affect the file size, so version number or signature won't be needed. However,
I think it's safer to detect the format change as early as possible.
I have no idea if embedded version number is the best way. Because segment
files are looked up through the operation links, the version number could be
stored there and/or the "segments" directory could be versioned. If we want to
support multiple format versions and clients, it might be better to split the
tables into data chunks (e.g. graph entries, commit id table, change id table),
and add per-chunk version/type tag. I choose the per-file version just because
it's simple and would be non-controversial.
As I'm going to introduce format change pretty soon, this patch doesn't
implement data migration. The existing index files will be deleted and new
files will be created from scratch.
Planned index format changes include:
1. remove unused "flags" field
2. inline commit parents up to two
3. add sorted change ids table
Since hidden commits can be looked up by remote_branches() revset for example,
reindexing should traverse ancestors from all named refs in addition to the
visible heads.
change_id_index() is only used by Readonly/MutableRepo, so we don't need an
abstraction at Index. evaluate_revset() is somewhat similar, but the callers
rely on &dyn Repo.
We current have `Revset::change_id_index()` for creating a
`ChangeIdIndex` for a given revset. I think it will be hard to make it
performant for general revsets, especially in very large repos and
with custom index implementations, like the one we have at Google. If
we instead restrict it to including all ancestors of a set of heads, I
think it will be much easier to implement. We only use
`Revset::change_id_index()` with revsets including all visible commits
today, so we won't lose any current functionality by making it more
restricted.
I'm going to add a prefix resolution method to OpStore, but OpStore is
unrelated to the index. I think ObjectId, HexPrefix, and PrefixResolution can
be extracted to this module.
My jj repo contains such head commits, and "jj debug reindex" fails. To address
this problem, we'll probably need to implement GC, and the user will discard
operations before the first bad op id.
Spotted while adding error propagation there. This wouldn't likely be a real
problem because "jj debug reindex" removes all of the operation links.
The "} else {" condition is removed because it doesn't make sense to exclude
only the exact parent_op_id operation. This can be optimized to not walk
ancestors of the parent_op_id operation, but I don't see a motivation to add
tests covering such scenarios. It's pretty rare that an intermediate operation
link is missing.
It seems better to have the caller pass the transaction description
when we finish the transaction than when we start it. That way we have
all the information we want to include more readily available.
default_index_store.rs is relatively big, and it contains types and impls in
arbitrary order. Let's split them into sub modules. After everything moved,
mod.rs will only contain tests.
The wrapper type isn't needed for the mutable layer, but this mirrors the
readonly type structure. Test cases are also migrated to be using the index
wrapper so long as we don't have to care for the nesting of the segment files.
I don't think the backend should matter for any of these tests, so
let's test with only one, and let's make that the strictest one - the
new test backend.
This reduces the number of tests by 74 (from 974 to 900), but saves no
measurable run time.
It makes the call sites clearer if we pass the `TestRepoBackend` enum
instead of the boolean `use_git` value. It's also more extensible (I
plan to add another backend for tests).
Almost everyone calls the project "jj", and there seeems to be
consensus that we should rename the crates. I originally wanted the
crates to be called `jj` and `jj-lib`, but `jj` was already
taken. `jj-cli` is probably at least as good for it anyway.
Once we've published a 0.8.0 under the new names, we'll release 0.7.1
versions under the old names with pointers to the new crates names.
This gets rid of round-trip conversion from queries like "(main..)-". I have
such expression in my default log/disambiguation revset, and the query could
take ~150ms to convert head positions back and forth if the repository had
tons of unmerged commits.
We can't get rid of the other "impl Index"es because .as_composite() must
return a real reference type. Maybe we could turn CompositeIndex into an
owned wrapper, but I don't know if that would be worth the effort.
It might sound scary to add public .mutable_index() accessor, but I think
it's okay because immutable MutableIndex reference has no more power than
Index.
This allows us to implement Index for lifetime-bound type such as
CompositeIndex<'_>.
The idea is that .as_composite() is equivalent to .as_index(), but for the
implementation type. I'm going to add "impl Index for CompositeIndex" to
clean up index references passed to revset engine.
It no longer needs to be on the `Index` trait, thereby removing the
last direct use of `IndexEntry` in the trait (it's still used
indirectly in `walk_revs()`).
This is another step towards allowing a custom `jj` binary to have its
own index type. We're going to have a server-backed index
implementation at Google, for example.
I'm about to make `RepoLoader::init()` return a `Result`, and I don't
want to have to wrap that in a new error in
`ReadonlyRepo::load_at_head()` since that's only used in tests.
This is just a little preparation for extracting a `Repo` trait that's
implemented by both `ReadonlyRepo` and `MutableRepo`. The `index()`
function in that trait will of course have to return the same type in
both implementations, and that type will be `&dyn Index`.
Even though we don't know the details yet, we know that we want to
make the index pluggable like the commit and opstore
backends. Defining a trait for it should be a good step. We can refine
the trait later.
We already have `create_random_commit()`, which returns a
`CommitBuilder`. Most callers directly write that to a
`MutableRepo`. That currently returns a `Commit`, but I'm about to
make it propagate errors from the backend. That would add an
`unwrap()` to this sequence, making it longer. Let's create a simple
helper for these callers to simplify this common pattern.
When you're done with the `CommitBuilder`, you're going to have to
call `write_to_repo()`, passing it a mutable `MutableRepo`
reference. It's a bit simpler to pass that reference when we create
the `CommitBuilder` instead, so that's what this patch does.
A drawback of passing in the mutable reference when we create the
builder is that we can't have multiple unfinished `CommitBuilder`
instance live at the same time. We don't have any such use cases yet,
and it's not hard to work around them, so I think this change is worth
it.
This will be a building block of 'parents(base)' revset. 'base---' will
be .filter_by_generation(3..4) for example. I think 'ancestors(base)' can
also have an optional generation parameter, but I haven't considered any
particular syntax yet.
Let's acknowledge everyone's contributions by replacing "Google LLC"
in the copyright header by "The Jujutsu Authors". If I understand
correctly, it won't have any legal effect, but maybe it still helps
reduce concerns from contributors (though I haven't heard any
concerns).
Google employees can read about Google's policy at
go/releasing/contributions#copyright.
We currently get the hostname and username from the `whoami` crate. We
do that in lib crate, without giving the caller a way to override
them. That seems wrong since it might be used in a server and
performing operations on behalf of some other user. This commit makes
the hostname and username configurable, so the calling crate can pass
them in. If they have not been passed in, we still default to the
values from the `whoami` crate.
These calls often appear in expressions long enough that not having to
qualify it means that we can sometimes avoid wrapping a line. I
noticed because IntelliJ told me that `test_git.rs` had some
unnecessary qualificiations (the function was already imported there).
The `testutils` module should ideally not be part of the library
dependencies. Since they're used by the integration tests (and the CLI
tests), we need to move them to a separate crate to achieve that.
