[wasm][coreclr] Dispatch R2R-compiled UnmanagedCallersOnly callbacks to their native entrypoint by pavelsavara · Pull Request #134355 · dotnet/runtime

pavelsavara · 2026-09-21T17:23:03Z

Problem

Under a partial ReadyToRun image, an [UnmanagedCallersOnly] callback can be crossgen'd to native (R2R) code. The generated native-to-interpreter reverse thunk (callhelpers-reverse.cpp) routed it unconditionally through ExecuteInterpretedMethodFromUnmanaged, which:

handed a null interpreter body to the interpreter, causing Fatal error. Unimplemented or invalid interpreter opcode (INTOP_INVALID), or
via InvokeManagedMethod, hit the interp-to-R2R RuntimeError: null function or function signature mismatch trap ([wasm][R2R] Interpreter→R2R call to a shared-generic method with no R2R body traps with "null function or function signature mismatch" at startup #134200).

The design already intends R2R-compiled [UnmanagedCallersOnly] methods to be dispatched directly to their native entrypoint (GetUnmanagedCallersOnlyThunk does this), but generated export reverse thunks bypassed that path. The browser event loop calls native C exports directly, for example SystemJS_ExecuteBackgroundJobCallback to System.Threading.ThreadPool.BackgroundJobHandler, which can be R2R-compiled under a partial image.

Fix

The call-helpers generator now emits a direct-R2R dispatch in each supported reverse thunk: resolve and cache the R2R native entrypoint with GetR2RNativeCodeForUnmanagedCallersOnly, call it with the native ABI, and fall back to the interpreter path only when there is no R2R body.

void* r2r = __atomic_load_n(&R2RCode_..., __ATOMIC_ACQUIRE);
if (r2r == (void*)(intptr_t)-1)
{
    r2r = GetR2RNativeCodeForUnmanagedCallersOnly(MD_...);
    __atomic_store_n(&R2RCode_..., r2r, __ATOMIC_RELEASE);
}
if (r2r != nullptr)
{
    ((void(*)())r2r)();
    return;
}
ExecuteInterpretedMethodFromUnmanaged(MD_..., ...);

Additional hardening from review:

Multi-slot UCO signatures such as Int128 parameters or returns are explicitly rejected before generating a wrapper whose C function type would not match the lowered WASM signature.
Vector128<T> parameters and returns using the v128 ABI are explicitly rejected because the generated C wrapper does not yet model that value type.
Aggregate returns that require a hidden return buffer are explicitly rejected because the generated native wrapper does not yet model that ABI shape.
Cached entrypoint publication emits acquire/release atomic builtins directly, so VM-compiled and app-linked helpers have identical thread-safe behavior.
HasNativeEntryPoint excludes interpreter-preferred entrypoints, so the generated interpreter fallback thunk cannot be returned recursively as R2R code.

GetUnmanagedCallersOnlyThunk is refactored onto the same shared lookup helper. The checked-in browser and WASI call helpers are regenerated.

End-to-end coverage

WasmInterpreterTransitions now exercises:

R2R managed code -> P/Invoke -> echo.c -> generated UCO export -> R2R managed callback

The native round trips cover:

int32_t -> int32_t
mixed int64_t, float, double -> double
int32_t*, int32_t -> void
a single-field int32_t struct return, scalarized to i32 by the WASM C ABI

Runtime-test support now compiles raw .c/.cpp NativeFileReference items, includes their module names in generated P/Invoke tables, and links their objects into the test-specific browser corerun.

Validation

The original partial-R2R repro crashed before the fix and runs to completion after it.
Removing only the generated direct-R2R branch makes the new native round-trip test fail at ExecuteInterpretedMethodFromUnmanaged with targetIp != NULL; restoring it passes.
Browser/CoreCLR clr+libs+host build: zero warnings and errors.
ILCompiler.ReadyToRun.Tests WasmArgumentLayoutTests: 72 passed, 0 failed.
Canonical Browser and WASI helper regeneration completed; both generated reverse-helper sets use direct atomic publication.
Browser/CoreCLR runtime rebuild and WASI/CoreCLR runtime compilation with the regenerated helpers: zero warnings and errors.
Targeted Browser/CoreCLR WasmInterpreterTransitions build: zero warnings and errors.
Node execution of WasmInterpreterTransitions: expected 100, actual 100.

Main files

src/coreclr/tools/aot/ILCompiler.ReadyToRun/PortableCallHelpers/PInvokeTableGenerator.cs - emits direct R2R dispatch and rejects unsupported multi-slot callbacks
src/coreclr/vm/wasm/helpers.cpp - shared R2R UCO entrypoint lookup
src/mono/browser/build/coreclr_compat.h - app-link compatibility declarations for generated helpers
src/coreclr/vm/wasm/browser/callhelpers-reverse.cpp, src/coreclr/vm/wasm/wasi/callhelpers-reverse.cpp - regenerated helpers
src/tests/Common/CLRTest.WasmCorerun.targets - native source support for Browser/CoreCLR runtime tests
src/tests/readytorun/wasm/WasmInterpreterTransitions/ - managed and native end-to-end coverage

…to their native entrypoint A partial R2R image can crossgen an [UnmanagedCallersOnly] callback to native code. The generated native->interpreter reverse thunk routed it unconditionally through ExecuteInterpretedMethodFromUnmanaged, which either handed a null interpreter body to the interpreter (fatal 'Unimplemented or invalid interpreter opcode') or, via InvokeManagedMethod, hit the interp->R2R 'null function or function signature mismatch' trap (dotnet#134200). Repro: the browser event loop calls the SystemJS_ExecuteBackgroundJobCallback export -> ThreadPool.BackgroundJobHandler, which is R2R-compiled under a partial image. Fix: the call-helpers generator now emits a direct-R2R dispatch in each reverse thunk - resolve the R2R native entrypoint via GetR2RNativeCodeForUnmanagedCallersOnly and call it with the native ABI, falling back to the interpreter only when there is no R2R body. GetUnmanagedCallersOnlyThunk is refactored onto the same shared helper. Regenerated the checked-in browser and wasi call helpers.

azure-pipelines · 2026-09-21T17:23:49Z

Azure Pipelines: Successfully started running 3 pipeline(s). 13 pipeline(s) were filtered out due to trigger conditions. There may be pipelines that require an authorized user to comment /azp run to run.

dotnet-policy-service · 2026-09-21T17:24:17Z

Tagging subscribers to this area: @JulieLeeMSFT, @BrzVlad, @janvorli
See info in area-owners.md if you want to be subscribed.

dotnet-policy-service · 2026-09-21T17:24:19Z

Tagging subscribers to 'arch-wasm': @lewing, @pavelsavara
See info in area-owners.md if you want to be subscribed.

lewing · 2026-09-21T18:58:26Z

you'll want to remove

runtime/src/libraries/tests.proj

Lines 91 to 93 in 3cd7514

    
           <!-- ActiveIssue https://github.com/dotnet/runtime/issues/134200 --> 
        
           <ProjectExclusions Include="$(MSBuildThisFileDirectory)Microsoft.CSharp\tests\Microsoft.CSharp.Tests.csproj" /> 
        
           <ProjectExclusions Include="$(MSBuildThisFileDirectory)System.Runtime\tests\System.Dynamic.Runtime.Tests\System.Dynamic.Runtime.Tests.csproj" />

pavelsavara · 2026-09-22T09:25:54Z

you'll want to remove
#134200

That's still failing after this PR, it's slightly different.
See #134200 (comment)

Copilot

Copilot review overview

🟡 Changes recommended

Address the lazy-R2R retry issue and add generator-level coverage for direct dispatch generation.

Get a fresh assessment by requesting another Copilot review.

Review effort: Lite
Findings: 1

Open (1)

Add generator tests for emitted direct R2R dispatch · New

What changed in this PR

Fixes WebAssembly reverse thunks so R2R-compiled [UnmanagedCallersOnly] callbacks dispatch directly to native entrypoints, with interpreter fallback.

Changes:

Adds shared R2R entrypoint resolution.
Updates callback generation and regenerates WASI helpers.
Preserves interpreter fallback for non-R2R callbacks.

File	Summary
`src/coreclr/vm/wasm/wasi/callhelpers-reverse.cpp`	Regenerated WASI reverse callback helpers.
`src/coreclr/vm/wasm/helpers.cpp`	Adds shared R2R callback resolution. Moderate concern (1 vote): lazy R2R may not be retried after interpreter preparation.
`src/coreclr/tools/aot/ILCompiler.ReadyToRun/PortableCallHelpers/PInvokeTableGenerator.cs`	Emits direct R2R dispatch. Nit (3 votes): add generator coverage for emitted R2R and fallback branches.

Addresses PR review: cache GetR2RNativeCodeForUnmanagedCallersOnly result in a per-callback static (like MD_*) so it runs at most once, and move the blank line to separate the R2R fast path from the interpreter fallback.

Copilot

Copilot review overview

🟡 Changes recommended

Critical ABI and concurrency findings, plus missing regression coverage, remain unresolved.

Get a fresh assessment by requesting another Copilot review.

Review effort: Lite
Findings: 2

Open (2)

Handle hidden return-buffer parameter in direct Wasm calls · New
Synchronize concurrent R2R function-pointer publication · New

Resolved since last review (1)

Add generator tests for emitted direct R2R dispatch

…sh cache Addresses PR review: skip the direct R2R call for callbacks whose return is passed by a hidden buffer (their wasm function type would not match the declared cast), and publish the cached entrypoint with VolatileLoad/VolatileStore so concurrent first calls cannot tear the pointer.

Copilot

Copilot review overview

🟡 Changes recommended

Direct dispatch can mishandle multi-slot Wasm signatures, and null R2R results may be cached permanently.

Get a fresh assessment by requesting another Copilot review.

Review effort: Lite
Findings: 1

Open (1)

Handle multi-slot aggregate parameters and returns in Wasm callbacks · New

Resolved since last review (2)

Synchronize concurrent R2R function-pointer publication
Handle hidden return-buffer parameter in direct Wasm calls

Copilot

Copilot review overview

🟡 Changes recommended

One or more issues must be addressed before approval.

Get a fresh assessment by requesting another Copilot review.

Review effort: Lite
Findings: 1

Open (1)

Interpreter fallback uses incorrect ABI for by-reference struct returns · New

Resolved since last review (3)

Threaded entrypoint cache publication uses non-atomic accesses
Fallback reverse thunk can recurse instead of interpreting
Handle multi-slot aggregate parameters and returns in Wasm callbacks

Reject UnmanagedCallersOnly callbacks whose aggregate return requires a hidden return buffer instead of emitting an ABI-incompatible native wrapper, and cover the diagnostic in generator tests. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Copilot

Copilot review overview

🟡 Changes recommended

Critical ABI validation is missing for v128 callback signatures, risking invalid calls or corrupted data.

Get a fresh assessment by requesting another Copilot review.

Review effort: Lite
Findings: 1

Open (1)

Reject unsupported v128 callback signatures to prevent ABI mismatches · New

Resolved since last review (1)

Interpreter fallback uses incorrect ABI for by-reference struct returns

Reject Vector128 UnmanagedCallersOnly parameters and returns before generating native wrappers whose C declarations cannot represent the v128 ABI, and cover both signature positions. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Copilot

Copilot review overview

🟡 Changes recommended

One or more issues must be addressed before approval.

Get a fresh assessment by requesting another Copilot review.

Review effort: Lite
Findings: 1

Open (1)

Fix WASM volatile helpers to preserve acquire/release ordering · New

Resolved since last review (1)

Reject unsupported v128 callback signatures to prevent ABI mismatches

Generate acquire/release builtins directly so both VM-compiled and app-linked reverse helpers publish cached R2R entrypoints atomically. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Copilot

Copilot review overview

🟡 Changes recommended

One or more issues must be addressed before approval.

Get a fresh assessment by requesting another Copilot review.

Review effort: Lite
Findings: 1 · 1

Open (2)

Use separate compiler flags for user native sources · New
Update comment to describe acquire/release atomics · New

Resolved since last review (1)

Fix WASM volatile helpers to preserve acquire/release ordering

User NativeFileReference .c/.cpp now use a dedicated corerun-compile-user.rsp that carries only the shared compile flags (optimization, exception model, SIMD, GEN_PINVOKE) and omits the coreclr_compat.h force-include and kit header path, mirroring the app build's _EmccCFlags/_EmccCFlagsGenerated split so raw sources don't inherit CoreCLR typedefs/macros. Also correct a stale generator comment that still referenced a volatile store. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Copilot

Copilot review overview

🔵 Needs a closer look

The changes span runtime dispatch, generated native ABI wrappers, and test-host linking paths, warranting final human review.

Review effort: Lite
Findings: None

Resolved since last review (2)

Use separate compiler flags for user native sources
Update comment to describe acquire/release atomics

The generated wasm reverse thunk cannot model hidden-return-buffer struct-returning UnmanagedCallersOnly callbacks (GetSize -> SizeF), so crossgen2 now rejects them and fails the whole test build leg. Opt these four callconv tests out of the test-specific corerun on browser-coreclr, matching the existing EmptyThisCallTest precedent (dotnet#131811). Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Copilot

Copilot review overview

🟡 Changes recommended

One or more issues must be addressed before approval.

Get a fresh assessment by requesting another Copilot review.

Review effort: Lite
Findings: 1

Open (1)

Make MethodDesc publication thread-safe · New

…134686) ## Problem R2R Wasm code keeps the shadow SP in a local and leaves the `__stack_pointer` global stale. `JIT_PInvokeBegin` and the SuppressGCTransition publish (#130924) set `__stack_pointer` to the caller's shadow SP and leave it there. So when a reverse P/Invoke (`UnmanagedCallersOnly`) method's body, or any R2R code it calls, did an inlined P/Invoke, the method returned to its native caller with the global still lowered. That breaks the native ABI, which requires `__stack_pointer` to be restored on return. In the observed case the P/Invoke was CoreLib's `CastHelpers.ChkCastAny_NoCacheLookup` under `GetUserData<T>`. When the UCO was reached through an R2R inlined P/Invoke (a `delegate* unmanaged` calli), the Debug check in `JIT_PInvokeEndImpl` fired: `sp == stack_pointer_global_value` at `src/coreclr/vm/wasm/helpers.cpp:420`. After that came GetFrame asserts and an out-of-bounds access. See #134681 for the full analysis. The bug is reachable on main whenever an R2R method calls a UCO that itself does an inlined P/Invoke (see the test below). #134606 exposes it more broadly by making `TestEntryPoint` in `readytorun/wasm/WasmInterpreterTransitions` run as R2R instead of interpreted, which also routes the existing `StreamLengthProxy` case through R2R. ## Fix In `CodeGen::genFnEpilog` (`codegenwasm.cpp`), reverse P/Invoke methods now emit this sequence before `return`/`end`: ``` local.get <FP local, or SP local if there is no frame pointer> i32.const genTotalFrameSize() i32.add global.set __stack_pointer ``` This restores the global to its entry value, the post-prolog SP plus the frame size. The sequence leaves the Wasm operand stack unchanged, so an already-pushed return value is preserved. It replaces the `TODO-WASM: shadow stack maintenance` comment. The emit pattern matches the SuppressGCTransition publish from #130924. ## Test `WasmInterpreterTransitions` gains `R2RCallsNestingUco`: an R2R method does a `delegate* unmanaged` calli into `UcoWithInlinedPInvoke`, which does a calli into a second UCO, `UcoLeaf`. Each calli is an inlined P/Invoke. The inner one lowers `__stack_pointer`, and the outer one's `JIT_PInvokeEnd` asserts that the global is back at the caller's SP. Run against the Checked browser Core_Root from main CI build 1612746 (without #134606), with crossgen2 and the universal wasm JIT built from main: | test | main JIT | fixed JIT | |---|---|---| | main `WasmInterpreterTransitions` | pass | pass | | with `R2RCallsNestingUco` | **assert** `sp == stack_pointer_global_value` (helpers.cpp:420), then GetFrame assert | pass (exit 100) | The test IL for this run was compiled locally with csc rather than through the repo's test build. ## Validation - Rebuilt `clrjit_universal_wasm_arm64`, swapped it into crossgen2, and recompiled `WasmInterpreterTransitions` against the CI Checked browser runtime from build 1612477, which includes #134606's runtime changes. - Old JIT: reproduces the CI failure (the helpers.cpp:420 assert, then GetFrame asserts and OOB). - New JIT: passes with exit 100 and no asserts. - Before the new test was added, a disassembly diff of `composite-r2r.wasm` showed exactly two changes: the two UCO epilogs (`StreamLengthProxy` and `UnmanagedCallerCallsInterpreted`), each with the 4-instruction restore added. Everything else was identical. - The other 6 tests in the Helix `readytorun` work item pass: readytorun, fieldlayout, crossgen2smoke_donotalwaysusecrossgen2, Breadth1Test, Depth1Test, DynamicMethodGCStress. - `./build.sh clr.alljits -c checked` on osx-arm64, which includes the universal wasm JIT, builds cleanly, and jit-format reports no changes. ## Related - Unblocks the `WasmInterpreterTransitions` failure in #134606. This PR is independent of #134606 and has its own regression test on main, and #134606 doesn't need to be stacked on it. - Helps the open #134355, where native callers dispatch directly to R2R UCO bodies. There the UCO epilog is the only place that can restore `__stack_pointer` for the native caller. Resolves #134681 > [!NOTE] > This PR description was generated with the help of GitHub Copilot. --------- Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com> Co-authored-by: Andy Ayers <andya@microsoft.com> Copilot-Session: eea66043-479c-4397-ae20-f4c776bc3782

radekdoulik

The MD assignment is pre-existing and can be done as followup.

lewing · 2026-09-26T17:17:32Z

+                // Cache the resolved entrypoint in a per-callback static, published with acquire/release
+                // atomics: these are native entry points that can be entered concurrently, and the value is
+                // computed identically on every call, so the racing read/write is benign but must not tear.
+                string r2rStaticDecl = $"{w.NewLine}static void* {r2rVar} = (void*)(intptr_t)-1;";


could the new r2r section stuff not be inline below where it is much more readable?

github-actions Bot added the area-CodeGen-Interpreter-coreclr label Sep 21, 2026

dotnet-policy-service Bot assigned pavelsavara Sep 21, 2026

pavelsavara added the arch-wasm WebAssembly architecture label Sep 21, 2026

pavelsavara requested review from lewing and radekdoulik September 21, 2026 17:24

pavelsavara added this to the 12.0.0 milestone Sep 21, 2026

Merge branch 'main' into wasm_UCO_TO_R2R_dispatch

29a46ff

pavelsavara marked this pull request as ready for review September 22, 2026 09:25

Copilot AI lite review requested due to automatic review settings September 22, 2026 09:25

Copilot started reviewing on behalf of pavelsavara September 22, 2026 09:26 View session

Copilot AI reviewed Sep 22, 2026

View reviewed changes

Comment thread src/coreclr/tools/aot/ILCompiler.ReadyToRun/PortableCallHelpers/PInvokeTableGenerator.cs Outdated

radekdoulik reviewed Sep 22, 2026

View reviewed changes

Comment thread src/coreclr/vm/wasm/browser/callhelpers-reverse.cpp Outdated

radekdoulik reviewed Sep 22, 2026

View reviewed changes

Comment thread src/coreclr/vm/wasm/browser/callhelpers-reverse.cpp

pavelsavara added 2 commits September 23, 2026 09:59

Merge branch 'main' into wasm_UCO_TO_R2R_dispatch

c895bf4

Copilot AI review requested due to automatic review settings September 23, 2026 08:41

Merge branch 'main' into wasm_UCO_TO_R2R_dispatch

fa2adc5

Copilot started reviewing on behalf of pavelsavara September 23, 2026 08:41 View session

Copilot AI reviewed Sep 23, 2026

View reviewed changes

Comment thread src/coreclr/tools/aot/ILCompiler.ReadyToRun/PortableCallHelpers/PInvokeTableGenerator.cs Outdated

Comment thread src/coreclr/tools/aot/ILCompiler.ReadyToRun/PortableCallHelpers/PInvokeTableGenerator.cs Outdated

Copilot AI review requested due to automatic review settings September 23, 2026 10:44

Copilot started reviewing on behalf of pavelsavara September 23, 2026 10:44 View session

Copilot AI reviewed Sep 23, 2026

View reviewed changes

Comment thread src/coreclr/tools/aot/ILCompiler.ReadyToRun/PortableCallHelpers/PInvokeTableGenerator.cs

build-analysis Bot mentioned this pull request Sep 23, 2026

[ci-scan] Build break: unable to pull Helix prerequisite image #134176

Open

Copilot AI reviewed Sep 24, 2026

View reviewed changes

Comment thread src/coreclr/tools/aot/ILCompiler.ReadyToRun/PortableCallHelpers/PInvokeTableGenerator.cs Outdated

Copilot AI review requested due to automatic review settings September 24, 2026 15:01

Copilot started reviewing on behalf of pavelsavara September 24, 2026 15:02 View session

Copilot AI reviewed Sep 24, 2026

View reviewed changes

Comment thread src/coreclr/tools/aot/ILCompiler.ReadyToRun/PortableCallHelpers/PInvokeTableGenerator.cs

Copilot AI review requested due to automatic review settings September 24, 2026 15:58

Copilot started reviewing on behalf of pavelsavara September 24, 2026 15:59 View session

Copilot AI reviewed Sep 24, 2026

View reviewed changes

Comment thread src/coreclr/tools/aot/ILCompiler.ReadyToRun/PortableCallHelpers/PInvokeTableGenerator.cs Outdated

Emit atomic wasm R2R callback cache operations

dfdc276

Generate acquire/release builtins directly so both VM-compiled and app-linked reverse helpers publish cached R2R entrypoints atomically. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Copilot AI review requested due to automatic review settings September 24, 2026 16:42

Copilot started reviewing on behalf of pavelsavara September 24, 2026 16:43 View session

Copilot AI reviewed Sep 24, 2026

View reviewed changes

Comment thread src/tests/Common/CLRTest.WasmCorerun.targets Outdated

Comment thread src/coreclr/tools/aot/ILCompiler.ReadyToRun/PortableCallHelpers/PInvokeTableGenerator.cs Outdated

Copilot AI review requested due to automatic review settings September 25, 2026 10:20

Copilot started reviewing on behalf of pavelsavara September 25, 2026 10:21 View session

pavelsavara mentioned this pull request Sep 25, 2026

[wasm][R2R] Fix closed static delegate struct returns #134108

Merged

Copilot AI reviewed Sep 25, 2026

View reviewed changes

radekdoulik reviewed Sep 25, 2026

View reviewed changes

Comment thread src/coreclr/tools/aot/ILCompiler.ReadyToRun/PortableCallHelpers/PInvokeTableGenerator.cs

build-analysis Bot mentioned this pull request Sep 25, 2026

System.Net.Quic.QuicException : Connection handshake was canceled due to the configured timeout #133645

Open

Copilot AI review requested due to automatic review settings September 25, 2026 16:08

Copilot started reviewing on behalf of pavelsavara September 25, 2026 16:09 View session

Copilot AI reviewed Sep 25, 2026

View reviewed changes

Comment thread src/coreclr/tools/aot/ILCompiler.ReadyToRun/PortableCallHelpers/PInvokeTableGenerator.cs

build-analysis Bot mentioned this pull request Sep 25, 2026

Async2TaskAdapters.TestValueTaskOfTaskValueVersusAsync fails with SanityCheck() in Wasm R2R test #133953

Open

This was referenced Sep 25, 2026

[wasm][coreclr] R2R UnmanagedCallersOnly methods return with __stack_pointer lowered by nested inlined P/Invokes #134681

Closed

[wasm][RyuJIT] Restore __stack_pointer in reverse P/Invoke epilogs #134686

Merged

radekdoulik approved these changes Sep 26, 2026

View reviewed changes

lewing reviewed Sep 26, 2026

View reviewed changes

Conversation

pavelsavara commented Sep 21, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Problem

Fix

End-to-end coverage

Validation

Main files

Related

Uh oh!

azure-pipelines Bot commented Sep 21, 2026

Uh oh!

dotnet-policy-service Bot commented Sep 21, 2026

Uh oh!

dotnet-policy-service Bot commented Sep 21, 2026

Uh oh!

lewing commented Sep 21, 2026

Uh oh!

pavelsavara commented Sep 22, 2026

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Copilot review overview

🟡 Changes recommended

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Copilot review overview

🟡 Changes recommended

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Copilot review overview

🟡 Changes recommended

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Copilot review overview

🟡 Changes recommended

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Copilot review overview

🟡 Changes recommended

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Copilot review overview

🟡 Changes recommended

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Copilot review overview

🟡 Changes recommended

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Copilot review overview

🔵 Needs a closer look

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Copilot review overview

🟡 Changes recommended

Uh oh!

Uh oh!

radekdoulik left a comment

Choose a reason for hiding this comment

pavelsavara commented Sep 21, 2026 •

edited

Loading