Skip to content

core: cut codex-core compile time 57% by type-erasing ToolHandler#16627

Closed
bolinfest wants to merge 1 commit intomainfrom
pr16627
Closed

core: cut codex-core compile time 57% by type-erasing ToolHandler#16627
bolinfest wants to merge 1 commit intomainfrom
pr16627

Conversation

@bolinfest
Copy link
Copy Markdown
Collaborator

@bolinfest bolinfest commented Apr 2, 2026 •
edited
Loading

Why

codex-core compile time was dominated by repeated trait solving and monomorphization work from the old ToolHandler<Output = T> shape plus the blanket async adapter into AnyToolHandler.

That pattern instantiated one async wrapper per concrete tool handler, and the compiler then had to prove Send/outlives obligations across a large set of generated future types. The profile showed this was not an LLVM/codegen problem; it was a trait-solving + monomorphization problem.

The win from this PR is large enough that I want the numbers in the PR body, because this is a broad API change and reviewers are right to ask whether the tradeoff is worth it.

Headline result

Cold codex-core lib compile time dropped from 187.15s to 79.90s on my machine after this refactor: -57.3% wall-clock in rustc's total pass.

The biggest hot passes dropped by roughly 70%:

Metric Before After Delta
total 187.15s 79.90s -57.3%
generate_crate_metadata 84.53s 25.88s -69.4%
MIR_borrow_checking 84.13s 25.69s -69.5%
monomorphization_collector_graph_walk 79.74s 23.54s -70.5%
evaluate_obligation self-time 180.62s 43.29s -76.0%

Important caveat: -Z time-passes timings are nested, so generate_crate_metadata and monomorphization_collector_graph_walk are mostly overlapping, not additive.

Profile data

Baseline before this change

cargo +nightly build -p codex-core --lib -Z unstable-options --timings=json after cargo clean -p codex-core:

Crate duration rmeta_time
codex-core 187.380776583s 174.474507208s
starlark 17.90s n/a

So codex-core was far and away the dominant crate in the build.

cargo +nightly rustc -p codex-core --lib -- -Z time-passes -Z time-passes-format=json after cargo clean -p codex-core:

Pass Time
total 187.150662083s
generate_crate_metadata 84.531864625s
MIR_borrow_checking 84.131389375s
monomorphization_collector_graph_walk 79.737515042s
codegen_crate 12.362532292s
type_check_crate 4.4765405s
coherence_checking 3.311121208s
process real / user / sys 187.70s / 201.87s / 4.99s

-Z self-profile + measureme summarize -p 0.5 before this change:

Query / phase Self time Total time % total CPU Item count Cache hits
evaluate_obligation 180.62s 182.08s 70.821% 572,234 1,130,998
mir_borrowck 1.42s 93.77s n/a n/a n/a
typeck 1.84s 2.38s n/a n/a n/a
LLVM_module_codegen_emit_obj n/a 17.01s n/a n/a n/a
LLVM_passes n/a 12.95s n/a n/a n/a
codegen_module n/a 12.22s n/a n/a n/a
self-profile CPU total 255.042999997s n/a n/a n/a n/a
process real / user / sys 220.96s / 235.02s / 7.09s n/a n/a n/a n/a

Top normalized obligation buckets before this change:

Obligation bucket Samples Duration
outlives:tasks::review::ReviewTask 1,067 6.33s
outlives:tools::handlers::unified_exec::UnifiedExecHandler 896 5.63s
trait:T as tools::registry::ToolHandler 876 5.45s
outlives:tools::handlers::shell::ShellHandler 888 5.37s
outlives:tools::handlers::shell::ShellCommandHandler 870 5.29s
outlives:tools::runtimes::shell::unix_escalation::CoreShellActionProvider 637 3.73s
outlives:tools::handlers::mcp::McpHandler 695 3.61s
outlives:tasks::regular::RegularTask 726 3.57s

Top items_of_instance entries before this change were mostly concrete async handler/task impls:

Instance Duration
tasks::regular::{impl#2}::run 3.79s
tools::handlers::mcp::{impl#0}::handle 3.27s
tools::runtimes::shell::unix_escalation::{impl#2}::determine_action 3.09s
tools::handlers::agent_jobs::{impl#11}::handle 3.07s
tools::handlers::multi_agents::spawn::{impl#1}::handle 2.84s
tasks::review::{impl#4}::run 2.82s
tools::handlers::multi_agents_v2::spawn::{impl#2}::handle 2.80s
tools::handlers::multi_agents::resume_agent::{impl#1}::handle 2.73s
tools::handlers::unified_exec::{impl#2}::handle 2.54s
tasks::compact::{impl#4}::run 2.45s

After this change

cargo +nightly rustc -p codex-core --lib -- -Z time-passes -Z time-passes-format=json after cargo clean -p codex-core:

Pass Time
total 79.896377542s
generate_crate_metadata 25.882309084s
MIR_borrow_checking 25.694140042s
monomorphization_collector_graph_walk 23.542859708s
codegen_crate 12.729979125s
process real / user / sys 81.06s / 87.82s / 2.76s

-Z self-profile + measureme summarize -p 0.5 after this change:

Query / phase Self time Total time % total CPU Item count Cache hits
evaluate_obligation 43.29s 44.61s 33.319% 366,783 1,056,994
mir_borrowck 1.42s 27.77s n/a n/a n/a
typeck 1.83s 2.36s n/a n/a n/a
type_op_prove_predicate 886.60ms 23.38s n/a n/a n/a
LLVM_module_codegen_emit_obj n/a 17.88s n/a n/a n/a
LLVM_passes n/a 17.33s n/a n/a n/a
codegen_module n/a 16.07s n/a n/a n/a
self-profile CPU total 129.915927377s n/a n/a n/a n/a
process real / user / sys 90.10s / 101.70s / 2.68s n/a n/a n/a n/a

Artifact size deltas from the self-profile summaries:

Artifact Before After Delta
crate_metadata 26,534,471 26,475,240 -59,231
dep_graph 253,181,425 236,801,785 -16,379,640
linked_artifact 565,366,624 557,429,424 -7,937,200
object_file 513,127,264 505,350,688 -7,776,576
query_cache 137,440,945 136,793,279 -647,666
cgu_instructions 3,586,307 3,565,670 -20,637
codegen_unit_size_estimate 2,084,846 2,073,774 -11,072
work_product_index 19,565 19,565 0

What changed

The main refactor is in codex-rs/core/src/tools/registry.rs:

  • ToolHandler is now object-safe and returns BoxFuture<'_, Result<AnyToolResult, FunctionCallError>> directly.
  • ToolRegistry stores Arc<dyn ToolHandler> directly.
  • AnyToolResult now wraps Box<dyn ToolOutput> plus the invocation metadata.
  • The blanket generic impl<T: ToolHandler> AnyToolHandler for T adapter is gone, which is the point: stop generating one generic async adapter per concrete handler type.

Representative handler updates:

Tradeoff

This intentionally moves a bit of work to runtime: each tool call now boxes the returned future and boxes the concrete ToolOutput behind AnyToolResult.

That is a reasonable tradeoff here because these handlers are typically dominated by process spawning, MCP I/O, agent orchestration, or filesystem work, while the compile-time savings are very large and directly address the crate's current scaling problem.

Validation

  • cargo test -p codex-core --lib --no-run passed
  • cargo test -p codex-core --lib ran 1,541 tests: 1,533 passed, 3 ignored, 5 failed
  • The 5 failures were all in untouched config tests in codex-rs/core/src/config/config_tests.rs, so I do not believe they are caused by this refactor:
    • config::tests::approvals_reviewer_defaults_to_manual_only_without_guardian_feature
    • config::tests::approvals_reviewer_stays_manual_only_when_guardian_feature_is_enabled
    • config::tests::approvals_reviewer_can_be_set_in_config_without_guardian_approval
    • config::tests::smart_approvals_alias_is_ignored
    • config::tests::smart_approvals_alias_is_ignored_in_profiles
  • just fix -p codex-core
  • just fmt

I intentionally did not rerun tests after just fix / just fmt, per the repo instructions in this checkout.

@bolinfest bolinfest changed the title core: type-erase ToolHandler outputs core: cut codex-core compile time 57% by type-erasing ToolHandler Apr 2, 2026
@bolinfest bolinfest mentioned this pull request Apr 2, 2026
Merged
bolinfest added a commit that referenced this pull request Apr 2, 2026
@bolinfest
core: cut codex-core compile time 63% with native async ToolHandler (#…
3c7f013 
…16630)

## Why

`ToolHandler` was still paying a large compile-time tax from
`#[async_trait]` on every concrete handler impl, even though the only
object-safe boundary the registry actually stores is the internal
`AnyToolHandler` adapter.

This PR removes that macro-generated async wrapper layer from concrete
`ToolHandler` impls while keeping the existing object-safe shim in
`AnyToolHandler`. In practice, that gets essentially the same
compile-time win as the larger type-erasure refactor in #16627, but with
a much smaller diff and without changing the public shape of
`ToolHandler<Output = T>`.

That tradeoff matters here because this is a broad `codex-core` hotspot
and reviewers should be able to judge the compile-time impact from hard
numbers, not vibes.

## Headline result

On a clean `codex-core` package rebuild (`cargo clean -p codex-core`
before each command), rustc `total` dropped from **187.15s to 68.98s**
versus the shared `0bd31dc382bd` baseline: **-63.1%**.

The biggest hot passes dropped by roughly **71-72%**:

| Metric | Baseline `0bd31dc382bd` | This PR `41f7ac0adeac` | Delta |
|---|---:|---:|---:|
| `total` | 187.15s | 68.98s | **-63.1%** |
| `generate_crate_metadata` | 84.53s | 24.49s | **-71.0%** |
| `MIR_borrow_checking` | 84.13s | 24.58s | **-70.8%** |
| `monomorphization_collector_graph_walk` | 79.74s | 22.19s | **-72.2%**
|
| `evaluate_obligation` self-time | 180.62s | 46.91s | **-74.0%** |

Important caveat: `-Z time-passes` timings are nested, so
`generate_crate_metadata` and `monomorphization_collector_graph_walk`
are mostly overlapping, not additive.

## Why this PR over #16627

#16627 already proved that the `ToolHandler` stack was the right
hotspot, but it got there by making `ToolHandler` object-safe and
changing every handler to return `BoxFuture<Result<AnyToolResult, _>>`
directly.

This PR keeps the lower-churn shape:

- `ToolHandler` remains generic over `type Output`.
- Concrete handlers use native RPITIT futures with explicit `Send`
bounds.
- `AnyToolHandler` remains the only object-safe adapter and still does
the boxing at the registry boundary, as before.
- The implementation diff is only **33 files, +28/-77**.

The measurements are at least comparable, and in this run this PR is
slightly faster than #16627 on the pass-level total:

| Metric | #16627 | This PR | Delta |
|---|---:|---:|---:|
| `total` | 79.90s | 68.98s | **-13.7%** |
| `generate_crate_metadata` | 25.88s | 24.49s | **-5.4%** |
| `monomorphization_collector_graph_walk` | 23.54s | 22.19s | **-5.7%**
|
| `evaluate_obligation` self-time | 43.29s | 46.91s | +8.4% |

## Profile data

### Crate-level timings

`cargo +nightly build -p codex-core --lib -Z unstable-options
--timings=json` after `cargo clean -p codex-core`.

Baseline data below is reused from the shared parent `0bd31dc382bd`
profile because this PR and #16627 are both one commit on top of that
same parent.

| Crate | Baseline `duration` | This PR `duration` | Delta | Baseline
`rmeta_time` | This PR `rmeta_time` | Delta |
|---|---:|---:|---:|---:|---:|---:|
| `codex_core` | 187.380776583s | 69.171113833s | **-63.1%** |
174.474507208s | 55.873015583s | **-68.0%** |
| `starlark` | 17.90s | 16.773824125s | -6.3% | n/a | 8.8999965s | n/a |

### Pass-level timings

`cargo +nightly rustc -p codex-core --lib -- -Z time-passes -Z
time-passes-format=json` after `cargo clean -p codex-core`.

| Pass | Baseline | This PR | Delta |
|---|---:|---:|---:|
| `total` | 187.150662083s | 68.978770375s | **-63.1%** |
| `generate_crate_metadata` | 84.531864625s | 24.487462958s | **-71.0%**
|
| `MIR_borrow_checking` | 84.131389375s | 24.575553875s | **-70.8%** |
| `monomorphization_collector_graph_walk` | 79.737515042s |
22.190207417s | **-72.2%** |
| `codegen_crate` | 12.362532292s | 12.695237625s | +2.7% |
| `type_check_crate` | 4.4765405s | 5.442019542s | +21.6% |
| `coherence_checking` | 3.311121208s | 4.239935292s | +28.0% |
| process `real` / `user` / `sys` | 187.70s / 201.87s / 4.99s | 69.52s /
85.90s / 2.92s | n/a |

### Self-profile query summary

`cargo +nightly rustc -p codex-core --lib -- -Z self-profile=... -Z
self-profile-events=default,query-keys,args,llvm,artifact-sizes` after
`cargo clean -p codex-core`, summarized with `measureme summarize -p
0.5`.

| Query / phase | Baseline self time | This PR self time | Delta |
Baseline total time | This PR total time | Baseline item count | This PR
item count | Baseline cache hits | This PR cache hits |
|---|---:|---:|---:|---:|---:|---:|---:|---:|---:|
| `evaluate_obligation` | 180.62s | 46.91s | **-74.0%** | 182.08s |
48.37s | 572,234 | 388,659 | 1,130,998 | 1,058,553 |
| `mir_borrowck` | 1.42s | 1.49s | +4.9% | 93.77s | 29.59s | n/a | 6,184
| n/a | 15,298 |
| `typeck` | 1.84s | 1.87s | +1.6% | 2.38s | 2.44s | n/a | 9,367 | n/a |
79,247 |
| `LLVM_module_codegen_emit_obj` | n/a | 17.12s | n/a | 17.01s | 17.12s
| n/a | 256 | n/a | 0 |
| `LLVM_passes` | n/a | 13.07s | n/a | 12.95s | 13.07s | n/a | 1 | n/a |
0 |
| `codegen_module` | n/a | 12.33s | n/a | 12.22s | 13.64s | n/a | 256 |
n/a | 0 |
| `items_of_instance` | n/a | 676.00ms | n/a | n/a | 24.96s | n/a |
99,990 | n/a | 0 |
| `type_op_prove_predicate` | n/a | 660.79ms | n/a | n/a | 24.78s | n/a
| 78,762 | n/a | 235,877 |

| Summary | Baseline | This PR |
|---|---:|---:|
| `evaluate_obligation` % of total CPU | 70.821% | 38.880% |
| self-profile total CPU time | 255.042999997s | 120.661175956s |
| process `real` / `user` / `sys` | 220.96s / 235.02s / 7.09s | 86.35s /
103.66s / 3.54s |

### Artifact sizes

From the same `measureme summarize` output:

| Artifact | Baseline | This PR | Delta |
|---|---:|---:|---:|
| `crate_metadata` | 26,534,471 bytes | 26,545,248 bytes | +10,777 |
| `dep_graph` | 253,181,425 bytes | 239,240,806 bytes | -13,940,619 |
| `linked_artifact` | 565,366,624 bytes | 562,673,176 bytes | -2,693,448
|
| `object_file` | 513,127,264 bytes | 510,464,096 bytes | -2,663,168 |
| `query_cache` | 137,440,945 bytes | 136,982,566 bytes | -458,379 |
| `cgu_instructions` | 3,586,307 bytes | 3,575,121 bytes | -11,186 |
| `codegen_unit_size_estimate` | 2,084,846 bytes | 2,078,773 bytes |
-6,073 |
| `work_product_index` | 19,565 bytes | 19,565 bytes | 0 |

### Baseline hotspots before this change

These are the top normalized obligation buckets from the shared baseline
profile:

| Obligation bucket | Samples | Duration |
|---|---:|---:|
| `outlives:tasks::review::ReviewTask` | 1,067 | 6.33s |
| `outlives:tools::handlers::unified_exec::UnifiedExecHandler` | 896 |
5.63s |
| `trait:T as tools::registry::ToolHandler` | 876 | 5.45s |
| `outlives:tools::handlers::shell::ShellHandler` | 888 | 5.37s |
| `outlives:tools::handlers::shell::ShellCommandHandler` | 870 | 5.29s |
|
`outlives:tools::runtimes::shell::unix_escalation::CoreShellActionProvider`
| 637 | 3.73s |
| `outlives:tools::handlers::mcp::McpHandler` | 695 | 3.61s |
| `outlives:tasks::regular::RegularTask` | 726 | 3.57s |

Top `items_of_instance` entries before this change were mostly concrete
async handler/task impls:

| Instance | Duration |
|---|---:|
| `tasks::regular::{impl#2}::run` | 3.79s |
| `tools::handlers::mcp::{impl#0}::handle` | 3.27s |
| `tools::runtimes::shell::unix_escalation::{impl#2}::determine_action`
| 3.09s |
| `tools::handlers::agent_jobs::{impl#11}::handle` | 3.07s |
| `tools::handlers::multi_agents::spawn::{impl#1}::handle` | 2.84s |
| `tasks::review::{impl#4}::run` | 2.82s |
| `tools::handlers::multi_agents_v2::spawn::{impl#2}::handle` | 2.80s |
| `tools::handlers::multi_agents::resume_agent::{impl#1}::handle` |
2.73s |
| `tools::handlers::unified_exec::{impl#2}::handle` | 2.54s |
| `tasks::compact::{impl#4}::run` | 2.45s |

## What changed

Relevant pre-change registry shape:
[`codex-rs/core/src/tools/registry.rs`](https://github.com/openai/codex/blob/0bd31dc382bd1c33dc2bb6b97069c76aa10ba14b/codex-rs/core/src/tools/registry.rs#L38-L219)

Current registry shape in this PR:
[`codex-rs/core/src/tools/registry.rs`](https://github.com/openai/codex/blob/41f7ac0adeac81d667541853d6546267d6083613/codex-rs/core/src/tools/registry.rs#L38-L203)

- `ToolHandler::{is_mutating, handle}` now return native `impl Future +
Send` futures instead of using `#[async_trait]`.
- `AnyToolHandler` remains the object-safe adapter and boxes those
futures at the registry boundary with explicit lifetimes.
- Concrete handlers and the registry test handler drop `#[async_trait]`
but otherwise keep their async method bodies intact.
- Representative examples:
[`codex-rs/core/src/tools/handlers/shell.rs`](https://github.com/openai/codex/blob/41f7ac0adeac81d667541853d6546267d6083613/codex-rs/core/src/tools/handlers/shell.rs#L223-L379),
[`codex-rs/core/src/tools/handlers/unified_exec.rs`](https://github.com/openai/codex/blob/41f7ac0adeac81d667541853d6546267d6083613/codex-rs/core/src/tools/handlers/unified_exec.rs),
[`codex-rs/core/src/tools/registry_tests.rs`](https://github.com/openai/codex/blob/41f7ac0adeac81d667541853d6546267d6083613/codex-rs/core/src/tools/registry_tests.rs)

## Tradeoff

This is intentionally less invasive than #16627: it does **not** move
result boxing into every concrete handler and does **not** change
`ToolHandler` into an object-safe trait.

Instead, it keeps the existing registry-level type-erasure boundary and
only removes the macro-generated async wrapper layer from concrete
impls. So the runtime boxing story stays basically the same as before,
while the compile-time savings are still large.

## Verification

Existing verification for this branch still applies:

- Ran `cargo test -p codex-core`; this change compiled and the suite
reached the known unrelated `config::tests::*guardian*` failures, with
no local diff under `codex-rs/core/src/config/`.

Profiling commands used for the tables above:

- `cargo clean -p codex-core`
- `cargo +nightly build -p codex-core --lib -Z unstable-options
--timings=json`
- `cargo +nightly rustc -p codex-core --lib -- -Z time-passes -Z
time-passes-format=json`
- `cargo +nightly rustc -p codex-core --lib -- -Z self-profile=... -Z
self-profile-events=default,query-keys,args,llvm,artifact-sizes`
- `measureme summarize -p 0.5`
@bolinfest
Copy link
Copy Markdown
Collaborator Author

bolinfest commented Apr 2, 2026

Closing in favor of #16630.

@bolinfest bolinfest closed this Apr 2, 2026
@github-actions github-actions bot mentioned this pull request Apr 3, 2026
Open
Oreoxp pushed a commit to Oreoxp/codex-cli that referenced this pull request Apr 6, 2026
@bolinfest @Oreoxp
core: cut codex-core compile time 63% with native async ToolHandler (o…
96cb34f 
…penai#16630)

## Why

`ToolHandler` was still paying a large compile-time tax from
`#[async_trait]` on every concrete handler impl, even though the only
object-safe boundary the registry actually stores is the internal
`AnyToolHandler` adapter.

This PR removes that macro-generated async wrapper layer from concrete
`ToolHandler` impls while keeping the existing object-safe shim in
`AnyToolHandler`. In practice, that gets essentially the same
compile-time win as the larger type-erasure refactor in openai#16627, but with
a much smaller diff and without changing the public shape of
`ToolHandler<Output = T>`.

That tradeoff matters here because this is a broad `codex-core` hotspot
and reviewers should be able to judge the compile-time impact from hard
numbers, not vibes.

## Headline result

On a clean `codex-core` package rebuild (`cargo clean -p codex-core`
before each command), rustc `total` dropped from **187.15s to 68.98s**
versus the shared `0bd31dc382bd` baseline: **-63.1%**.

The biggest hot passes dropped by roughly **71-72%**:

| Metric | Baseline `0bd31dc382bd` | This PR `41f7ac0adeac` | Delta |
|---|---:|---:|---:|
| `total` | 187.15s | 68.98s | **-63.1%** |
| `generate_crate_metadata` | 84.53s | 24.49s | **-71.0%** |
| `MIR_borrow_checking` | 84.13s | 24.58s | **-70.8%** |
| `monomorphization_collector_graph_walk` | 79.74s | 22.19s | **-72.2%**
|
| `evaluate_obligation` self-time | 180.62s | 46.91s | **-74.0%** |

Important caveat: `-Z time-passes` timings are nested, so
`generate_crate_metadata` and `monomorphization_collector_graph_walk`
are mostly overlapping, not additive.

## Why this PR over openai#16627

openai#16627 already proved that the `ToolHandler` stack was the right
hotspot, but it got there by making `ToolHandler` object-safe and
changing every handler to return `BoxFuture<Result<AnyToolResult, _>>`
directly.

This PR keeps the lower-churn shape:

- `ToolHandler` remains generic over `type Output`.
- Concrete handlers use native RPITIT futures with explicit `Send`
bounds.
- `AnyToolHandler` remains the only object-safe adapter and still does
the boxing at the registry boundary, as before.
- The implementation diff is only **33 files, +28/-77**.

The measurements are at least comparable, and in this run this PR is
slightly faster than openai#16627 on the pass-level total:

| Metric | openai#16627 | This PR | Delta |
|---|---:|---:|---:|
| `total` | 79.90s | 68.98s | **-13.7%** |
| `generate_crate_metadata` | 25.88s | 24.49s | **-5.4%** |
| `monomorphization_collector_graph_walk` | 23.54s | 22.19s | **-5.7%**
|
| `evaluate_obligation` self-time | 43.29s | 46.91s | +8.4% |

## Profile data

### Crate-level timings

`cargo +nightly build -p codex-core --lib -Z unstable-options
--timings=json` after `cargo clean -p codex-core`.

Baseline data below is reused from the shared parent `0bd31dc382bd`
profile because this PR and openai#16627 are both one commit on top of that
same parent.

| Crate | Baseline `duration` | This PR `duration` | Delta | Baseline
`rmeta_time` | This PR `rmeta_time` | Delta |
|---|---:|---:|---:|---:|---:|---:|
| `codex_core` | 187.380776583s | 69.171113833s | **-63.1%** |
174.474507208s | 55.873015583s | **-68.0%** |
| `starlark` | 17.90s | 16.773824125s | -6.3% | n/a | 8.8999965s | n/a |

### Pass-level timings

`cargo +nightly rustc -p codex-core --lib -- -Z time-passes -Z
time-passes-format=json` after `cargo clean -p codex-core`.

| Pass | Baseline | This PR | Delta |
|---|---:|---:|---:|
| `total` | 187.150662083s | 68.978770375s | **-63.1%** |
| `generate_crate_metadata` | 84.531864625s | 24.487462958s | **-71.0%**
|
| `MIR_borrow_checking` | 84.131389375s | 24.575553875s | **-70.8%** |
| `monomorphization_collector_graph_walk` | 79.737515042s |
22.190207417s | **-72.2%** |
| `codegen_crate` | 12.362532292s | 12.695237625s | +2.7% |
| `type_check_crate` | 4.4765405s | 5.442019542s | +21.6% |
| `coherence_checking` | 3.311121208s | 4.239935292s | +28.0% |
| process `real` / `user` / `sys` | 187.70s / 201.87s / 4.99s | 69.52s /
85.90s / 2.92s | n/a |

### Self-profile query summary

`cargo +nightly rustc -p codex-core --lib -- -Z self-profile=... -Z
self-profile-events=default,query-keys,args,llvm,artifact-sizes` after
`cargo clean -p codex-core`, summarized with `measureme summarize -p
0.5`.

| Query / phase | Baseline self time | This PR self time | Delta |
Baseline total time | This PR total time | Baseline item count | This PR
item count | Baseline cache hits | This PR cache hits |
|---|---:|---:|---:|---:|---:|---:|---:|---:|---:|
| `evaluate_obligation` | 180.62s | 46.91s | **-74.0%** | 182.08s |
48.37s | 572,234 | 388,659 | 1,130,998 | 1,058,553 |
| `mir_borrowck` | 1.42s | 1.49s | +4.9% | 93.77s | 29.59s | n/a | 6,184
| n/a | 15,298 |
| `typeck` | 1.84s | 1.87s | +1.6% | 2.38s | 2.44s | n/a | 9,367 | n/a |
79,247 |
| `LLVM_module_codegen_emit_obj` | n/a | 17.12s | n/a | 17.01s | 17.12s
| n/a | 256 | n/a | 0 |
| `LLVM_passes` | n/a | 13.07s | n/a | 12.95s | 13.07s | n/a | 1 | n/a |
0 |
| `codegen_module` | n/a | 12.33s | n/a | 12.22s | 13.64s | n/a | 256 |
n/a | 0 |
| `items_of_instance` | n/a | 676.00ms | n/a | n/a | 24.96s | n/a |
99,990 | n/a | 0 |
| `type_op_prove_predicate` | n/a | 660.79ms | n/a | n/a | 24.78s | n/a
| 78,762 | n/a | 235,877 |

| Summary | Baseline | This PR |
|---|---:|---:|
| `evaluate_obligation` % of total CPU | 70.821% | 38.880% |
| self-profile total CPU time | 255.042999997s | 120.661175956s |
| process `real` / `user` / `sys` | 220.96s / 235.02s / 7.09s | 86.35s /
103.66s / 3.54s |

### Artifact sizes

From the same `measureme summarize` output:

| Artifact | Baseline | This PR | Delta |
|---|---:|---:|---:|
| `crate_metadata` | 26,534,471 bytes | 26,545,248 bytes | +10,777 |
| `dep_graph` | 253,181,425 bytes | 239,240,806 bytes | -13,940,619 |
| `linked_artifact` | 565,366,624 bytes | 562,673,176 bytes | -2,693,448
|
| `object_file` | 513,127,264 bytes | 510,464,096 bytes | -2,663,168 |
| `query_cache` | 137,440,945 bytes | 136,982,566 bytes | -458,379 |
| `cgu_instructions` | 3,586,307 bytes | 3,575,121 bytes | -11,186 |
| `codegen_unit_size_estimate` | 2,084,846 bytes | 2,078,773 bytes |
-6,073 |
| `work_product_index` | 19,565 bytes | 19,565 bytes | 0 |

### Baseline hotspots before this change

These are the top normalized obligation buckets from the shared baseline
profile:

| Obligation bucket | Samples | Duration |
|---|---:|---:|
| `outlives:tasks::review::ReviewTask` | 1,067 | 6.33s |
| `outlives:tools::handlers::unified_exec::UnifiedExecHandler` | 896 |
5.63s |
| `trait:T as tools::registry::ToolHandler` | 876 | 5.45s |
| `outlives:tools::handlers::shell::ShellHandler` | 888 | 5.37s |
| `outlives:tools::handlers::shell::ShellCommandHandler` | 870 | 5.29s |
|
`outlives:tools::runtimes::shell::unix_escalation::CoreShellActionProvider`
| 637 | 3.73s |
| `outlives:tools::handlers::mcp::McpHandler` | 695 | 3.61s |
| `outlives:tasks::regular::RegularTask` | 726 | 3.57s |

Top `items_of_instance` entries before this change were mostly concrete
async handler/task impls:

| Instance | Duration |
|---|---:|
| `tasks::regular::{impl#2}::run` | 3.79s |
| `tools::handlers::mcp::{impl#0}::handle` | 3.27s |
| `tools::runtimes::shell::unix_escalation::{impl#2}::determine_action`
| 3.09s |
| `tools::handlers::agent_jobs::{impl#11}::handle` | 3.07s |
| `tools::handlers::multi_agents::spawn::{impl#1}::handle` | 2.84s |
| `tasks::review::{impl#4}::run` | 2.82s |
| `tools::handlers::multi_agents_v2::spawn::{impl#2}::handle` | 2.80s |
| `tools::handlers::multi_agents::resume_agent::{impl#1}::handle` |
2.73s |
| `tools::handlers::unified_exec::{impl#2}::handle` | 2.54s |
| `tasks::compact::{impl#4}::run` | 2.45s |

## What changed

Relevant pre-change registry shape:
[`codex-rs/core/src/tools/registry.rs`](https://github.com/openai/codex/blob/0bd31dc382bd1c33dc2bb6b97069c76aa10ba14b/codex-rs/core/src/tools/registry.rs#L38-L219)

Current registry shape in this PR:
[`codex-rs/core/src/tools/registry.rs`](https://github.com/openai/codex/blob/41f7ac0adeac81d667541853d6546267d6083613/codex-rs/core/src/tools/registry.rs#L38-L203)

- `ToolHandler::{is_mutating, handle}` now return native `impl Future +
Send` futures instead of using `#[async_trait]`.
- `AnyToolHandler` remains the object-safe adapter and boxes those
futures at the registry boundary with explicit lifetimes.
- Concrete handlers and the registry test handler drop `#[async_trait]`
but otherwise keep their async method bodies intact.
- Representative examples:
[`codex-rs/core/src/tools/handlers/shell.rs`](https://github.com/openai/codex/blob/41f7ac0adeac81d667541853d6546267d6083613/codex-rs/core/src/tools/handlers/shell.rs#L223-L379),
[`codex-rs/core/src/tools/handlers/unified_exec.rs`](https://github.com/openai/codex/blob/41f7ac0adeac81d667541853d6546267d6083613/codex-rs/core/src/tools/handlers/unified_exec.rs),
[`codex-rs/core/src/tools/registry_tests.rs`](https://github.com/openai/codex/blob/41f7ac0adeac81d667541853d6546267d6083613/codex-rs/core/src/tools/registry_tests.rs)

## Tradeoff

This is intentionally less invasive than openai#16627: it does **not** move
result boxing into every concrete handler and does **not** change
`ToolHandler` into an object-safe trait.

Instead, it keeps the existing registry-level type-erasure boundary and
only removes the macro-generated async wrapper layer from concrete
impls. So the runtime boxing story stays basically the same as before,
while the compile-time savings are still large.

## Verification

Existing verification for this branch still applies:

- Ran `cargo test -p codex-core`; this change compiled and the suite
reached the known unrelated `config::tests::*guardian*` failures, with
no local diff under `codex-rs/core/src/config/`.

Profiling commands used for the tables above:

- `cargo clean -p codex-core`
- `cargo +nightly build -p codex-core --lib -Z unstable-options
--timings=json`
- `cargo +nightly rustc -p codex-core --lib -- -Z time-passes -Z
time-passes-format=json`
- `cargo +nightly rustc -p codex-core --lib -- -Z self-profile=... -Z
self-profile-events=default,query-keys,args,llvm,artifact-sizes`
- `measureme summarize -p 0.5`
rebroad pushed a commit to rebroad/codex that referenced this pull request Apr 9, 2026
@bolinfest @rebroad
core: cut codex-core compile time 63% with native async ToolHandler (o…
15fda4e 
…penai#16630)

`ToolHandler` was still paying a large compile-time tax from
`#[async_trait]` on every concrete handler impl, even though the only
object-safe boundary the registry actually stores is the internal
`AnyToolHandler` adapter.

This PR removes that macro-generated async wrapper layer from concrete
`ToolHandler` impls while keeping the existing object-safe shim in
`AnyToolHandler`. In practice, that gets essentially the same
compile-time win as the larger type-erasure refactor in openai#16627, but with
a much smaller diff and without changing the public shape of
`ToolHandler<Output = T>`.

That tradeoff matters here because this is a broad `codex-core` hotspot
and reviewers should be able to judge the compile-time impact from hard
numbers, not vibes.

On a clean `codex-core` package rebuild (`cargo clean -p codex-core`
before each command), rustc `total` dropped from **187.15s to 68.98s**
versus the shared `0bd31dc382bd` baseline: **-63.1%**.

The biggest hot passes dropped by roughly **71-72%**:

| Metric | Baseline `0bd31dc382bd` | This PR `41f7ac0adeac` | Delta |
|---|---:|---:|---:|
| `total` | 187.15s | 68.98s | **-63.1%** |
| `generate_crate_metadata` | 84.53s | 24.49s | **-71.0%** |
| `MIR_borrow_checking` | 84.13s | 24.58s | **-70.8%** |
| `monomorphization_collector_graph_walk` | 79.74s | 22.19s | **-72.2%**
|
| `evaluate_obligation` self-time | 180.62s | 46.91s | **-74.0%** |

Important caveat: `-Z time-passes` timings are nested, so
`generate_crate_metadata` and `monomorphization_collector_graph_walk`
are mostly overlapping, not additive.

hotspot, but it got there by making `ToolHandler` object-safe and
changing every handler to return `BoxFuture<Result<AnyToolResult, _>>`
directly.

This PR keeps the lower-churn shape:

- `ToolHandler` remains generic over `type Output`.
- Concrete handlers use native RPITIT futures with explicit `Send`
bounds.
- `AnyToolHandler` remains the only object-safe adapter and still does
the boxing at the registry boundary, as before.
- The implementation diff is only **33 files, +28/-77**.

The measurements are at least comparable, and in this run this PR is
slightly faster than openai#16627 on the pass-level total:

| Metric | openai#16627 | This PR | Delta |
|---|---:|---:|---:|
| `total` | 79.90s | 68.98s | **-13.7%** |
| `generate_crate_metadata` | 25.88s | 24.49s | **-5.4%** |
| `monomorphization_collector_graph_walk` | 23.54s | 22.19s | **-5.7%**
|
| `evaluate_obligation` self-time | 43.29s | 46.91s | +8.4% |

`cargo +nightly build -p codex-core --lib -Z unstable-options
--timings=json` after `cargo clean -p codex-core`.

Baseline data below is reused from the shared parent `0bd31dc382bd`
profile because this PR and openai#16627 are both one commit on top of that
same parent.

| Crate | Baseline `duration` | This PR `duration` | Delta | Baseline
`rmeta_time` | This PR `rmeta_time` | Delta |
|---|---:|---:|---:|---:|---:|---:|
| `codex_core` | 187.380776583s | 69.171113833s | **-63.1%** |
174.474507208s | 55.873015583s | **-68.0%** |
| `starlark` | 17.90s | 16.773824125s | -6.3% | n/a | 8.8999965s | n/a |

`cargo +nightly rustc -p codex-core --lib -- -Z time-passes -Z
time-passes-format=json` after `cargo clean -p codex-core`.

| Pass | Baseline | This PR | Delta |
|---|---:|---:|---:|
| `total` | 187.150662083s | 68.978770375s | **-63.1%** |
| `generate_crate_metadata` | 84.531864625s | 24.487462958s | **-71.0%**
|
| `MIR_borrow_checking` | 84.131389375s | 24.575553875s | **-70.8%** |
| `monomorphization_collector_graph_walk` | 79.737515042s |
22.190207417s | **-72.2%** |
| `codegen_crate` | 12.362532292s | 12.695237625s | +2.7% |
| `type_check_crate` | 4.4765405s | 5.442019542s | +21.6% |
| `coherence_checking` | 3.311121208s | 4.239935292s | +28.0% |
| process `real` / `user` / `sys` | 187.70s / 201.87s / 4.99s | 69.52s /
85.90s / 2.92s | n/a |

`cargo +nightly rustc -p codex-core --lib -- -Z self-profile=... -Z
self-profile-events=default,query-keys,args,llvm,artifact-sizes` after
`cargo clean -p codex-core`, summarized with `measureme summarize -p
0.5`.

| Query / phase | Baseline self time | This PR self time | Delta |
Baseline total time | This PR total time | Baseline item count | This PR
item count | Baseline cache hits | This PR cache hits |
|---|---:|---:|---:|---:|---:|---:|---:|---:|---:|
| `evaluate_obligation` | 180.62s | 46.91s | **-74.0%** | 182.08s |
48.37s | 572,234 | 388,659 | 1,130,998 | 1,058,553 |
| `mir_borrowck` | 1.42s | 1.49s | +4.9% | 93.77s | 29.59s | n/a | 6,184
| n/a | 15,298 |
| `typeck` | 1.84s | 1.87s | +1.6% | 2.38s | 2.44s | n/a | 9,367 | n/a |
79,247 |
| `LLVM_module_codegen_emit_obj` | n/a | 17.12s | n/a | 17.01s | 17.12s
| n/a | 256 | n/a | 0 |
| `LLVM_passes` | n/a | 13.07s | n/a | 12.95s | 13.07s | n/a | 1 | n/a |
0 |
| `codegen_module` | n/a | 12.33s | n/a | 12.22s | 13.64s | n/a | 256 |
n/a | 0 |
| `items_of_instance` | n/a | 676.00ms | n/a | n/a | 24.96s | n/a |
99,990 | n/a | 0 |
| `type_op_prove_predicate` | n/a | 660.79ms | n/a | n/a | 24.78s | n/a
| 78,762 | n/a | 235,877 |

| Summary | Baseline | This PR |
|---|---:|---:|
| `evaluate_obligation` % of total CPU | 70.821% | 38.880% |
| self-profile total CPU time | 255.042999997s | 120.661175956s |
| process `real` / `user` / `sys` | 220.96s / 235.02s / 7.09s | 86.35s /
103.66s / 3.54s |

From the same `measureme summarize` output:

| Artifact | Baseline | This PR | Delta |
|---|---:|---:|---:|
| `crate_metadata` | 26,534,471 bytes | 26,545,248 bytes | +10,777 |
| `dep_graph` | 253,181,425 bytes | 239,240,806 bytes | -13,940,619 |
| `linked_artifact` | 565,366,624 bytes | 562,673,176 bytes | -2,693,448
|
| `object_file` | 513,127,264 bytes | 510,464,096 bytes | -2,663,168 |
| `query_cache` | 137,440,945 bytes | 136,982,566 bytes | -458,379 |
| `cgu_instructions` | 3,586,307 bytes | 3,575,121 bytes | -11,186 |
| `codegen_unit_size_estimate` | 2,084,846 bytes | 2,078,773 bytes |
-6,073 |
| `work_product_index` | 19,565 bytes | 19,565 bytes | 0 |

These are the top normalized obligation buckets from the shared baseline
profile:

| Obligation bucket | Samples | Duration |
|---|---:|---:|
| `outlives:tasks::review::ReviewTask` | 1,067 | 6.33s |
| `outlives:tools::handlers::unified_exec::UnifiedExecHandler` | 896 |
5.63s |
| `trait:T as tools::registry::ToolHandler` | 876 | 5.45s |
| `outlives:tools::handlers::shell::ShellHandler` | 888 | 5.37s |
| `outlives:tools::handlers::shell::ShellCommandHandler` | 870 | 5.29s |
|
`outlives:tools::runtimes::shell::unix_escalation::CoreShellActionProvider`
| 637 | 3.73s |
| `outlives:tools::handlers::mcp::McpHandler` | 695 | 3.61s |
| `outlives:tasks::regular::RegularTask` | 726 | 3.57s |

Top `items_of_instance` entries before this change were mostly concrete
async handler/task impls:

| Instance | Duration |
|---|---:|
| `tasks::regular::{impl#2}::run` | 3.79s |
| `tools::handlers::mcp::{impl#0}::handle` | 3.27s |
| `tools::runtimes::shell::unix_escalation::{impl#2}::determine_action`
| 3.09s |
| `tools::handlers::agent_jobs::{impl#11}::handle` | 3.07s |
| `tools::handlers::multi_agents::spawn::{impl#1}::handle` | 2.84s |
| `tasks::review::{impl#4}::run` | 2.82s |
| `tools::handlers::multi_agents_v2::spawn::{impl#2}::handle` | 2.80s |
| `tools::handlers::multi_agents::resume_agent::{impl#1}::handle` |
2.73s |
| `tools::handlers::unified_exec::{impl#2}::handle` | 2.54s |
| `tasks::compact::{impl#4}::run` | 2.45s |

Relevant pre-change registry shape:
[`codex-rs/core/src/tools/registry.rs`](https://github.com/openai/codex/blob/0bd31dc382bd1c33dc2bb6b97069c76aa10ba14b/codex-rs/core/src/tools/registry.rs#L38-L219)

Current registry shape in this PR:
[`codex-rs/core/src/tools/registry.rs`](https://github.com/openai/codex/blob/41f7ac0adeac81d667541853d6546267d6083613/codex-rs/core/src/tools/registry.rs#L38-L203)

- `ToolHandler::{is_mutating, handle}` now return native `impl Future +
Send` futures instead of using `#[async_trait]`.
- `AnyToolHandler` remains the object-safe adapter and boxes those
futures at the registry boundary with explicit lifetimes.
- Concrete handlers and the registry test handler drop `#[async_trait]`
but otherwise keep their async method bodies intact.
- Representative examples:
[`codex-rs/core/src/tools/handlers/shell.rs`](https://github.com/openai/codex/blob/41f7ac0adeac81d667541853d6546267d6083613/codex-rs/core/src/tools/handlers/shell.rs#L223-L379),
[`codex-rs/core/src/tools/handlers/unified_exec.rs`](https://github.com/openai/codex/blob/41f7ac0adeac81d667541853d6546267d6083613/codex-rs/core/src/tools/handlers/unified_exec.rs),
[`codex-rs/core/src/tools/registry_tests.rs`](https://github.com/openai/codex/blob/41f7ac0adeac81d667541853d6546267d6083613/codex-rs/core/src/tools/registry_tests.rs)

This is intentionally less invasive than openai#16627: it does **not** move
result boxing into every concrete handler and does **not** change
`ToolHandler` into an object-safe trait.

Instead, it keeps the existing registry-level type-erasure boundary and
only removes the macro-generated async wrapper layer from concrete
impls. So the runtime boxing story stays basically the same as before,
while the compile-time savings are still large.

Existing verification for this branch still applies:

- Ran `cargo test -p codex-core`; this change compiled and the suite
reached the known unrelated `config::tests::*guardian*` failures, with
no local diff under `codex-rs/core/src/config/`.

Profiling commands used for the tables above:

- `cargo clean -p codex-core`
- `cargo +nightly build -p codex-core --lib -Z unstable-options
--timings=json`
- `cargo +nightly rustc -p codex-core --lib -- -Z time-passes -Z
time-passes-format=json`
- `cargo +nightly rustc -p codex-core --lib -- -Z self-profile=... -Z
self-profile-events=default,query-keys,args,llvm,artifact-sizes`
- `measureme summarize -p 0.5`
rebroad pushed a commit to rebroad/codex that referenced this pull request Apr 9, 2026
@bolinfest @rebroad
core: cut codex-core compile time 63% with native async ToolHandler (o…
0143743 
…penai#16630)

`ToolHandler` was still paying a large compile-time tax from
`#[async_trait]` on every concrete handler impl, even though the only
object-safe boundary the registry actually stores is the internal
`AnyToolHandler` adapter.

This PR removes that macro-generated async wrapper layer from concrete
`ToolHandler` impls while keeping the existing object-safe shim in
`AnyToolHandler`. In practice, that gets essentially the same
compile-time win as the larger type-erasure refactor in openai#16627, but with
a much smaller diff and without changing the public shape of
`ToolHandler<Output = T>`.

That tradeoff matters here because this is a broad `codex-core` hotspot
and reviewers should be able to judge the compile-time impact from hard
numbers, not vibes.

On a clean `codex-core` package rebuild (`cargo clean -p codex-core`
before each command), rustc `total` dropped from **187.15s to 68.98s**
versus the shared `0bd31dc382bd` baseline: **-63.1%**.

The biggest hot passes dropped by roughly **71-72%**:

| Metric | Baseline `0bd31dc382bd` | This PR `41f7ac0adeac` | Delta |
|---|---:|---:|---:|
| `total` | 187.15s | 68.98s | **-63.1%** |
| `generate_crate_metadata` | 84.53s | 24.49s | **-71.0%** |
| `MIR_borrow_checking` | 84.13s | 24.58s | **-70.8%** |
| `monomorphization_collector_graph_walk` | 79.74s | 22.19s | **-72.2%**
|
| `evaluate_obligation` self-time | 180.62s | 46.91s | **-74.0%** |

Important caveat: `-Z time-passes` timings are nested, so
`generate_crate_metadata` and `monomorphization_collector_graph_walk`
are mostly overlapping, not additive.

hotspot, but it got there by making `ToolHandler` object-safe and
changing every handler to return `BoxFuture<Result<AnyToolResult, _>>`
directly.

This PR keeps the lower-churn shape:

- `ToolHandler` remains generic over `type Output`.
- Concrete handlers use native RPITIT futures with explicit `Send`
bounds.
- `AnyToolHandler` remains the only object-safe adapter and still does
the boxing at the registry boundary, as before.
- The implementation diff is only **33 files, +28/-77**.

The measurements are at least comparable, and in this run this PR is
slightly faster than openai#16627 on the pass-level total:

| Metric | openai#16627 | This PR | Delta |
|---|---:|---:|---:|
| `total` | 79.90s | 68.98s | **-13.7%** |
| `generate_crate_metadata` | 25.88s | 24.49s | **-5.4%** |
| `monomorphization_collector_graph_walk` | 23.54s | 22.19s | **-5.7%**
|
| `evaluate_obligation` self-time | 43.29s | 46.91s | +8.4% |

`cargo +nightly build -p codex-core --lib -Z unstable-options
--timings=json` after `cargo clean -p codex-core`.

Baseline data below is reused from the shared parent `0bd31dc382bd`
profile because this PR and openai#16627 are both one commit on top of that
same parent.

| Crate | Baseline `duration` | This PR `duration` | Delta | Baseline
`rmeta_time` | This PR `rmeta_time` | Delta |
|---|---:|---:|---:|---:|---:|---:|
| `codex_core` | 187.380776583s | 69.171113833s | **-63.1%** |
174.474507208s | 55.873015583s | **-68.0%** |
| `starlark` | 17.90s | 16.773824125s | -6.3% | n/a | 8.8999965s | n/a |

`cargo +nightly rustc -p codex-core --lib -- -Z time-passes -Z
time-passes-format=json` after `cargo clean -p codex-core`.

| Pass | Baseline | This PR | Delta |
|---|---:|---:|---:|
| `total` | 187.150662083s | 68.978770375s | **-63.1%** |
| `generate_crate_metadata` | 84.531864625s | 24.487462958s | **-71.0%**
|
| `MIR_borrow_checking` | 84.131389375s | 24.575553875s | **-70.8%** |
| `monomorphization_collector_graph_walk` | 79.737515042s |
22.190207417s | **-72.2%** |
| `codegen_crate` | 12.362532292s | 12.695237625s | +2.7% |
| `type_check_crate` | 4.4765405s | 5.442019542s | +21.6% |
| `coherence_checking` | 3.311121208s | 4.239935292s | +28.0% |
| process `real` / `user` / `sys` | 187.70s / 201.87s / 4.99s | 69.52s /
85.90s / 2.92s | n/a |

`cargo +nightly rustc -p codex-core --lib -- -Z self-profile=... -Z
self-profile-events=default,query-keys,args,llvm,artifact-sizes` after
`cargo clean -p codex-core`, summarized with `measureme summarize -p
0.5`.

| Query / phase | Baseline self time | This PR self time | Delta |
Baseline total time | This PR total time | Baseline item count | This PR
item count | Baseline cache hits | This PR cache hits |
|---|---:|---:|---:|---:|---:|---:|---:|---:|---:|
| `evaluate_obligation` | 180.62s | 46.91s | **-74.0%** | 182.08s |
48.37s | 572,234 | 388,659 | 1,130,998 | 1,058,553 |
| `mir_borrowck` | 1.42s | 1.49s | +4.9% | 93.77s | 29.59s | n/a | 6,184
| n/a | 15,298 |
| `typeck` | 1.84s | 1.87s | +1.6% | 2.38s | 2.44s | n/a | 9,367 | n/a |
79,247 |
| `LLVM_module_codegen_emit_obj` | n/a | 17.12s | n/a | 17.01s | 17.12s
| n/a | 256 | n/a | 0 |
| `LLVM_passes` | n/a | 13.07s | n/a | 12.95s | 13.07s | n/a | 1 | n/a |
0 |
| `codegen_module` | n/a | 12.33s | n/a | 12.22s | 13.64s | n/a | 256 |
n/a | 0 |
| `items_of_instance` | n/a | 676.00ms | n/a | n/a | 24.96s | n/a |
99,990 | n/a | 0 |
| `type_op_prove_predicate` | n/a | 660.79ms | n/a | n/a | 24.78s | n/a
| 78,762 | n/a | 235,877 |

| Summary | Baseline | This PR |
|---|---:|---:|
| `evaluate_obligation` % of total CPU | 70.821% | 38.880% |
| self-profile total CPU time | 255.042999997s | 120.661175956s |
| process `real` / `user` / `sys` | 220.96s / 235.02s / 7.09s | 86.35s /
103.66s / 3.54s |

From the same `measureme summarize` output:

| Artifact | Baseline | This PR | Delta |
|---|---:|---:|---:|
| `crate_metadata` | 26,534,471 bytes | 26,545,248 bytes | +10,777 |
| `dep_graph` | 253,181,425 bytes | 239,240,806 bytes | -13,940,619 |
| `linked_artifact` | 565,366,624 bytes | 562,673,176 bytes | -2,693,448
|
| `object_file` | 513,127,264 bytes | 510,464,096 bytes | -2,663,168 |
| `query_cache` | 137,440,945 bytes | 136,982,566 bytes | -458,379 |
| `cgu_instructions` | 3,586,307 bytes | 3,575,121 bytes | -11,186 |
| `codegen_unit_size_estimate` | 2,084,846 bytes | 2,078,773 bytes |
-6,073 |
| `work_product_index` | 19,565 bytes | 19,565 bytes | 0 |

These are the top normalized obligation buckets from the shared baseline
profile:

| Obligation bucket | Samples | Duration |
|---|---:|---:|
| `outlives:tasks::review::ReviewTask` | 1,067 | 6.33s |
| `outlives:tools::handlers::unified_exec::UnifiedExecHandler` | 896 |
5.63s |
| `trait:T as tools::registry::ToolHandler` | 876 | 5.45s |
| `outlives:tools::handlers::shell::ShellHandler` | 888 | 5.37s |
| `outlives:tools::handlers::shell::ShellCommandHandler` | 870 | 5.29s |
|
`outlives:tools::runtimes::shell::unix_escalation::CoreShellActionProvider`
| 637 | 3.73s |
| `outlives:tools::handlers::mcp::McpHandler` | 695 | 3.61s |
| `outlives:tasks::regular::RegularTask` | 726 | 3.57s |

Top `items_of_instance` entries before this change were mostly concrete
async handler/task impls:

| Instance | Duration |
|---|---:|
| `tasks::regular::{impl#2}::run` | 3.79s |
| `tools::handlers::mcp::{impl#0}::handle` | 3.27s |
| `tools::runtimes::shell::unix_escalation::{impl#2}::determine_action`
| 3.09s |
| `tools::handlers::agent_jobs::{impl#11}::handle` | 3.07s |
| `tools::handlers::multi_agents::spawn::{impl#1}::handle` | 2.84s |
| `tasks::review::{impl#4}::run` | 2.82s |
| `tools::handlers::multi_agents_v2::spawn::{impl#2}::handle` | 2.80s |
| `tools::handlers::multi_agents::resume_agent::{impl#1}::handle` |
2.73s |
| `tools::handlers::unified_exec::{impl#2}::handle` | 2.54s |
| `tasks::compact::{impl#4}::run` | 2.45s |

Relevant pre-change registry shape:
[`codex-rs/core/src/tools/registry.rs`](https://github.com/openai/codex/blob/0bd31dc382bd1c33dc2bb6b97069c76aa10ba14b/codex-rs/core/src/tools/registry.rs#L38-L219)

Current registry shape in this PR:
[`codex-rs/core/src/tools/registry.rs`](https://github.com/openai/codex/blob/41f7ac0adeac81d667541853d6546267d6083613/codex-rs/core/src/tools/registry.rs#L38-L203)

- `ToolHandler::{is_mutating, handle}` now return native `impl Future +
Send` futures instead of using `#[async_trait]`.
- `AnyToolHandler` remains the object-safe adapter and boxes those
futures at the registry boundary with explicit lifetimes.
- Concrete handlers and the registry test handler drop `#[async_trait]`
but otherwise keep their async method bodies intact.
- Representative examples:
[`codex-rs/core/src/tools/handlers/shell.rs`](https://github.com/openai/codex/blob/41f7ac0adeac81d667541853d6546267d6083613/codex-rs/core/src/tools/handlers/shell.rs#L223-L379),
[`codex-rs/core/src/tools/handlers/unified_exec.rs`](https://github.com/openai/codex/blob/41f7ac0adeac81d667541853d6546267d6083613/codex-rs/core/src/tools/handlers/unified_exec.rs),
[`codex-rs/core/src/tools/registry_tests.rs`](https://github.com/openai/codex/blob/41f7ac0adeac81d667541853d6546267d6083613/codex-rs/core/src/tools/registry_tests.rs)

This is intentionally less invasive than openai#16627: it does **not** move
result boxing into every concrete handler and does **not** change
`ToolHandler` into an object-safe trait.

Instead, it keeps the existing registry-level type-erasure boundary and
only removes the macro-generated async wrapper layer from concrete
impls. So the runtime boxing story stays basically the same as before,
while the compile-time savings are still large.

Existing verification for this branch still applies:

- Ran `cargo test -p codex-core`; this change compiled and the suite
reached the known unrelated `config::tests::*guardian*` failures, with
no local diff under `codex-rs/core/src/config/`.

Profiling commands used for the tables above:

- `cargo clean -p codex-core`
- `cargo +nightly build -p codex-core --lib -Z unstable-options
--timings=json`
- `cargo +nightly rustc -p codex-core --lib -- -Z time-passes -Z
time-passes-format=json`
- `cargo +nightly rustc -p codex-core --lib -- -Z self-profile=... -Z
self-profile-events=default,query-keys,args,llvm,artifact-sizes`
- `measureme summarize -p 0.5`
rebroad pushed a commit to rebroad/codex that referenced this pull request Apr 9, 2026
@bolinfest @rebroad
core: cut codex-core compile time 63% with native async ToolHandler (o…
87a8f70 
…penai#16630)

`ToolHandler` was still paying a large compile-time tax from
`#[async_trait]` on every concrete handler impl, even though the only
object-safe boundary the registry actually stores is the internal
`AnyToolHandler` adapter.

This PR removes that macro-generated async wrapper layer from concrete
`ToolHandler` impls while keeping the existing object-safe shim in
`AnyToolHandler`. In practice, that gets essentially the same
compile-time win as the larger type-erasure refactor in openai#16627, but with
a much smaller diff and without changing the public shape of
`ToolHandler<Output = T>`.

That tradeoff matters here because this is a broad `codex-core` hotspot
and reviewers should be able to judge the compile-time impact from hard
numbers, not vibes.

On a clean `codex-core` package rebuild (`cargo clean -p codex-core`
before each command), rustc `total` dropped from **187.15s to 68.98s**
versus the shared `0bd31dc382bd` baseline: **-63.1%**.

The biggest hot passes dropped by roughly **71-72%**:

| Metric | Baseline `0bd31dc382bd` | This PR `41f7ac0adeac` | Delta |
|---|---:|---:|---:|
| `total` | 187.15s | 68.98s | **-63.1%** |
| `generate_crate_metadata` | 84.53s | 24.49s | **-71.0%** |
| `MIR_borrow_checking` | 84.13s | 24.58s | **-70.8%** |
| `monomorphization_collector_graph_walk` | 79.74s | 22.19s | **-72.2%**
|
| `evaluate_obligation` self-time | 180.62s | 46.91s | **-74.0%** |

Important caveat: `-Z time-passes` timings are nested, so
`generate_crate_metadata` and `monomorphization_collector_graph_walk`
are mostly overlapping, not additive.

hotspot, but it got there by making `ToolHandler` object-safe and
changing every handler to return `BoxFuture<Result<AnyToolResult, _>>`
directly.

This PR keeps the lower-churn shape:

- `ToolHandler` remains generic over `type Output`.
- Concrete handlers use native RPITIT futures with explicit `Send`
bounds.
- `AnyToolHandler` remains the only object-safe adapter and still does
the boxing at the registry boundary, as before.
- The implementation diff is only **33 files, +28/-77**.

The measurements are at least comparable, and in this run this PR is
slightly faster than openai#16627 on the pass-level total:

| Metric | openai#16627 | This PR | Delta |
|---|---:|---:|---:|
| `total` | 79.90s | 68.98s | **-13.7%** |
| `generate_crate_metadata` | 25.88s | 24.49s | **-5.4%** |
| `monomorphization_collector_graph_walk` | 23.54s | 22.19s | **-5.7%**
|
| `evaluate_obligation` self-time | 43.29s | 46.91s | +8.4% |

`cargo +nightly build -p codex-core --lib -Z unstable-options
--timings=json` after `cargo clean -p codex-core`.

Baseline data below is reused from the shared parent `0bd31dc382bd`
profile because this PR and openai#16627 are both one commit on top of that
same parent.

| Crate | Baseline `duration` | This PR `duration` | Delta | Baseline
`rmeta_time` | This PR `rmeta_time` | Delta |
|---|---:|---:|---:|---:|---:|---:|
| `codex_core` | 187.380776583s | 69.171113833s | **-63.1%** |
174.474507208s | 55.873015583s | **-68.0%** |
| `starlark` | 17.90s | 16.773824125s | -6.3% | n/a | 8.8999965s | n/a |

`cargo +nightly rustc -p codex-core --lib -- -Z time-passes -Z
time-passes-format=json` after `cargo clean -p codex-core`.

| Pass | Baseline | This PR | Delta |
|---|---:|---:|---:|
| `total` | 187.150662083s | 68.978770375s | **-63.1%** |
| `generate_crate_metadata` | 84.531864625s | 24.487462958s | **-71.0%**
|
| `MIR_borrow_checking` | 84.131389375s | 24.575553875s | **-70.8%** |
| `monomorphization_collector_graph_walk` | 79.737515042s |
22.190207417s | **-72.2%** |
| `codegen_crate` | 12.362532292s | 12.695237625s | +2.7% |
| `type_check_crate` | 4.4765405s | 5.442019542s | +21.6% |
| `coherence_checking` | 3.311121208s | 4.239935292s | +28.0% |
| process `real` / `user` / `sys` | 187.70s / 201.87s / 4.99s | 69.52s /
85.90s / 2.92s | n/a |

`cargo +nightly rustc -p codex-core --lib -- -Z self-profile=... -Z
self-profile-events=default,query-keys,args,llvm,artifact-sizes` after
`cargo clean -p codex-core`, summarized with `measureme summarize -p
0.5`.

| Query / phase | Baseline self time | This PR self time | Delta |
Baseline total time | This PR total time | Baseline item count | This PR
item count | Baseline cache hits | This PR cache hits |
|---|---:|---:|---:|---:|---:|---:|---:|---:|---:|
| `evaluate_obligation` | 180.62s | 46.91s | **-74.0%** | 182.08s |
48.37s | 572,234 | 388,659 | 1,130,998 | 1,058,553 |
| `mir_borrowck` | 1.42s | 1.49s | +4.9% | 93.77s | 29.59s | n/a | 6,184
| n/a | 15,298 |
| `typeck` | 1.84s | 1.87s | +1.6% | 2.38s | 2.44s | n/a | 9,367 | n/a |
79,247 |
| `LLVM_module_codegen_emit_obj` | n/a | 17.12s | n/a | 17.01s | 17.12s
| n/a | 256 | n/a | 0 |
| `LLVM_passes` | n/a | 13.07s | n/a | 12.95s | 13.07s | n/a | 1 | n/a |
0 |
| `codegen_module` | n/a | 12.33s | n/a | 12.22s | 13.64s | n/a | 256 |
n/a | 0 |
| `items_of_instance` | n/a | 676.00ms | n/a | n/a | 24.96s | n/a |
99,990 | n/a | 0 |
| `type_op_prove_predicate` | n/a | 660.79ms | n/a | n/a | 24.78s | n/a
| 78,762 | n/a | 235,877 |

| Summary | Baseline | This PR |
|---|---:|---:|
| `evaluate_obligation` % of total CPU | 70.821% | 38.880% |
| self-profile total CPU time | 255.042999997s | 120.661175956s |
| process `real` / `user` / `sys` | 220.96s / 235.02s / 7.09s | 86.35s /
103.66s / 3.54s |

From the same `measureme summarize` output:

| Artifact | Baseline | This PR | Delta |
|---|---:|---:|---:|
| `crate_metadata` | 26,534,471 bytes | 26,545,248 bytes | +10,777 |
| `dep_graph` | 253,181,425 bytes | 239,240,806 bytes | -13,940,619 |
| `linked_artifact` | 565,366,624 bytes | 562,673,176 bytes | -2,693,448
|
| `object_file` | 513,127,264 bytes | 510,464,096 bytes | -2,663,168 |
| `query_cache` | 137,440,945 bytes | 136,982,566 bytes | -458,379 |
| `cgu_instructions` | 3,586,307 bytes | 3,575,121 bytes | -11,186 |
| `codegen_unit_size_estimate` | 2,084,846 bytes | 2,078,773 bytes |
-6,073 |
| `work_product_index` | 19,565 bytes | 19,565 bytes | 0 |

These are the top normalized obligation buckets from the shared baseline
profile:

| Obligation bucket | Samples | Duration |
|---|---:|---:|
| `outlives:tasks::review::ReviewTask` | 1,067 | 6.33s |
| `outlives:tools::handlers::unified_exec::UnifiedExecHandler` | 896 |
5.63s |
| `trait:T as tools::registry::ToolHandler` | 876 | 5.45s |
| `outlives:tools::handlers::shell::ShellHandler` | 888 | 5.37s |
| `outlives:tools::handlers::shell::ShellCommandHandler` | 870 | 5.29s |
|
`outlives:tools::runtimes::shell::unix_escalation::CoreShellActionProvider`
| 637 | 3.73s |
| `outlives:tools::handlers::mcp::McpHandler` | 695 | 3.61s |
| `outlives:tasks::regular::RegularTask` | 726 | 3.57s |

Top `items_of_instance` entries before this change were mostly concrete
async handler/task impls:

| Instance | Duration |
|---|---:|
| `tasks::regular::{impl#2}::run` | 3.79s |
| `tools::handlers::mcp::{impl#0}::handle` | 3.27s |
| `tools::runtimes::shell::unix_escalation::{impl#2}::determine_action`
| 3.09s |
| `tools::handlers::agent_jobs::{impl#11}::handle` | 3.07s |
| `tools::handlers::multi_agents::spawn::{impl#1}::handle` | 2.84s |
| `tasks::review::{impl#4}::run` | 2.82s |
| `tools::handlers::multi_agents_v2::spawn::{impl#2}::handle` | 2.80s |
| `tools::handlers::multi_agents::resume_agent::{impl#1}::handle` |
2.73s |
| `tools::handlers::unified_exec::{impl#2}::handle` | 2.54s |
| `tasks::compact::{impl#4}::run` | 2.45s |

Relevant pre-change registry shape:
[`codex-rs/core/src/tools/registry.rs`](https://github.com/openai/codex/blob/0bd31dc382bd1c33dc2bb6b97069c76aa10ba14b/codex-rs/core/src/tools/registry.rs#L38-L219)

Current registry shape in this PR:
[`codex-rs/core/src/tools/registry.rs`](https://github.com/openai/codex/blob/41f7ac0adeac81d667541853d6546267d6083613/codex-rs/core/src/tools/registry.rs#L38-L203)

- `ToolHandler::{is_mutating, handle}` now return native `impl Future +
Send` futures instead of using `#[async_trait]`.
- `AnyToolHandler` remains the object-safe adapter and boxes those
futures at the registry boundary with explicit lifetimes.
- Concrete handlers and the registry test handler drop `#[async_trait]`
but otherwise keep their async method bodies intact.
- Representative examples:
[`codex-rs/core/src/tools/handlers/shell.rs`](https://github.com/openai/codex/blob/41f7ac0adeac81d667541853d6546267d6083613/codex-rs/core/src/tools/handlers/shell.rs#L223-L379),
[`codex-rs/core/src/tools/handlers/unified_exec.rs`](https://github.com/openai/codex/blob/41f7ac0adeac81d667541853d6546267d6083613/codex-rs/core/src/tools/handlers/unified_exec.rs),
[`codex-rs/core/src/tools/registry_tests.rs`](https://github.com/openai/codex/blob/41f7ac0adeac81d667541853d6546267d6083613/codex-rs/core/src/tools/registry_tests.rs)

This is intentionally less invasive than openai#16627: it does **not** move
result boxing into every concrete handler and does **not** change
`ToolHandler` into an object-safe trait.

Instead, it keeps the existing registry-level type-erasure boundary and
only removes the macro-generated async wrapper layer from concrete
impls. So the runtime boxing story stays basically the same as before,
while the compile-time savings are still large.

Existing verification for this branch still applies:

- Ran `cargo test -p codex-core`; this change compiled and the suite
reached the known unrelated `config::tests::*guardian*` failures, with
no local diff under `codex-rs/core/src/config/`.

Profiling commands used for the tables above:

- `cargo clean -p codex-core`
- `cargo +nightly build -p codex-core --lib -Z unstable-options
--timings=json`
- `cargo +nightly rustc -p codex-core --lib -- -Z time-passes -Z
time-passes-format=json`
- `cargo +nightly rustc -p codex-core --lib -- -Z self-profile=... -Z
self-profile-events=default,query-keys,args,llvm,artifact-sizes`
- `measureme summarize -p 0.5`
rebroad pushed a commit to rebroad/codex that referenced this pull request Apr 9, 2026
@bolinfest @rebroad
core: cut codex-core compile time 63% with native async ToolHandler (o…
9d0efd4 
…penai#16630)

`ToolHandler` was still paying a large compile-time tax from
`#[async_trait]` on every concrete handler impl, even though the only
object-safe boundary the registry actually stores is the internal
`AnyToolHandler` adapter.

This PR removes that macro-generated async wrapper layer from concrete
`ToolHandler` impls while keeping the existing object-safe shim in
`AnyToolHandler`. In practice, that gets essentially the same
compile-time win as the larger type-erasure refactor in openai#16627, but with
a much smaller diff and without changing the public shape of
`ToolHandler<Output = T>`.

That tradeoff matters here because this is a broad `codex-core` hotspot
and reviewers should be able to judge the compile-time impact from hard
numbers, not vibes.

On a clean `codex-core` package rebuild (`cargo clean -p codex-core`
before each command), rustc `total` dropped from **187.15s to 68.98s**
versus the shared `0bd31dc382bd` baseline: **-63.1%**.

The biggest hot passes dropped by roughly **71-72%**:

| Metric | Baseline `0bd31dc382bd` | This PR `41f7ac0adeac` | Delta |
|---|---:|---:|---:|
| `total` | 187.15s | 68.98s | **-63.1%** |
| `generate_crate_metadata` | 84.53s | 24.49s | **-71.0%** |
| `MIR_borrow_checking` | 84.13s | 24.58s | **-70.8%** |
| `monomorphization_collector_graph_walk` | 79.74s | 22.19s | **-72.2%**
|
| `evaluate_obligation` self-time | 180.62s | 46.91s | **-74.0%** |

Important caveat: `-Z time-passes` timings are nested, so
`generate_crate_metadata` and `monomorphization_collector_graph_walk`
are mostly overlapping, not additive.

hotspot, but it got there by making `ToolHandler` object-safe and
changing every handler to return `BoxFuture<Result<AnyToolResult, _>>`
directly.

This PR keeps the lower-churn shape:

- `ToolHandler` remains generic over `type Output`.
- Concrete handlers use native RPITIT futures with explicit `Send`
bounds.
- `AnyToolHandler` remains the only object-safe adapter and still does
the boxing at the registry boundary, as before.
- The implementation diff is only **33 files, +28/-77**.

The measurements are at least comparable, and in this run this PR is
slightly faster than openai#16627 on the pass-level total:

| Metric | openai#16627 | This PR | Delta |
|---|---:|---:|---:|
| `total` | 79.90s | 68.98s | **-13.7%** |
| `generate_crate_metadata` | 25.88s | 24.49s | **-5.4%** |
| `monomorphization_collector_graph_walk` | 23.54s | 22.19s | **-5.7%**
|
| `evaluate_obligation` self-time | 43.29s | 46.91s | +8.4% |

`cargo +nightly build -p codex-core --lib -Z unstable-options
--timings=json` after `cargo clean -p codex-core`.

Baseline data below is reused from the shared parent `0bd31dc382bd`
profile because this PR and openai#16627 are both one commit on top of that
same parent.

| Crate | Baseline `duration` | This PR `duration` | Delta | Baseline
`rmeta_time` | This PR `rmeta_time` | Delta |
|---|---:|---:|---:|---:|---:|---:|
| `codex_core` | 187.380776583s | 69.171113833s | **-63.1%** |
174.474507208s | 55.873015583s | **-68.0%** |
| `starlark` | 17.90s | 16.773824125s | -6.3% | n/a | 8.8999965s | n/a |

`cargo +nightly rustc -p codex-core --lib -- -Z time-passes -Z
time-passes-format=json` after `cargo clean -p codex-core`.

| Pass | Baseline | This PR | Delta |
|---|---:|---:|---:|
| `total` | 187.150662083s | 68.978770375s | **-63.1%** |
| `generate_crate_metadata` | 84.531864625s | 24.487462958s | **-71.0%**
|
| `MIR_borrow_checking` | 84.131389375s | 24.575553875s | **-70.8%** |
| `monomorphization_collector_graph_walk` | 79.737515042s |
22.190207417s | **-72.2%** |
| `codegen_crate` | 12.362532292s | 12.695237625s | +2.7% |
| `type_check_crate` | 4.4765405s | 5.442019542s | +21.6% |
| `coherence_checking` | 3.311121208s | 4.239935292s | +28.0% |
| process `real` / `user` / `sys` | 187.70s / 201.87s / 4.99s | 69.52s /
85.90s / 2.92s | n/a |

`cargo +nightly rustc -p codex-core --lib -- -Z self-profile=... -Z
self-profile-events=default,query-keys,args,llvm,artifact-sizes` after
`cargo clean -p codex-core`, summarized with `measureme summarize -p
0.5`.

| Query / phase | Baseline self time | This PR self time | Delta |
Baseline total time | This PR total time | Baseline item count | This PR
item count | Baseline cache hits | This PR cache hits |
|---|---:|---:|---:|---:|---:|---:|---:|---:|---:|
| `evaluate_obligation` | 180.62s | 46.91s | **-74.0%** | 182.08s |
48.37s | 572,234 | 388,659 | 1,130,998 | 1,058,553 |
| `mir_borrowck` | 1.42s | 1.49s | +4.9% | 93.77s | 29.59s | n/a | 6,184
| n/a | 15,298 |
| `typeck` | 1.84s | 1.87s | +1.6% | 2.38s | 2.44s | n/a | 9,367 | n/a |
79,247 |
| `LLVM_module_codegen_emit_obj` | n/a | 17.12s | n/a | 17.01s | 17.12s
| n/a | 256 | n/a | 0 |
| `LLVM_passes` | n/a | 13.07s | n/a | 12.95s | 13.07s | n/a | 1 | n/a |
0 |
| `codegen_module` | n/a | 12.33s | n/a | 12.22s | 13.64s | n/a | 256 |
n/a | 0 |
| `items_of_instance` | n/a | 676.00ms | n/a | n/a | 24.96s | n/a |
99,990 | n/a | 0 |
| `type_op_prove_predicate` | n/a | 660.79ms | n/a | n/a | 24.78s | n/a
| 78,762 | n/a | 235,877 |

| Summary | Baseline | This PR |
|---|---:|---:|
| `evaluate_obligation` % of total CPU | 70.821% | 38.880% |
| self-profile total CPU time | 255.042999997s | 120.661175956s |
| process `real` / `user` / `sys` | 220.96s / 235.02s / 7.09s | 86.35s /
103.66s / 3.54s |

From the same `measureme summarize` output:

| Artifact | Baseline | This PR | Delta |
|---|---:|---:|---:|
| `crate_metadata` | 26,534,471 bytes | 26,545,248 bytes | +10,777 |
| `dep_graph` | 253,181,425 bytes | 239,240,806 bytes | -13,940,619 |
| `linked_artifact` | 565,366,624 bytes | 562,673,176 bytes | -2,693,448
|
| `object_file` | 513,127,264 bytes | 510,464,096 bytes | -2,663,168 |
| `query_cache` | 137,440,945 bytes | 136,982,566 bytes | -458,379 |
| `cgu_instructions` | 3,586,307 bytes | 3,575,121 bytes | -11,186 |
| `codegen_unit_size_estimate` | 2,084,846 bytes | 2,078,773 bytes |
-6,073 |
| `work_product_index` | 19,565 bytes | 19,565 bytes | 0 |

These are the top normalized obligation buckets from the shared baseline
profile:

| Obligation bucket | Samples | Duration |
|---|---:|---:|
| `outlives:tasks::review::ReviewTask` | 1,067 | 6.33s |
| `outlives:tools::handlers::unified_exec::UnifiedExecHandler` | 896 |
5.63s |
| `trait:T as tools::registry::ToolHandler` | 876 | 5.45s |
| `outlives:tools::handlers::shell::ShellHandler` | 888 | 5.37s |
| `outlives:tools::handlers::shell::ShellCommandHandler` | 870 | 5.29s |
|
`outlives:tools::runtimes::shell::unix_escalation::CoreShellActionProvider`
| 637 | 3.73s |
| `outlives:tools::handlers::mcp::McpHandler` | 695 | 3.61s |
| `outlives:tasks::regular::RegularTask` | 726 | 3.57s |

Top `items_of_instance` entries before this change were mostly concrete
async handler/task impls:

| Instance | Duration |
|---|---:|
| `tasks::regular::{impl#2}::run` | 3.79s |
| `tools::handlers::mcp::{impl#0}::handle` | 3.27s |
| `tools::runtimes::shell::unix_escalation::{impl#2}::determine_action`
| 3.09s |
| `tools::handlers::agent_jobs::{impl#11}::handle` | 3.07s |
| `tools::handlers::multi_agents::spawn::{impl#1}::handle` | 2.84s |
| `tasks::review::{impl#4}::run` | 2.82s |
| `tools::handlers::multi_agents_v2::spawn::{impl#2}::handle` | 2.80s |
| `tools::handlers::multi_agents::resume_agent::{impl#1}::handle` |
2.73s |
| `tools::handlers::unified_exec::{impl#2}::handle` | 2.54s |
| `tasks::compact::{impl#4}::run` | 2.45s |

Relevant pre-change registry shape:
[`codex-rs/core/src/tools/registry.rs`](https://github.com/openai/codex/blob/0bd31dc382bd1c33dc2bb6b97069c76aa10ba14b/codex-rs/core/src/tools/registry.rs#L38-L219)

Current registry shape in this PR:
[`codex-rs/core/src/tools/registry.rs`](https://github.com/openai/codex/blob/41f7ac0adeac81d667541853d6546267d6083613/codex-rs/core/src/tools/registry.rs#L38-L203)

- `ToolHandler::{is_mutating, handle}` now return native `impl Future +
Send` futures instead of using `#[async_trait]`.
- `AnyToolHandler` remains the object-safe adapter and boxes those
futures at the registry boundary with explicit lifetimes.
- Concrete handlers and the registry test handler drop `#[async_trait]`
but otherwise keep their async method bodies intact.
- Representative examples:
[`codex-rs/core/src/tools/handlers/shell.rs`](https://github.com/openai/codex/blob/41f7ac0adeac81d667541853d6546267d6083613/codex-rs/core/src/tools/handlers/shell.rs#L223-L379),
[`codex-rs/core/src/tools/handlers/unified_exec.rs`](https://github.com/openai/codex/blob/41f7ac0adeac81d667541853d6546267d6083613/codex-rs/core/src/tools/handlers/unified_exec.rs),
[`codex-rs/core/src/tools/registry_tests.rs`](https://github.com/openai/codex/blob/41f7ac0adeac81d667541853d6546267d6083613/codex-rs/core/src/tools/registry_tests.rs)

This is intentionally less invasive than openai#16627: it does **not** move
result boxing into every concrete handler and does **not** change
`ToolHandler` into an object-safe trait.

Instead, it keeps the existing registry-level type-erasure boundary and
only removes the macro-generated async wrapper layer from concrete
impls. So the runtime boxing story stays basically the same as before,
while the compile-time savings are still large.

Existing verification for this branch still applies:

- Ran `cargo test -p codex-core`; this change compiled and the suite
reached the known unrelated `config::tests::*guardian*` failures, with
no local diff under `codex-rs/core/src/config/`.

Profiling commands used for the tables above:

- `cargo clean -p codex-core`
- `cargo +nightly build -p codex-core --lib -Z unstable-options
--timings=json`
- `cargo +nightly rustc -p codex-core --lib -- -Z time-passes -Z
time-passes-format=json`
- `cargo +nightly rustc -p codex-core --lib -- -Z self-profile=... -Z
self-profile-events=default,query-keys,args,llvm,artifact-sizes`
- `measureme summarize -p 0.5`
rebroad pushed a commit to rebroad/codex that referenced this pull request Apr 10, 2026
@bolinfest @rebroad
core: cut codex-core compile time 63% with native async ToolHandler (o…
0e7d8c0 
…penai#16630)

`ToolHandler` was still paying a large compile-time tax from
`#[async_trait]` on every concrete handler impl, even though the only
object-safe boundary the registry actually stores is the internal
`AnyToolHandler` adapter.

This PR removes that macro-generated async wrapper layer from concrete
`ToolHandler` impls while keeping the existing object-safe shim in
`AnyToolHandler`. In practice, that gets essentially the same
compile-time win as the larger type-erasure refactor in openai#16627, but with
a much smaller diff and without changing the public shape of
`ToolHandler<Output = T>`.

That tradeoff matters here because this is a broad `codex-core` hotspot
and reviewers should be able to judge the compile-time impact from hard
numbers, not vibes.

On a clean `codex-core` package rebuild (`cargo clean -p codex-core`
before each command), rustc `total` dropped from **187.15s to 68.98s**
versus the shared `0bd31dc382bd` baseline: **-63.1%**.

The biggest hot passes dropped by roughly **71-72%**:

| Metric | Baseline `0bd31dc382bd` | This PR `41f7ac0adeac` | Delta |
|---|---:|---:|---:|
| `total` | 187.15s | 68.98s | **-63.1%** |
| `generate_crate_metadata` | 84.53s | 24.49s | **-71.0%** |
| `MIR_borrow_checking` | 84.13s | 24.58s | **-70.8%** |
| `monomorphization_collector_graph_walk` | 79.74s | 22.19s | **-72.2%**
|
| `evaluate_obligation` self-time | 180.62s | 46.91s | **-74.0%** |

Important caveat: `-Z time-passes` timings are nested, so
`generate_crate_metadata` and `monomorphization_collector_graph_walk`
are mostly overlapping, not additive.

hotspot, but it got there by making `ToolHandler` object-safe and
changing every handler to return `BoxFuture<Result<AnyToolResult, _>>`
directly.

This PR keeps the lower-churn shape:

- `ToolHandler` remains generic over `type Output`.
- Concrete handlers use native RPITIT futures with explicit `Send`
bounds.
- `AnyToolHandler` remains the only object-safe adapter and still does
the boxing at the registry boundary, as before.
- The implementation diff is only **33 files, +28/-77**.

The measurements are at least comparable, and in this run this PR is
slightly faster than openai#16627 on the pass-level total:

| Metric | openai#16627 | This PR | Delta |
|---|---:|---:|---:|
| `total` | 79.90s | 68.98s | **-13.7%** |
| `generate_crate_metadata` | 25.88s | 24.49s | **-5.4%** |
| `monomorphization_collector_graph_walk` | 23.54s | 22.19s | **-5.7%**
|
| `evaluate_obligation` self-time | 43.29s | 46.91s | +8.4% |

`cargo +nightly build -p codex-core --lib -Z unstable-options
--timings=json` after `cargo clean -p codex-core`.

Baseline data below is reused from the shared parent `0bd31dc382bd`
profile because this PR and openai#16627 are both one commit on top of that
same parent.

| Crate | Baseline `duration` | This PR `duration` | Delta | Baseline
`rmeta_time` | This PR `rmeta_time` | Delta |
|---|---:|---:|---:|---:|---:|---:|
| `codex_core` | 187.380776583s | 69.171113833s | **-63.1%** |
174.474507208s | 55.873015583s | **-68.0%** |
| `starlark` | 17.90s | 16.773824125s | -6.3% | n/a | 8.8999965s | n/a |

`cargo +nightly rustc -p codex-core --lib -- -Z time-passes -Z
time-passes-format=json` after `cargo clean -p codex-core`.

| Pass | Baseline | This PR | Delta |
|---|---:|---:|---:|
| `total` | 187.150662083s | 68.978770375s | **-63.1%** |
| `generate_crate_metadata` | 84.531864625s | 24.487462958s | **-71.0%**
|
| `MIR_borrow_checking` | 84.131389375s | 24.575553875s | **-70.8%** |
| `monomorphization_collector_graph_walk` | 79.737515042s |
22.190207417s | **-72.2%** |
| `codegen_crate` | 12.362532292s | 12.695237625s | +2.7% |
| `type_check_crate` | 4.4765405s | 5.442019542s | +21.6% |
| `coherence_checking` | 3.311121208s | 4.239935292s | +28.0% |
| process `real` / `user` / `sys` | 187.70s / 201.87s / 4.99s | 69.52s /
85.90s / 2.92s | n/a |

`cargo +nightly rustc -p codex-core --lib -- -Z self-profile=... -Z
self-profile-events=default,query-keys,args,llvm,artifact-sizes` after
`cargo clean -p codex-core`, summarized with `measureme summarize -p
0.5`.

| Query / phase | Baseline self time | This PR self time | Delta |
Baseline total time | This PR total time | Baseline item count | This PR
item count | Baseline cache hits | This PR cache hits |
|---|---:|---:|---:|---:|---:|---:|---:|---:|---:|
| `evaluate_obligation` | 180.62s | 46.91s | **-74.0%** | 182.08s |
48.37s | 572,234 | 388,659 | 1,130,998 | 1,058,553 |
| `mir_borrowck` | 1.42s | 1.49s | +4.9% | 93.77s | 29.59s | n/a | 6,184
| n/a | 15,298 |
| `typeck` | 1.84s | 1.87s | +1.6% | 2.38s | 2.44s | n/a | 9,367 | n/a |
79,247 |
| `LLVM_module_codegen_emit_obj` | n/a | 17.12s | n/a | 17.01s | 17.12s
| n/a | 256 | n/a | 0 |
| `LLVM_passes` | n/a | 13.07s | n/a | 12.95s | 13.07s | n/a | 1 | n/a |
0 |
| `codegen_module` | n/a | 12.33s | n/a | 12.22s | 13.64s | n/a | 256 |
n/a | 0 |
| `items_of_instance` | n/a | 676.00ms | n/a | n/a | 24.96s | n/a |
99,990 | n/a | 0 |
| `type_op_prove_predicate` | n/a | 660.79ms | n/a | n/a | 24.78s | n/a
| 78,762 | n/a | 235,877 |

| Summary | Baseline | This PR |
|---|---:|---:|
| `evaluate_obligation` % of total CPU | 70.821% | 38.880% |
| self-profile total CPU time | 255.042999997s | 120.661175956s |
| process `real` / `user` / `sys` | 220.96s / 235.02s / 7.09s | 86.35s /
103.66s / 3.54s |

From the same `measureme summarize` output:

| Artifact | Baseline | This PR | Delta |
|---|---:|---:|---:|
| `crate_metadata` | 26,534,471 bytes | 26,545,248 bytes | +10,777 |
| `dep_graph` | 253,181,425 bytes | 239,240,806 bytes | -13,940,619 |
| `linked_artifact` | 565,366,624 bytes | 562,673,176 bytes | -2,693,448
|
| `object_file` | 513,127,264 bytes | 510,464,096 bytes | -2,663,168 |
| `query_cache` | 137,440,945 bytes | 136,982,566 bytes | -458,379 |
| `cgu_instructions` | 3,586,307 bytes | 3,575,121 bytes | -11,186 |
| `codegen_unit_size_estimate` | 2,084,846 bytes | 2,078,773 bytes |
-6,073 |
| `work_product_index` | 19,565 bytes | 19,565 bytes | 0 |

These are the top normalized obligation buckets from the shared baseline
profile:

| Obligation bucket | Samples | Duration |
|---|---:|---:|
| `outlives:tasks::review::ReviewTask` | 1,067 | 6.33s |
| `outlives:tools::handlers::unified_exec::UnifiedExecHandler` | 896 |
5.63s |
| `trait:T as tools::registry::ToolHandler` | 876 | 5.45s |
| `outlives:tools::handlers::shell::ShellHandler` | 888 | 5.37s |
| `outlives:tools::handlers::shell::ShellCommandHandler` | 870 | 5.29s |
|
`outlives:tools::runtimes::shell::unix_escalation::CoreShellActionProvider`
| 637 | 3.73s |
| `outlives:tools::handlers::mcp::McpHandler` | 695 | 3.61s |
| `outlives:tasks::regular::RegularTask` | 726 | 3.57s |

Top `items_of_instance` entries before this change were mostly concrete
async handler/task impls:

| Instance | Duration |
|---|---:|
| `tasks::regular::{impl#2}::run` | 3.79s |
| `tools::handlers::mcp::{impl#0}::handle` | 3.27s |
| `tools::runtimes::shell::unix_escalation::{impl#2}::determine_action`
| 3.09s |
| `tools::handlers::agent_jobs::{impl#11}::handle` | 3.07s |
| `tools::handlers::multi_agents::spawn::{impl#1}::handle` | 2.84s |
| `tasks::review::{impl#4}::run` | 2.82s |
| `tools::handlers::multi_agents_v2::spawn::{impl#2}::handle` | 2.80s |
| `tools::handlers::multi_agents::resume_agent::{impl#1}::handle` |
2.73s |
| `tools::handlers::unified_exec::{impl#2}::handle` | 2.54s |
| `tasks::compact::{impl#4}::run` | 2.45s |

Relevant pre-change registry shape:
[`codex-rs/core/src/tools/registry.rs`](https://github.com/openai/codex/blob/0bd31dc382bd1c33dc2bb6b97069c76aa10ba14b/codex-rs/core/src/tools/registry.rs#L38-L219)

Current registry shape in this PR:
[`codex-rs/core/src/tools/registry.rs`](https://github.com/openai/codex/blob/41f7ac0adeac81d667541853d6546267d6083613/codex-rs/core/src/tools/registry.rs#L38-L203)

- `ToolHandler::{is_mutating, handle}` now return native `impl Future +
Send` futures instead of using `#[async_trait]`.
- `AnyToolHandler` remains the object-safe adapter and boxes those
futures at the registry boundary with explicit lifetimes.
- Concrete handlers and the registry test handler drop `#[async_trait]`
but otherwise keep their async method bodies intact.
- Representative examples:
[`codex-rs/core/src/tools/handlers/shell.rs`](https://github.com/openai/codex/blob/41f7ac0adeac81d667541853d6546267d6083613/codex-rs/core/src/tools/handlers/shell.rs#L223-L379),
[`codex-rs/core/src/tools/handlers/unified_exec.rs`](https://github.com/openai/codex/blob/41f7ac0adeac81d667541853d6546267d6083613/codex-rs/core/src/tools/handlers/unified_exec.rs),
[`codex-rs/core/src/tools/registry_tests.rs`](https://github.com/openai/codex/blob/41f7ac0adeac81d667541853d6546267d6083613/codex-rs/core/src/tools/registry_tests.rs)

This is intentionally less invasive than openai#16627: it does **not** move
result boxing into every concrete handler and does **not** change
`ToolHandler` into an object-safe trait.

Instead, it keeps the existing registry-level type-erasure boundary and
only removes the macro-generated async wrapper layer from concrete
impls. So the runtime boxing story stays basically the same as before,
while the compile-time savings are still large.

Existing verification for this branch still applies:

- Ran `cargo test -p codex-core`; this change compiled and the suite
reached the known unrelated `config::tests::*guardian*` failures, with
no local diff under `codex-rs/core/src/config/`.

Profiling commands used for the tables above:

- `cargo clean -p codex-core`
- `cargo +nightly build -p codex-core --lib -Z unstable-options
--timings=json`
- `cargo +nightly rustc -p codex-core --lib -- -Z time-passes -Z
time-passes-format=json`
- `cargo +nightly rustc -p codex-core --lib -- -Z self-profile=... -Z
self-profile-events=default,query-keys,args,llvm,artifact-sizes`
- `measureme summarize -p 0.5`
rebroad pushed a commit to rebroad/codex that referenced this pull request Apr 10, 2026
@bolinfest @rebroad
core: cut codex-core compile time 63% with native async ToolHandler (o…
85120d3 
…penai#16630)

`ToolHandler` was still paying a large compile-time tax from
`#[async_trait]` on every concrete handler impl, even though the only
object-safe boundary the registry actually stores is the internal
`AnyToolHandler` adapter.

This PR removes that macro-generated async wrapper layer from concrete
`ToolHandler` impls while keeping the existing object-safe shim in
`AnyToolHandler`. In practice, that gets essentially the same
compile-time win as the larger type-erasure refactor in openai#16627, but with
a much smaller diff and without changing the public shape of
`ToolHandler<Output = T>`.

That tradeoff matters here because this is a broad `codex-core` hotspot
and reviewers should be able to judge the compile-time impact from hard
numbers, not vibes.

On a clean `codex-core` package rebuild (`cargo clean -p codex-core`
before each command), rustc `total` dropped from **187.15s to 68.98s**
versus the shared `0bd31dc382bd` baseline: **-63.1%**.

The biggest hot passes dropped by roughly **71-72%**:

| Metric | Baseline `0bd31dc382bd` | This PR `41f7ac0adeac` | Delta |
|---|---:|---:|---:|
| `total` | 187.15s | 68.98s | **-63.1%** |
| `generate_crate_metadata` | 84.53s | 24.49s | **-71.0%** |
| `MIR_borrow_checking` | 84.13s | 24.58s | **-70.8%** |
| `monomorphization_collector_graph_walk` | 79.74s | 22.19s | **-72.2%**
|
| `evaluate_obligation` self-time | 180.62s | 46.91s | **-74.0%** |

Important caveat: `-Z time-passes` timings are nested, so
`generate_crate_metadata` and `monomorphization_collector_graph_walk`
are mostly overlapping, not additive.

hotspot, but it got there by making `ToolHandler` object-safe and
changing every handler to return `BoxFuture<Result<AnyToolResult, _>>`
directly.

This PR keeps the lower-churn shape:

- `ToolHandler` remains generic over `type Output`.
- Concrete handlers use native RPITIT futures with explicit `Send`
bounds.
- `AnyToolHandler` remains the only object-safe adapter and still does
the boxing at the registry boundary, as before.
- The implementation diff is only **33 files, +28/-77**.

The measurements are at least comparable, and in this run this PR is
slightly faster than openai#16627 on the pass-level total:

| Metric | openai#16627 | This PR | Delta |
|---|---:|---:|---:|
| `total` | 79.90s | 68.98s | **-13.7%** |
| `generate_crate_metadata` | 25.88s | 24.49s | **-5.4%** |
| `monomorphization_collector_graph_walk` | 23.54s | 22.19s | **-5.7%**
|
| `evaluate_obligation` self-time | 43.29s | 46.91s | +8.4% |

`cargo +nightly build -p codex-core --lib -Z unstable-options
--timings=json` after `cargo clean -p codex-core`.

Baseline data below is reused from the shared parent `0bd31dc382bd`
profile because this PR and openai#16627 are both one commit on top of that
same parent.

| Crate | Baseline `duration` | This PR `duration` | Delta | Baseline
`rmeta_time` | This PR `rmeta_time` | Delta |
|---|---:|---:|---:|---:|---:|---:|
| `codex_core` | 187.380776583s | 69.171113833s | **-63.1%** |
174.474507208s | 55.873015583s | **-68.0%** |
| `starlark` | 17.90s | 16.773824125s | -6.3% | n/a | 8.8999965s | n/a |

`cargo +nightly rustc -p codex-core --lib -- -Z time-passes -Z
time-passes-format=json` after `cargo clean -p codex-core`.

| Pass | Baseline | This PR | Delta |
|---|---:|---:|---:|
| `total` | 187.150662083s | 68.978770375s | **-63.1%** |
| `generate_crate_metadata` | 84.531864625s | 24.487462958s | **-71.0%**
|
| `MIR_borrow_checking` | 84.131389375s | 24.575553875s | **-70.8%** |
| `monomorphization_collector_graph_walk` | 79.737515042s |
22.190207417s | **-72.2%** |
| `codegen_crate` | 12.362532292s | 12.695237625s | +2.7% |
| `type_check_crate` | 4.4765405s | 5.442019542s | +21.6% |
| `coherence_checking` | 3.311121208s | 4.239935292s | +28.0% |
| process `real` / `user` / `sys` | 187.70s / 201.87s / 4.99s | 69.52s /
85.90s / 2.92s | n/a |

`cargo +nightly rustc -p codex-core --lib -- -Z self-profile=... -Z
self-profile-events=default,query-keys,args,llvm,artifact-sizes` after
`cargo clean -p codex-core`, summarized with `measureme summarize -p
0.5`.

| Query / phase | Baseline self time | This PR self time | Delta |
Baseline total time | This PR total time | Baseline item count | This PR
item count | Baseline cache hits | This PR cache hits |
|---|---:|---:|---:|---:|---:|---:|---:|---:|---:|
| `evaluate_obligation` | 180.62s | 46.91s | **-74.0%** | 182.08s |
48.37s | 572,234 | 388,659 | 1,130,998 | 1,058,553 |
| `mir_borrowck` | 1.42s | 1.49s | +4.9% | 93.77s | 29.59s | n/a | 6,184
| n/a | 15,298 |
| `typeck` | 1.84s | 1.87s | +1.6% | 2.38s | 2.44s | n/a | 9,367 | n/a |
79,247 |
| `LLVM_module_codegen_emit_obj` | n/a | 17.12s | n/a | 17.01s | 17.12s
| n/a | 256 | n/a | 0 |
| `LLVM_passes` | n/a | 13.07s | n/a | 12.95s | 13.07s | n/a | 1 | n/a |
0 |
| `codegen_module` | n/a | 12.33s | n/a | 12.22s | 13.64s | n/a | 256 |
n/a | 0 |
| `items_of_instance` | n/a | 676.00ms | n/a | n/a | 24.96s | n/a |
99,990 | n/a | 0 |
| `type_op_prove_predicate` | n/a | 660.79ms | n/a | n/a | 24.78s | n/a
| 78,762 | n/a | 235,877 |

| Summary | Baseline | This PR |
|---|---:|---:|
| `evaluate_obligation` % of total CPU | 70.821% | 38.880% |
| self-profile total CPU time | 255.042999997s | 120.661175956s |
| process `real` / `user` / `sys` | 220.96s / 235.02s / 7.09s | 86.35s /
103.66s / 3.54s |

From the same `measureme summarize` output:

| Artifact | Baseline | This PR | Delta |
|---|---:|---:|---:|
| `crate_metadata` | 26,534,471 bytes | 26,545,248 bytes | +10,777 |
| `dep_graph` | 253,181,425 bytes | 239,240,806 bytes | -13,940,619 |
| `linked_artifact` | 565,366,624 bytes | 562,673,176 bytes | -2,693,448
|
| `object_file` | 513,127,264 bytes | 510,464,096 bytes | -2,663,168 |
| `query_cache` | 137,440,945 bytes | 136,982,566 bytes | -458,379 |
| `cgu_instructions` | 3,586,307 bytes | 3,575,121 bytes | -11,186 |
| `codegen_unit_size_estimate` | 2,084,846 bytes | 2,078,773 bytes |
-6,073 |
| `work_product_index` | 19,565 bytes | 19,565 bytes | 0 |

These are the top normalized obligation buckets from the shared baseline
profile:

| Obligation bucket | Samples | Duration |
|---|---:|---:|
| `outlives:tasks::review::ReviewTask` | 1,067 | 6.33s |
| `outlives:tools::handlers::unified_exec::UnifiedExecHandler` | 896 |
5.63s |
| `trait:T as tools::registry::ToolHandler` | 876 | 5.45s |
| `outlives:tools::handlers::shell::ShellHandler` | 888 | 5.37s |
| `outlives:tools::handlers::shell::ShellCommandHandler` | 870 | 5.29s |
|
`outlives:tools::runtimes::shell::unix_escalation::CoreShellActionProvider`
| 637 | 3.73s |
| `outlives:tools::handlers::mcp::McpHandler` | 695 | 3.61s |
| `outlives:tasks::regular::RegularTask` | 726 | 3.57s |

Top `items_of_instance` entries before this change were mostly concrete
async handler/task impls:

| Instance | Duration |
|---|---:|
| `tasks::regular::{impl#2}::run` | 3.79s |
| `tools::handlers::mcp::{impl#0}::handle` | 3.27s |
| `tools::runtimes::shell::unix_escalation::{impl#2}::determine_action`
| 3.09s |
| `tools::handlers::agent_jobs::{impl#11}::handle` | 3.07s |
| `tools::handlers::multi_agents::spawn::{impl#1}::handle` | 2.84s |
| `tasks::review::{impl#4}::run` | 2.82s |
| `tools::handlers::multi_agents_v2::spawn::{impl#2}::handle` | 2.80s |
| `tools::handlers::multi_agents::resume_agent::{impl#1}::handle` |
2.73s |
| `tools::handlers::unified_exec::{impl#2}::handle` | 2.54s |
| `tasks::compact::{impl#4}::run` | 2.45s |

Relevant pre-change registry shape:
[`codex-rs/core/src/tools/registry.rs`](https://github.com/openai/codex/blob/0bd31dc382bd1c33dc2bb6b97069c76aa10ba14b/codex-rs/core/src/tools/registry.rs#L38-L219)

Current registry shape in this PR:
[`codex-rs/core/src/tools/registry.rs`](https://github.com/openai/codex/blob/41f7ac0adeac81d667541853d6546267d6083613/codex-rs/core/src/tools/registry.rs#L38-L203)

- `ToolHandler::{is_mutating, handle}` now return native `impl Future +
Send` futures instead of using `#[async_trait]`.
- `AnyToolHandler` remains the object-safe adapter and boxes those
futures at the registry boundary with explicit lifetimes.
- Concrete handlers and the registry test handler drop `#[async_trait]`
but otherwise keep their async method bodies intact.
- Representative examples:
[`codex-rs/core/src/tools/handlers/shell.rs`](https://github.com/openai/codex/blob/41f7ac0adeac81d667541853d6546267d6083613/codex-rs/core/src/tools/handlers/shell.rs#L223-L379),
[`codex-rs/core/src/tools/handlers/unified_exec.rs`](https://github.com/openai/codex/blob/41f7ac0adeac81d667541853d6546267d6083613/codex-rs/core/src/tools/handlers/unified_exec.rs),
[`codex-rs/core/src/tools/registry_tests.rs`](https://github.com/openai/codex/blob/41f7ac0adeac81d667541853d6546267d6083613/codex-rs/core/src/tools/registry_tests.rs)

This is intentionally less invasive than openai#16627: it does **not** move
result boxing into every concrete handler and does **not** change
`ToolHandler` into an object-safe trait.

Instead, it keeps the existing registry-level type-erasure boundary and
only removes the macro-generated async wrapper layer from concrete
impls. So the runtime boxing story stays basically the same as before,
while the compile-time savings are still large.

Existing verification for this branch still applies:

- Ran `cargo test -p codex-core`; this change compiled and the suite
reached the known unrelated `config::tests::*guardian*` failures, with
no local diff under `codex-rs/core/src/config/`.

Profiling commands used for the tables above:

- `cargo clean -p codex-core`
- `cargo +nightly build -p codex-core --lib -Z unstable-options
--timings=json`
- `cargo +nightly rustc -p codex-core --lib -- -Z time-passes -Z
time-passes-format=json`
- `cargo +nightly rustc -p codex-core --lib -- -Z self-profile=... -Z
self-profile-events=default,query-keys,args,llvm,artifact-sizes`
- `measureme summarize -p 0.5`
rebroad pushed a commit to rebroad/codex that referenced this pull request Apr 16, 2026
@bolinfest @rebroad
core: cut codex-core compile time 63% with native async ToolHandler (o…
a947fc8 
…penai#16630)

`ToolHandler` was still paying a large compile-time tax from
`#[async_trait]` on every concrete handler impl, even though the only
object-safe boundary the registry actually stores is the internal
`AnyToolHandler` adapter.

This PR removes that macro-generated async wrapper layer from concrete
`ToolHandler` impls while keeping the existing object-safe shim in
`AnyToolHandler`. In practice, that gets essentially the same
compile-time win as the larger type-erasure refactor in openai#16627, but with
a much smaller diff and without changing the public shape of
`ToolHandler<Output = T>`.

That tradeoff matters here because this is a broad `codex-core` hotspot
and reviewers should be able to judge the compile-time impact from hard
numbers, not vibes.

On a clean `codex-core` package rebuild (`cargo clean -p codex-core`
before each command), rustc `total` dropped from **187.15s to 68.98s**
versus the shared `0bd31dc382bd` baseline: **-63.1%**.

The biggest hot passes dropped by roughly **71-72%**:

| Metric | Baseline `0bd31dc382bd` | This PR `41f7ac0adeac` | Delta |
|---|---:|---:|---:|
| `total` | 187.15s | 68.98s | **-63.1%** |
| `generate_crate_metadata` | 84.53s | 24.49s | **-71.0%** |
| `MIR_borrow_checking` | 84.13s | 24.58s | **-70.8%** |
| `monomorphization_collector_graph_walk` | 79.74s | 22.19s | **-72.2%**
|
| `evaluate_obligation` self-time | 180.62s | 46.91s | **-74.0%** |

Important caveat: `-Z time-passes` timings are nested, so
`generate_crate_metadata` and `monomorphization_collector_graph_walk`
are mostly overlapping, not additive.

hotspot, but it got there by making `ToolHandler` object-safe and
changing every handler to return `BoxFuture<Result<AnyToolResult, _>>`
directly.

This PR keeps the lower-churn shape:

- `ToolHandler` remains generic over `type Output`.
- Concrete handlers use native RPITIT futures with explicit `Send`
bounds.
- `AnyToolHandler` remains the only object-safe adapter and still does
the boxing at the registry boundary, as before.
- The implementation diff is only **33 files, +28/-77**.

The measurements are at least comparable, and in this run this PR is
slightly faster than openai#16627 on the pass-level total:

| Metric | openai#16627 | This PR | Delta |
|---|---:|---:|---:|
| `total` | 79.90s | 68.98s | **-13.7%** |
| `generate_crate_metadata` | 25.88s | 24.49s | **-5.4%** |
| `monomorphization_collector_graph_walk` | 23.54s | 22.19s | **-5.7%**
|
| `evaluate_obligation` self-time | 43.29s | 46.91s | +8.4% |

`cargo +nightly build -p codex-core --lib -Z unstable-options
--timings=json` after `cargo clean -p codex-core`.

Baseline data below is reused from the shared parent `0bd31dc382bd`
profile because this PR and openai#16627 are both one commit on top of that
same parent.

| Crate | Baseline `duration` | This PR `duration` | Delta | Baseline
`rmeta_time` | This PR `rmeta_time` | Delta |
|---|---:|---:|---:|---:|---:|---:|
| `codex_core` | 187.380776583s | 69.171113833s | **-63.1%** |
174.474507208s | 55.873015583s | **-68.0%** |
| `starlark` | 17.90s | 16.773824125s | -6.3% | n/a | 8.8999965s | n/a |

`cargo +nightly rustc -p codex-core --lib -- -Z time-passes -Z
time-passes-format=json` after `cargo clean -p codex-core`.

| Pass | Baseline | This PR | Delta |
|---|---:|---:|---:|
| `total` | 187.150662083s | 68.978770375s | **-63.1%** |
| `generate_crate_metadata` | 84.531864625s | 24.487462958s | **-71.0%**
|
| `MIR_borrow_checking` | 84.131389375s | 24.575553875s | **-70.8%** |
| `monomorphization_collector_graph_walk` | 79.737515042s |
22.190207417s | **-72.2%** |
| `codegen_crate` | 12.362532292s | 12.695237625s | +2.7% |
| `type_check_crate` | 4.4765405s | 5.442019542s | +21.6% |
| `coherence_checking` | 3.311121208s | 4.239935292s | +28.0% |
| process `real` / `user` / `sys` | 187.70s / 201.87s / 4.99s | 69.52s /
85.90s / 2.92s | n/a |

`cargo +nightly rustc -p codex-core --lib -- -Z self-profile=... -Z
self-profile-events=default,query-keys,args,llvm,artifact-sizes` after
`cargo clean -p codex-core`, summarized with `measureme summarize -p
0.5`.

| Query / phase | Baseline self time | This PR self time | Delta |
Baseline total time | This PR total time | Baseline item count | This PR
item count | Baseline cache hits | This PR cache hits |
|---|---:|---:|---:|---:|---:|---:|---:|---:|---:|
| `evaluate_obligation` | 180.62s | 46.91s | **-74.0%** | 182.08s |
48.37s | 572,234 | 388,659 | 1,130,998 | 1,058,553 |
| `mir_borrowck` | 1.42s | 1.49s | +4.9% | 93.77s | 29.59s | n/a | 6,184
| n/a | 15,298 |
| `typeck` | 1.84s | 1.87s | +1.6% | 2.38s | 2.44s | n/a | 9,367 | n/a |
79,247 |
| `LLVM_module_codegen_emit_obj` | n/a | 17.12s | n/a | 17.01s | 17.12s
| n/a | 256 | n/a | 0 |
| `LLVM_passes` | n/a | 13.07s | n/a | 12.95s | 13.07s | n/a | 1 | n/a |
0 |
| `codegen_module` | n/a | 12.33s | n/a | 12.22s | 13.64s | n/a | 256 |
n/a | 0 |
| `items_of_instance` | n/a | 676.00ms | n/a | n/a | 24.96s | n/a |
99,990 | n/a | 0 |
| `type_op_prove_predicate` | n/a | 660.79ms | n/a | n/a | 24.78s | n/a
| 78,762 | n/a | 235,877 |

| Summary | Baseline | This PR |
|---|---:|---:|
| `evaluate_obligation` % of total CPU | 70.821% | 38.880% |
| self-profile total CPU time | 255.042999997s | 120.661175956s |
| process `real` / `user` / `sys` | 220.96s / 235.02s / 7.09s | 86.35s /
103.66s / 3.54s |

From the same `measureme summarize` output:

| Artifact | Baseline | This PR | Delta |
|---|---:|---:|---:|
| `crate_metadata` | 26,534,471 bytes | 26,545,248 bytes | +10,777 |
| `dep_graph` | 253,181,425 bytes | 239,240,806 bytes | -13,940,619 |
| `linked_artifact` | 565,366,624 bytes | 562,673,176 bytes | -2,693,448
|
| `object_file` | 513,127,264 bytes | 510,464,096 bytes | -2,663,168 |
| `query_cache` | 137,440,945 bytes | 136,982,566 bytes | -458,379 |
| `cgu_instructions` | 3,586,307 bytes | 3,575,121 bytes | -11,186 |
| `codegen_unit_size_estimate` | 2,084,846 bytes | 2,078,773 bytes |
-6,073 |
| `work_product_index` | 19,565 bytes | 19,565 bytes | 0 |

These are the top normalized obligation buckets from the shared baseline
profile:

| Obligation bucket | Samples | Duration |
|---|---:|---:|
| `outlives:tasks::review::ReviewTask` | 1,067 | 6.33s |
| `outlives:tools::handlers::unified_exec::UnifiedExecHandler` | 896 |
5.63s |
| `trait:T as tools::registry::ToolHandler` | 876 | 5.45s |
| `outlives:tools::handlers::shell::ShellHandler` | 888 | 5.37s |
| `outlives:tools::handlers::shell::ShellCommandHandler` | 870 | 5.29s |
|
`outlives:tools::runtimes::shell::unix_escalation::CoreShellActionProvider`
| 637 | 3.73s |
| `outlives:tools::handlers::mcp::McpHandler` | 695 | 3.61s |
| `outlives:tasks::regular::RegularTask` | 726 | 3.57s |

Top `items_of_instance` entries before this change were mostly concrete
async handler/task impls:

| Instance | Duration |
|---|---:|
| `tasks::regular::{impl#2}::run` | 3.79s |
| `tools::handlers::mcp::{impl#0}::handle` | 3.27s |
| `tools::runtimes::shell::unix_escalation::{impl#2}::determine_action`
| 3.09s |
| `tools::handlers::agent_jobs::{impl#11}::handle` | 3.07s |
| `tools::handlers::multi_agents::spawn::{impl#1}::handle` | 2.84s |
| `tasks::review::{impl#4}::run` | 2.82s |
| `tools::handlers::multi_agents_v2::spawn::{impl#2}::handle` | 2.80s |
| `tools::handlers::multi_agents::resume_agent::{impl#1}::handle` |
2.73s |
| `tools::handlers::unified_exec::{impl#2}::handle` | 2.54s |
| `tasks::compact::{impl#4}::run` | 2.45s |

Relevant pre-change registry shape:
[`codex-rs/core/src/tools/registry.rs`](https://github.com/openai/codex/blob/0bd31dc382bd1c33dc2bb6b97069c76aa10ba14b/codex-rs/core/src/tools/registry.rs#L38-L219)

Current registry shape in this PR:
[`codex-rs/core/src/tools/registry.rs`](https://github.com/openai/codex/blob/41f7ac0adeac81d667541853d6546267d6083613/codex-rs/core/src/tools/registry.rs#L38-L203)

- `ToolHandler::{is_mutating, handle}` now return native `impl Future +
Send` futures instead of using `#[async_trait]`.
- `AnyToolHandler` remains the object-safe adapter and boxes those
futures at the registry boundary with explicit lifetimes.
- Concrete handlers and the registry test handler drop `#[async_trait]`
but otherwise keep their async method bodies intact.
- Representative examples:
[`codex-rs/core/src/tools/handlers/shell.rs`](https://github.com/openai/codex/blob/41f7ac0adeac81d667541853d6546267d6083613/codex-rs/core/src/tools/handlers/shell.rs#L223-L379),
[`codex-rs/core/src/tools/handlers/unified_exec.rs`](https://github.com/openai/codex/blob/41f7ac0adeac81d667541853d6546267d6083613/codex-rs/core/src/tools/handlers/unified_exec.rs),
[`codex-rs/core/src/tools/registry_tests.rs`](https://github.com/openai/codex/blob/41f7ac0adeac81d667541853d6546267d6083613/codex-rs/core/src/tools/registry_tests.rs)

This is intentionally less invasive than openai#16627: it does **not** move
result boxing into every concrete handler and does **not** change
`ToolHandler` into an object-safe trait.

Instead, it keeps the existing registry-level type-erasure boundary and
only removes the macro-generated async wrapper layer from concrete
impls. So the runtime boxing story stays basically the same as before,
while the compile-time savings are still large.

Existing verification for this branch still applies:

- Ran `cargo test -p codex-core`; this change compiled and the suite
reached the known unrelated `config::tests::*guardian*` failures, with
no local diff under `codex-rs/core/src/config/`.

Profiling commands used for the tables above:

- `cargo clean -p codex-core`
- `cargo +nightly build -p codex-core --lib -Z unstable-options
--timings=json`
- `cargo +nightly rustc -p codex-core --lib -- -Z time-passes -Z
time-passes-format=json`
- `cargo +nightly rustc -p codex-core --lib -- -Z self-profile=... -Z
self-profile-events=default,query-keys,args,llvm,artifact-sizes`
- `measureme summarize -p 0.5`
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

1 participant

@bolinfest