refactor(pacquet/cli): split cli_args.rs#12690
Merged
Merged
Conversation
The single cli_args.rs file had grown to ~1500 lines and mixed several unrelated concerns, making it a frequent source of merge conflicts. Split it into focused submodules under cli_args/, leaving cli_args.rs as a thin module index: - cli_command.rs: the CliArgs and CliCommand argument types - dispatch.rs: the impl CliArgs command-dispatch logic - pipelines.rs: the install/deploy/dedupe/prune reporter-generic pipelines - package_manager.rs: packageManager / devEngines parsing helpers - reporter.rs: ReporterType and its emit/configure helpers No behavior change; the move is mechanical. Subcommand modules that adding a new command still touches (the enum and the dispatch match) now live in their own files, so unrelated changes to pipelines or package-manager parsing no longer conflict with them. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01WLA4Nx1wDMQCaGzhcbXKVV
|
No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configurationConfiguration used: Path: .coderabbit.yaml Review profile: CHILL Plan: Pro Plus Run ID: 📒 Files selected for processing (3)
🚧 Files skipped from review as they are similar to previous changes (2)
📝 WalkthroughWalkthroughThe CLI argument surface was split into dedicated modules for command definitions, reporter setup, package-manager sync detection, pipelines, and dispatch handlers. The main run path now uses shared context, completion handling, install fast-path execution, and per-command install/query/script routing. ChangesCLI module split and orchestration
Estimated code review effort🎯 4 (Complex) | ⏱️ ~60 minutes Possibly related PRs
Suggested labels
Suggested reviewers
✨ Finishing Touches📝 Generate docstrings
🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
Contributor
Micro-Benchmark ResultsLinux |
There was a problem hiding this comment.
Actionable comments posted: 3
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@pacquet/crates/cli/src/cli_args/cli_command.rs`:
- Around line 108-109: Restore the install shortcut by ensuring
CliCommand::Install has the same alias/argv handling for "i" as before, so pnpm
i maps to install again. Update the CliCommand::Install variant in
cli_command.rs or the command parsing/rewrite logic that feeds it, and add or
adjust a test covering the pnpm i invocation so the shortcut behavior stays
locked in.
In `@pacquet/crates/cli/src/cli_args/package_manager.rs`:
- Around line 106-116: The packageManager parser in parse_package_manager
currently uses split_once('@'), which incorrectly breaks scoped packages like
`@pnpm/exe`@10.0.0 and prevents pnpm-matching behavior. Update
parse_package_manager to treat the last '@' as the version separator while
preserving the existing reference normalization logic (including the ':' guard
and '+' version trimming), so scoped names are parsed into the correct name and
optional reference.
In `@pacquet/crates/cli/src/cli_args/pipelines.rs`:
- Around line 196-213: The prune boundary check in pipelines::Config validation
is only lexical, so a relative escape like modules_dir pointing through ".." can
still bypass the workspace guard. Update the containment check around
cfg.modules_dir and config_root to first resolve/normalize the prune target
against the workspace root (or canonicalize both paths) before using
starts_with, so the final comparison in the prune pipeline reflects the real
filesystem location. Keep the safety behavior aligned with the existing prune
validation and the WorkspaceSettings::apply_to flow.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro Plus
Run ID: cf15d503-53e1-463f-a966-431db773ddf8
📒 Files selected for processing (7)
pacquet/crates/cli/src/cli_args.rspacquet/crates/cli/src/cli_args/cli_command.rspacquet/crates/cli/src/cli_args/dispatch.rspacquet/crates/cli/src/cli_args/package_manager.rspacquet/crates/cli/src/cli_args/pipelines.rspacquet/crates/cli/src/cli_args/reporter.rspacquet/crates/cli/src/cli_args/tests.rs
Codecov Report❌ Patch coverage is Additional details and impacted files@@ Coverage Diff @@
## main #12690 +/- ##
==========================================
+ Coverage 86.92% 86.96% +0.04%
==========================================
Files 376 382 +6
Lines 57967 58161 +194
==========================================
+ Hits 50386 50579 +193
- Misses 7581 7582 +1 ☔ View full report in Codecov by Harness. 🚀 New features to boost your workflow:
|
The dylint `perfectionist::import-granularity` lint (style = "crate") requires all imports sharing a crate root to be merged into a single `use` with nested braces. The split introduced several files with one `use super::x::Y;` per line, which the lint rejects. Collapse the top-level imports in the new cli_args submodules accordingly. This only affects module-level imports; function-body `use` statements are left as-is (the lint does not flag them).
Contributor
Integrated-Benchmark Report (Linux)Commit: Each scenario reports direct installs and pnpr installs. Bencher consumes pacquet@HEAD and pnpr@HEAD. Scenario: Isolated linker: fresh restore, cold cache + cold store
BENCHMARK_REPORT.json{
"results": [
{
"command": "pacquet@HEAD",
"mean": 4.447792795700001,
"stddev": 0.18275526009310783,
"median": 4.4216461875,
"user": 3.1564263200000005,
"system": 2.69355482,
"min": 4.2338434845,
"max": 4.7911472985,
"times": [
4.3387218275,
4.7037978015,
4.5060775225,
4.7911472985,
4.4837955705,
4.4156382415,
4.3253193685,
4.2338434845,
4.4276541335,
4.2519327085
]
},
{
"command": "pacquet@main",
"mean": 4.3573895379,
"stddev": 0.09534042186825586,
"median": 4.3450163565,
"user": 3.11221572,
"system": 2.68461512,
"min": 4.2308315875,
"max": 4.5113799225,
"times": [
4.3160126475,
4.2308315875,
4.4701038745,
4.5113799225,
4.4423555445,
4.2865208735,
4.2847877455,
4.3931828335,
4.2647002845,
4.3740200655
]
},
{
"command": "pnpr@HEAD",
"mean": 2.7723649518,
"stddev": 0.13926223788651496,
"median": 2.7459913030000003,
"user": 2.15608792,
"system": 2.29063022,
"min": 2.6339354155,
"max": 3.0777243005,
"times": [
2.8145266815000003,
2.7914335975,
3.0777243005,
2.6653991425,
2.9197551855,
2.6339354155,
2.7717508895000003,
2.6441885155000002,
2.6847040735,
2.7202317165
]
},
{
"command": "pnpr@main",
"mean": 2.8265720696,
"stddev": 0.13549471201373717,
"median": 2.819267559,
"user": 2.1518121199999998,
"system": 2.2817162200000003,
"min": 2.6605615465000003,
"max": 3.0340871155,
"times": [
3.0041397425,
2.7087355215,
2.6605615465000003,
2.8736339375,
2.8104075265,
2.6973736265,
2.7018305445,
3.0340871155,
2.8281275915000004,
2.9468235435
]
}
]
}Scenario: Isolated linker: fresh restore, hot cache + hot store
BENCHMARK_REPORT.json{
"results": [
{
"command": "pacquet@HEAD",
"mean": 0.66172196354,
"stddev": 0.059068447446870356,
"median": 0.62269658134,
"user": 0.31228500000000003,
"system": 0.9968055999999998,
"min": 0.60609254134,
"max": 0.73707992234,
"times": [
0.73707992234,
0.60609254134,
0.72451361534,
0.62453763034,
0.62085553234,
0.61578014234,
0.72983615134,
0.72875969234,
0.61434805034,
0.61541635734
]
},
{
"command": "pacquet@main",
"mean": 0.6887148209399998,
"stddev": 0.10662663409789597,
"median": 0.6243599773399999,
"user": 0.3102204,
"system": 1.0107624,
"min": 0.61410262434,
"max": 0.94536169334,
"times": [
0.62709467434,
0.74462745434,
0.74544247534,
0.73095496334,
0.62162528034,
0.61410262434,
0.62098911734,
0.61647735534,
0.62047257134,
0.94536169334
]
},
{
"command": "pnpr@HEAD",
"mean": 0.6746956512400001,
"stddev": 0.048602157268984675,
"median": 0.66214945534,
"user": 0.3194484,
"system": 1.0508707999999998,
"min": 0.64664181334,
"max": 0.81101163534,
"times": [
0.81101163534,
0.66986963234,
0.67129434934,
0.64664181334,
0.6554904903400001,
0.66221325034,
0.66208566034,
0.65577282334,
0.64803563934,
0.66454121834
]
},
{
"command": "pnpr@main",
"mean": 0.73674742634,
"stddev": 0.08288742324011034,
"median": 0.7213200578400001,
"user": 0.313863,
"system": 1.0716912,
"min": 0.65863031434,
"max": 0.8866616003400001,
"times": [
0.8866616003400001,
0.82566390534,
0.67314482034,
0.66166310434,
0.76949529534,
0.78337292134,
0.66652793534,
0.78033507834,
0.66197928834,
0.65863031434
]
}
]
}Scenario: Isolated linker: fresh install, cold cache + cold store
BENCHMARK_REPORT.json{
"results": [
{
"command": "pacquet@HEAD",
"mean": 4.372174682159999,
"stddev": 0.03621498662470635,
"median": 4.373118291259999,
"user": 3.0117563999999994,
"system": 2.57807866,
"min": 4.32097334876,
"max": 4.42067719076,
"times": [
4.42067719076,
4.4057405337599995,
4.369264525759999,
4.38537686676,
4.333578419759999,
4.4153380877599995,
4.37697205676,
4.32607851576,
4.32097334876,
4.367747275759999
]
},
{
"command": "pacquet@main",
"mean": 4.36711845886,
"stddev": 0.04774616200077435,
"median": 4.35254588676,
"user": 3.0169043999999996,
"system": 2.5885437599999994,
"min": 4.32497560476,
"max": 4.490834952759999,
"times": [
4.490834952759999,
4.3359203507599995,
4.34684501576,
4.3677389797599995,
4.32497560476,
4.37698405176,
4.3553096257599995,
4.33449009576,
4.38830376376,
4.34978214776
]
},
{
"command": "pnpr@HEAD",
"mean": 3.19014051926,
"stddev": 0.28640823314099706,
"median": 3.15395868476,
"user": 2.3306801999999998,
"system": 2.50195616,
"min": 2.83335520876,
"max": 3.64535346876,
"times": [
2.9028590287600005,
3.0310783327600004,
2.93856126376,
3.4106970977600004,
2.99284047176,
3.64535346876,
3.52902485276,
3.27683903676,
2.83335520876,
3.34079643076
]
},
{
"command": "pnpr@main",
"mean": 3.0107072500600003,
"stddev": 0.07808220919743038,
"median": 3.01484328026,
"user": 2.2531264,
"system": 2.4638980599999996,
"min": 2.8526531747600004,
"max": 3.11598033476,
"times": [
2.98396406776,
3.06202461576,
2.95517767176,
3.03958954376,
2.99009701676,
2.8526531747600004,
3.11598033476,
2.9577104237600005,
3.0695993027600004,
3.08027634876
]
}
]
}Scenario: Isolated linker: fresh install, hot cache + hot store
BENCHMARK_REPORT.json{
"results": [
{
"command": "pacquet@HEAD",
"mean": 1.2173982998800001,
"stddev": 0.046434694261724765,
"median": 1.2084169721800002,
"user": 1.0478673400000003,
"system": 1.3428523599999997,
"min": 1.1808719541800001,
"max": 1.34356061718,
"times": [
1.1894855561800002,
1.19495789418,
1.2058517911800002,
1.34356061718,
1.1918730691800001,
1.22258994818,
1.21993519618,
1.1808719541800001,
1.2138748191800002,
1.21098215318
]
},
{
"command": "pacquet@main",
"mean": 1.2360448218800002,
"stddev": 0.05567251701808611,
"median": 1.21891279518,
"user": 1.05147264,
"system": 1.35857646,
"min": 1.18936041118,
"max": 1.39003698518,
"times": [
1.21351817618,
1.21603725718,
1.39003698518,
1.2406332561800002,
1.2187585871800002,
1.2182780531800002,
1.2305520111800001,
1.2242064781800002,
1.18936041118,
1.2190670031800002
]
},
{
"command": "pnpr@HEAD",
"mean": 1.29377457118,
"stddev": 0.08683528164826192,
"median": 1.26880941868,
"user": 0.47109724000000003,
"system": 1.23323196,
"min": 1.24321835618,
"max": 1.53892552218,
"times": [
1.2769735611800002,
1.26401759118,
1.24321835618,
1.53892552218,
1.2851129671800001,
1.26241796418,
1.26965755218,
1.26910280918,
1.26851602818,
1.25980336018
]
},
{
"command": "pnpr@main",
"mean": 1.3027630579799998,
"stddev": 0.08178019920853473,
"median": 1.2762448276800002,
"user": 0.4703504399999999,
"system": 1.23528336,
"min": 1.2674990511800002,
"max": 1.53404241918,
"times": [
1.2674990511800002,
1.2759295021800001,
1.27034117818,
1.53404241918,
1.29655454018,
1.28443685418,
1.2676618161800002,
1.2765601531800002,
1.28428594918,
1.27031911618
]
}
]
}Scenario: Isolated linker: fresh install, cold cache + hot store
BENCHMARK_REPORT.json{
"results": [
{
"command": "pacquet@HEAD",
"mean": 2.73257122908,
"stddev": 0.09612468063140177,
"median": 2.69897953078,
"user": 1.42091592,
"system": 1.5559568,
"min": 2.67205262178,
"max": 2.98468350878,
"times": [
2.67205262178,
2.70893233178,
2.71415048678,
2.68902672978,
2.68236379778,
2.70973631378,
2.80181139178,
2.68798497878,
2.98468350878,
2.6749701297799997
]
},
{
"command": "pacquet@main",
"mean": 2.7883534810799993,
"stddev": 0.10582113157462848,
"median": 2.75880803828,
"user": 1.41869622,
"system": 1.5706883999999999,
"min": 2.68343498078,
"max": 3.04116522278,
"times": [
2.75327429478,
2.74231415778,
2.68741887178,
2.7643417817800002,
2.83474692478,
2.71252260278,
2.68343498078,
2.83395655878,
3.04116522278,
2.8303594147799997
]
},
{
"command": "pnpr@HEAD",
"mean": 1.26957674558,
"stddev": 0.013369905847460993,
"median": 1.2712781187799997,
"user": 0.47410682,
"system": 1.2271338999999999,
"min": 1.2364300317799999,
"max": 1.28345058678,
"times": [
1.2692539087799999,
1.28345058678,
1.2761994857799999,
1.2678293327799999,
1.2733023287799998,
1.26560749278,
1.27790496378,
1.2642467817799998,
1.2364300317799999,
1.28154254278
]
},
{
"command": "pnpr@main",
"mean": 1.29202259678,
"stddev": 0.09646948115966539,
"median": 1.2574064647799998,
"user": 0.48208912000000004,
"system": 1.2066268,
"min": 1.23469731678,
"max": 1.55800953978,
"times": [
1.2787816967799999,
1.24474720078,
1.2710480517799998,
1.55800953978,
1.31971029578,
1.26201427878,
1.25279865078,
1.23469731678,
1.2488660757799999,
1.24955286078
]
}
]
} |
Contributor
|
Bencher Report| Branch | pr/12690 |
| Testbed | pacquet |
Click to view all benchmark results
| Benchmark | Latency | Benchmark Result milliseconds (ms) (Result Δ%) | Upper Boundary milliseconds (ms) (Limit %) |
|---|---|---|---|
| isolated-linker.fresh-install.cold-cache.cold-store | 📈 view plot 🚷 view threshold | 4,372.17 ms(-8.11%)Baseline: 4,758.04 ms | 5,709.65 ms (76.58%) |
| isolated-linker.fresh-install.cold-cache.hot-store | 📈 view plot 🚷 view threshold | 2,732.57 ms(-10.41%)Baseline: 3,050.23 ms | 3,660.28 ms (74.65%) |
| isolated-linker.fresh-install.hot-cache.hot-store | 📈 view plot 🚷 view threshold | 1,217.40 ms(-10.39%)Baseline: 1,358.57 ms | 1,630.28 ms (74.67%) |
| isolated-linker.fresh-restore.cold-cache.cold-store | 📈 view plot 🚷 view threshold | 4,447.79 ms(-7.43%)Baseline: 4,805.00 ms | 5,766.00 ms (77.14%) |
| isolated-linker.fresh-restore.hot-cache.hot-store | 📈 view plot 🚷 view threshold | 661.72 ms(+0.91%)Baseline: 655.72 ms | 786.87 ms (84.10%) |
Contributor
|
| Branch | pr/12690 |
| Testbed | pnpr |
⚠️ WARNING: No Threshold found!Without a Threshold, no Alerts will ever be generated.
Click here to create a new Threshold
For more information, see the Threshold documentation.
To only post results if a Threshold exists, set the--ci-only-thresholdsflag.
Click to view all benchmark results
| Benchmark | Latency | milliseconds (ms) |
|---|---|---|
| isolated-linker.fresh-install.cold-cache.cold-store | 📈 view plot | 3,190.14 ms |
| isolated-linker.fresh-install.cold-cache.hot-store | 📈 view plot | 1,269.58 ms |
| isolated-linker.fresh-install.hot-cache.hot-store | 📈 view plot | 1,293.77 ms |
| isolated-linker.fresh-restore.cold-cache.cold-store | 📈 view plot | 2,772.36 ms |
| isolated-linker.fresh-restore.hot-cache.hot-store | 📈 view plot | 674.70 ms |
`dispatch.rs` was dominated by one ~600-line `match` over every `CliCommand` variant. Extract a `RunCtx` context (the canonicalized dir, the manifest path, the reporter, the `--recursive` flag, and the lazily loaded `config` / `state` closures as `&dyn Fn`) and move each command's body into one of three handler modules grouped by what the command does: - dispatch_install.rs: commands that mutate the install graph (add, update, remove, install, deploy, dedupe, prune, fetch, import, link, unlink, rebuild, patch*, dlx, create, runtime, approve-builds) - dispatch_query.rs: read-only queries (outdated, audit, list, why, whoami, dist-tag, ping, pack, root, config, pack-app, docs, store, cache, cat-file, cat-index, ignored-builds, find-hash) - dispatch_script.rs: package.json script / lifecycle commands (init, test, run, exec, start, stop, restart, set-script) `dispatch.rs` keeps only `RunCtx`, the three `CliArgs` methods, and a thin `route` match that wires each variant to its handler. No behavior change. The `config` / `state` closures are shared with each handler as `&dyn Fn(...) + Sync`; the `+ Sync` is required so the async handler futures that capture them stay `Send`, and it is mirrored onto `ApproveBuildsArgs::prepare`, which takes the same closures.
…ger parsing
Two pnpm-parity fixes surfaced by review of the cli_args split (both
pre-existing on the original cli_args.rs, not refactor regressions):
- Expose `i` as a visible alias of `install`, matching pnpm's
`commandNames = ['install', 'i']`, so `pacquet i` dispatches to install.
- Parse the `packageManager` field the way pnpm does: split on the first
`@`, except skip a leading scope `@` and split on the next one. The
previous `split_once('@')` mis-parsed scoped names like
`@scope/pnpm@1.0.0` into an empty name. The first `@` (not the last) is
used deliberately so a reference that is a URL containing `@` (e.g.
credentials) stays intact — mirroring pnpm's parsePackageManager.
Adds tests for the `i` alias and for parse_package_manager (unscoped,
scoped, `+`-hash, missing separator, and URL-with-`@` cases).
zkochan
approved these changes
Jun 27, 2026
KSXGitHub
pushed a commit
that referenced
this pull request
Jun 27, 2026
Brings in 3 commits from main: the cli_args.rs split (#12690, CLI-only), plus the bytes and tower-http dependency-version bumps. Only Cargo.lock conflicted; resolved by taking main's lockfile and re-adding this branch's network-web-auth subtree (crossterm, qrcode, open) via cargo. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01RWejcxTU8n144a1KK4nRj8
2 tasks
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
pacquet/crates/cli/src/cli_args.rshad grown to ~1500 lines and mixedseveral unrelated concerns in one file, which made it a frequent source of
merge conflicts. This PR splits it into focused submodules under
cli_args/, leavingcli_args.rsas a thin module index, and then furtherbreaks up the command-dispatch
match.1. Split
cli_args.rsinto focused submodulesThe
pub mod <subcommand>;declarations stay incli_args.rs(they definethe module paths for the existing
cli_args/<cmd>.rsfiles), but everythingelse moves out:
cli_args.rsCliArgsre-exportcli_args/cli_command.rsCliArgsstruct andCliCommandenumcli_args/dispatch.rsimpl CliArgs— the install fast-path, completion handling, and theroutedispatchercli_args/pipelines.rscli_args/package_manager.rspackageManager/devEnginesparsingcli_args/reporter.rsReporterTypeand itsemit/configurehelpers2. Split the dispatch
matchdispatch.rsnow holds aRunCtxcontext (the canonicalized dir, themanifest path, the reporter, the
--recursiveflag, and the lazily-loadedconfig/stateclosures) and a thinroutematch that wires each variantto a handler. The per-command bodies move into three modules grouped by what
the command does:
cli_args/dispatch_install.rscli_args/dispatch_query.rscli_args/dispatch_script.rspackage.jsonscript / lifecycle: init, test, run, exec, start, stop, restart, set-scriptdispatch.rsdrops from ~800 lines to ~280; thematchitself is now a~50-line router.
3. Two pnpm-parity fixes surfaced by review
Both were pre-existing in the original
cli_args.rs(not refactorregressions), found while reviewing the split:
iinstall alias — pnpm exposescommandNames = ['install', 'i'];pacquet lacked it. Added
#[clap(visible_alias = "i")]sopacquet idispatches to install.
packageManagerparsing —parse_package_managernow mirrorspnpm's parser (split on the first
@, skipping a leading scope@),instead of
split_once('@')which mis-parsed@scope/pnpm@1.0.0. Thefirst
@(not the last) is deliberate so a URL reference containing@stays intact.
Notes
substantive edits there are visibility widenings (to
pub(crate)/pub(super)), the import collapsing required by theperfectionist::import-granularitydylint lint (oneuseper crate root),and one signature change:
ApproveBuildsArgs::preparenow takes itsconfig/stateclosures as&(dyn Fn(...) + Sync)so the async handlerfutures that capture those closures stay
Send.pacquet-cli:cargo check,cargo clippy -- --deny warnings, andcargo fmt --checkare clean, and the lib tests(356, incl. the two new parity tests) pass. Import granularity was verified
with
rustfmt --config imports_granularity=Crate(and the CI Dylint job isgreen).
Squash Commit Body
Checklist
pacquet/port, or the description notes what still needs porting.The refactor is pacquet-internal (nothing to port). The two parity
fixes align pacquet with the existing TypeScript pnpm behavior, so no
TS-side change is needed.
The refactor is a mechanical move; the existing
cli_argstests movedwith the code. Added
install_command_parses_i_aliasandparse_package_manager_handles_unscoped_scoped_and_url_referencesfor thetwo parity fixes.
Written by an agent (Claude Code, Claude Opus 4.8).
🤖 Generated with Claude Code
https://claude.ai/code/session_01WLA4Nx1wDMQCaGzhcbXKVV
Summary by CodeRabbit
New Features
Bug Fixes