[OMM] Implement front end diagnostics for OMM, including on TraceRayInline #7156

bob80905 · 2025-02-24T21:52:08Z

This PR addresses the front end part of OMM, defining the new flags defined in the spec, and implementing the relevant diagnostics should the flags be incompatible. It also adds the second template argument to the RayQuery object, which is set to have a default value of 0 if no explicit template argument is provided.
Fixes #7145

github-actions · 2025-02-26T19:04:27Z

✅ With the latest revision this PR passed the C/C++ code formatter.

damyanp

Couple of questions, but otherwise looks good to my untrained eye.

tools/clang/lib/Sema/SemaHLSL.cpp

tools/clang/lib/Sema/SemaHLSLDiagnoseTU.cpp

damyanp

Looking good - a few style nits and I think I spotted a few places where there's some unwanted copies happening.

tools/clang/lib/Sema/SemaHLSL.cpp

damyanp · 2025-02-28T02:07:10Z

tools/clang/lib/Sema/SemaHLSL.cpp

@@ -14161,6 +14161,97 @@ static bool IsUsageAttributeCompatible(AttributeList::Kind kind, bool &usageIn,
  return true;
 }

+bool DiagnoseRayQueryObject(QualType qt, Declarator &D, Sema &S) {


Suggested change

bool DiagnoseRayQueryObject(QualType qt, Declarator &D, Sema &S) {

static bool DiagnoseRayQueryObject(QualType qt, Declarator &D, Sema &S) {

tools/clang/lib/Sema/SemaHLSL.cpp

…iler into OMM01

tex3d

Partial review, see comments so far.

tex3d · 2025-03-04T01:06:45Z

lib/DxcSupport/FileIOHelper.cpp

@@ -1369,6 +1369,8 @@ class ReadOnlyBlobStream : public IStream {
      return E_POINTER;
    ULONG cbLeft = m_size - m_offset;
    *pcbRead = std::min(cb, cbLeft);
+    if (m_pMemory + m_offset == 0)
+      return S_FALSE;


Can you explain why this change is made at this time? Was m_pMemory uninitialized (nullptr), when attempting to read? If so, what code path hit this?

Can this just check m_pMemory for nullptr and abort with E_FAIL or something like that earlier, since we shouldn't find ourselves reading from an uninitialized stream object in the first place?

If we do want to return S_FALSE for this case, we could still check m_pMemory for nullpltr earlier and just explicitly set *pcbRead = 0, since it must be set that way for this code path. I realize that if other assumptions hold for an uninitialized object, m_offset should be zero (hopefully), and m_size should be zero, so cbLeft should end up being zero, which will set the *pcbRead to zero, but why bother with all those assumptive dependencies when m_pMemory is nullpltr?

In any case, we should have a unit test for this change. Perhaps this change can be a separate bugfix PR?

I believe there was some CompileTest that was failing, because the addition resulted in 0, so it was being treated as a nullptr when passed to memcpy. But I am no longer seeing this. I'll undo these 2 lines, and if it pops back up again I'll put up a separate PR.

Ok, it appeared again. Here's the output for the NIX Linux Clang release pipeline:

FAIL: Clang-Unit :: HLSL/ClangHLSLTests/CompilerTest.CodeGenHashStabilityHLSL (1674 of 4075) ******************** TEST 'Clang-Unit :: HLSL/ClangHLSLTests/CompilerTest.CodeGenHashStabilityHLSL' FAILED ******************** Note: Google Test filter = CompilerTest.CodeGenHashStabilityHLSL [==========] Running 1 test from 1 test case. [----------] Global test environment set-up. [----------] 1 test from CompilerTest /home/vsts/work/1/s/lib/DxcSupport/FileIOHelper.cpp:1372:16: runtime error: null pointer passed as argument 2, which is declared to never be null /usr/include/string.h:44:28: note: nonnull attribute specified here SUMMARY: UndefinedBehaviorSanitizer: undefined-behavior /home/vsts/work/1/s/lib/DxcSupport/FileIOHelper.cpp:1372:16 ********************

This points to the fact that the sum ends up as 0, and is being treated as a nullptr.
I don't see a straightforward way to test this, as this is deep deep into the call stack, and deals with a private member variable of a stream object type.
Also, I don't think setting pcbRead to 0 will solve the issue, since m_pMemory + m_offset still sums to 0, and the test is failing because the function specified that it should never receive a nullptr as the second argument.
I'll reintroduce the original addition check, and return S_FALSE, though switching to an E_FAIL may be the right move later on. I can do some more investigating as to why this specific change serves as a unit test for this if condition.

Further detail, here is the call stack:

hlsl::ReadOnlyBlobStream::Read(void * pv, unsigned long cb, unsigned long * pcbRead) Line 1373 C++ ReadAllBytes(IStream * pStream, void * pDst, unsigned __int64 uSize) Line 82 C++ PDBReader::ReadSuperblock(MSF_SuperBlock * pSB) Line 375 C++ PDBReader::PDBReader(IMalloc * pMalloc, IStream * pStream) Line 365 C++ hlsl::pdb::LoadDataFromStream(IMalloc * pMalloc, IStream * pIStream, IDxcBlob * * ppHash, IDxcBlob * * ppContainer) Line 494 C++ hlsl::pdb::LoadDataFromStream(IMalloc * pMalloc, IStream * pIStream, IDxcBlob * * ppContainer) Line 534 C++ DxilContainerReflection::Load(IDxcBlob * pContainer) Line 409 C++ CompileForHash(hlsl::options::DxcOpts & opts, const wchar_t * CommandFileName, dxc::DxcDllSupport & DllSupport, std::vector<wchar_t const *,std::allocator<wchar_t const *>> & flags, IDxcBlob * * ppHashBlob, std::string & output) Line 345 C++ FileRunCommandPart::RunDxcHashTest(dxc::DxcDllSupport & DllSupport) Line 410 C++ FileRunCommandPart::RunHashTests(dxc::DxcDllSupport & DllSupport) Line 64 C++ FileRunTestResultImpl::RunHashTestFromCommands(const char * commands, const wchar_t * fileName) Line 1280 C++ FileRunTestResultImpl::RunHashTestFromFileCommands(const wchar_t * fileName) Line 1375 C++ FileRunTestResult::RunHashTestFromFileCommands(const wchar_t * fileName) Line 1385 C++ > CompilerTest::CodeGenTestHashFullPath(const wchar_t * fullPath) Line 403 C++

The tools\clang\test\HLSLFileCheck\hlsl\objects\RayQuery\rayquery-trace-ray-inline-errors.hlsl is causing the error. Seems like something related to not initializing the PDBReader object correctly, which should stem from the incorrect pStream initialiation on IFR(hlsl::CreateReadOnlyBlobStream(pContainer, &pStream));, located at
D:\DXC\lib\HLSL\DxilContainerReflection.cpp::408
Turns out it was initialized incorrectly because in FileCheckerTest.cpp, on line 329, IFT(pCompiler2->CompileWithDebug(pSource, CommandFileName, entry.c_str(), profile.c_str(), flags.data(), flags.size(), nullptr, 0, pIncludeHandler, &pResult, &pDebugName, &pPDBBlob));
Didn't produce any output (because the -verify flag was given, no output was produced by the compiler)
So, with no output, pCompiledBlob is empty / has no memory. This emptiness is the nullptr that gets passed along to the point where the nullptr is detected.

TLDR: the -verify file rayquery-trace-ray-inline-errors.hlsl doesn't belong in the file-check directory, it needs to go somewhere where the CompilerTest doesn't check, not in a directory where all files are expected to have compiler output. Oops!

tex3d · 2025-03-04T01:58:56Z

tools/clang/include/clang/Basic/DiagnosticSemaKinds.td

@@ -7652,8 +7652,17 @@ def err_payload_fields_is_payload_and_overqualified : Error<
  "payload field '%0' is a payload struct. Payload access qualifiers are not allowed on payload types.">;
 def warn_hlsl_payload_qualifer_dropped : Warning<
  "payload access qualifiers ignored. These are only supported for lib_6_7+ targets and lib_6_6 with with the -enable-payload-qualifiers flag.">, InGroup<HLSLPayloadAccessQualifer>;
+def warn_hlsl_unexpected_value_for_ray_query_flags_template_arg : Warning<


The warning isn't what I expected from the name. How about this?

Suggested change

def warn_hlsl_unexpected_value_for_ray_query_flags_template_arg : Warning<

def warn_hlsl_rayquery_flags_disallowed: Warning<

I also joined rayquery for consistency with object and other warning names below.

tex3d · 2025-03-04T02:01:12Z

tools/clang/include/clang/Basic/DiagnosticSemaKinds.td

@@ -7652,8 +7652,17 @@ def err_payload_fields_is_payload_and_overqualified : Error<
  "payload field '%0' is a payload struct. Payload access qualifiers are not allowed on payload types.">;
 def warn_hlsl_payload_qualifer_dropped : Warning<
  "payload access qualifiers ignored. These are only supported for lib_6_7+ targets and lib_6_6 with with the -enable-payload-qualifiers flag.">, InGroup<HLSLPayloadAccessQualifer>;
+def warn_hlsl_unexpected_value_for_ray_query_flags_template_arg : Warning<
+  "A non-zero value for the RayQueryFlags template argument requires shader model 6.9 or above.">, DefaultError;
+def warn_hlsl_unexpected_rayquery_flag : Warning<


How about:

Suggested change

def warn_hlsl_unexpected_rayquery_flag : Warning<

def warn_hlsl_rayquery_flags_conflict : Warning<

tex3d · 2025-03-04T02:02:17Z

tools/clang/include/clang/Basic/DiagnosticSemaKinds.td

@@ -7652,8 +7652,17 @@ def err_payload_fields_is_payload_and_overqualified : Error<
  "payload field '%0' is a payload struct. Payload access qualifiers are not allowed on payload types.">;
 def warn_hlsl_payload_qualifer_dropped : Warning<
  "payload access qualifiers ignored. These are only supported for lib_6_7+ targets and lib_6_6 with with the -enable-payload-qualifiers flag.">, InGroup<HLSLPayloadAccessQualifer>;
+def warn_hlsl_unexpected_value_for_ray_query_flags_template_arg : Warning<
+  "A non-zero value for the RayQueryFlags template argument requires shader model 6.9 or above.">, DefaultError;


These may be DefaultError, but are kind of pointless as warnings unless you add them to a warning group to allow overriding of the setting with -W* options.

Ok, I've put the two warnings of interest into the HLSLAvailabilityConstant group

tex3d · 2025-03-04T02:09:50Z

tools/clang/include/clang/Basic/DiagnosticSemaKinds.td

+def warn_hlsl_unexpected_rayquery_flag : Warning<
+  "When using 'RAY_FLAG_FORCE_OMM_2_STATE' in RayFlags, RayQueryFlags must have RAYQUERY_FLAG_ALLOW_OPACITY_MICROMAPS set.">, DefaultError;
+def note_hlsl_unexpected_rayquery_flag : Note<
+  "RayQueryFlags must have RAYQUERY_FLAG_ALLOW_OPACITY_MICROMAPS set here.">;


I think it's sufficient to use the existing note_previous_decl for this instead.

def note_previous_decl : Note<"%0 declared here">;

tex3d · 2025-03-04T02:53:52Z

tools/clang/lib/Sema/SemaHLSLDiagnoseTU.cpp

+  bool IsRayFlagForceOMM2StateSet(const CallExpr *CE) {
+    const ImplicitCastExpr *RayFlagArg =
+        dyn_cast<ImplicitCastExpr>(CE->getArg(1));
+    if (!RayFlagArg)
+      return false;
+
+    const DeclRefExpr *DR = dyn_cast<DeclRefExpr>(RayFlagArg->getSubExpr());
+    if (DR) {
+      const ValueDecl *valueDecl = DR->getDecl();
+      const VarDecl *varDecl = dyn_cast<VarDecl>(valueDecl);
+
+      if (varDecl) {
+        const Expr *initializer = varDecl->getInit();
+        if (!initializer)
+          return false;
+
+        const IntegerLiteral *intLit = dyn_cast<IntegerLiteral>(initializer);
+        if (!intLit)
+          return false;
+
+        unsigned int rayFlagValue = intLit->getValue().getZExtValue();
+        if (rayFlagValue & (unsigned int)DXIL::RayFlag::ForceOMM2State) {
+          return true;
+        }
+      }
+    } else if (const IntegerLiteral *intLit =
+                   dyn_cast<IntegerLiteral>(RayFlagArg->getSubExpr())) {
+      unsigned int rayFlagValue = intLit->getValue().getZExtValue();
+      if (rayFlagValue & (unsigned int)DXIL::RayFlag::ForceOMM2State)
+        return true;
+    }
+    return false;
+  }


I think we should simply be trying to check a constant expression, and evaluating the expression to look for the flag value, not drilling specifically into a DeclRefExpr. Something like this?

Suggested change

bool IsRayFlagForceOMM2StateSet(const CallExpr *CE) {

const ImplicitCastExpr *RayFlagArg =

dyn_cast<ImplicitCastExpr>(CE->getArg(1));

if (!RayFlagArg)

return false;

const DeclRefExpr *DR = dyn_cast<DeclRefExpr>(RayFlagArg->getSubExpr());

if (DR) {

const ValueDecl *valueDecl = DR->getDecl();

const VarDecl *varDecl = dyn_cast<VarDecl>(valueDecl);

if (varDecl) {

const Expr *initializer = varDecl->getInit();

if (!initializer)

return false;

const IntegerLiteral *intLit = dyn_cast<IntegerLiteral>(initializer);

if (!intLit)

return false;

unsigned int rayFlagValue = intLit->getValue().getZExtValue();

if (rayFlagValue & (unsigned int)DXIL::RayFlag::ForceOMM2State) {

return true;

}

}

} else if (const IntegerLiteral *intLit =

dyn_cast<IntegerLiteral>(RayFlagArg->getSubExpr())) {

unsigned int rayFlagValue = intLit->getValue().getZExtValue();

if (rayFlagValue & (unsigned int)DXIL::RayFlag::ForceOMM2State)

return true;

}

return false;

}

bool IsRayFlagForceOMM2StateSet(const CallExpr *CE) {

Expr *E = CE->getArg(1);

llvm::APSInt constantResult;

return E->isIntegerConstantExpr(constantResult, sema->getASTContext()) &&

(constantResult.getLimitedValue() &

(uint64_t)DXIL::RayFlag::ForceOMM2State != 0);

}

Wait, after writing this, I don't think this function is even used. It should be deleted then.

Made use of the function and your suggestion 👍

tex3d · 2025-03-04T03:04:22Z

tools/clang/lib/Sema/SemaHLSLDiagnoseTU.cpp

+    // validate that if the ray flag parameter is
+    // RAY_FLAG_FORCE_OMM_2_STATE, the ray query object
+    // has the RAYQUERY_FLAG_ALLOW_OPACITY_MICROMAPS
+    // flag set, otherwise emit a diagnostic


Not "if the ray flag parameter is RAY_FLAG_FORCE_OMM_2_STATE", but "if the ray flag parameter has the RAY_FLAG_FORCE_OMM_2_STATE flag set".

Suggested change

// validate that if the ray flag parameter is

// RAY_FLAG_FORCE_OMM_2_STATE, the ray query object

// has the RAYQUERY_FLAG_ALLOW_OPACITY_MICROMAPS

// flag set, otherwise emit a diagnostic

// If the RayFlag parameter has RAY_FLAG_FORCE_OMM_2_STATE set,

// the RayQuery decl must have RAYQUERY_FLAG_ALLOW_OPACITY_MICROMAPS set,

// otherwise emit a diagnostic.

tex3d · 2025-03-04T03:09:51Z

tools/clang/lib/Sema/SemaHLSLDiagnoseTU.cpp

+    bool IsRayFlagForceOMM2State = false;
+    if (ImplicitCastExpr *RayFlagArg =
+            dyn_cast<ImplicitCastExpr>(callExpr->getArg(1))) {
+      if (DeclRefExpr *DR = dyn_cast<DeclRefExpr>(RayFlagArg->getSubExpr())) {
+        ValueDecl *valueDecl = DR->getDecl();
+        VarDecl *varDecl = dyn_cast<VarDecl>(valueDecl);
+
+        if (varDecl) {
+          Expr *initializer = varDecl->getInit();
+
+          if (initializer) {
+            if (const IntegerLiteral *intLit =
+                    dyn_cast<IntegerLiteral>(initializer)) {
+              unsigned int rayFlagValue = intLit->getValue().getZExtValue();
+              if (rayFlagValue & (unsigned int)DXIL::RayFlag::ForceOMM2State) {
+                IsRayFlagForceOMM2State = true;
+              }
+            }
+          }
+        }
+      } else if (const IntegerLiteral *intLit =
+                     dyn_cast<IntegerLiteral>(RayFlagArg->getSubExpr())) {
+        unsigned int rayFlagValue = intLit->getValue().getZExtValue();
+        if (rayFlagValue & (unsigned int)DXIL::RayFlag::ForceOMM2State) {
+          IsRayFlagForceOMM2State = true;
+        }
+      }
+    }


We should just use isIntegerConstantExpr to get the value from the ICE if possible.

Suggested change

bool IsRayFlagForceOMM2State = false;

if (ImplicitCastExpr *RayFlagArg =

dyn_cast<ImplicitCastExpr>(callExpr->getArg(1))) {

if (DeclRefExpr *DR = dyn_cast<DeclRefExpr>(RayFlagArg->getSubExpr())) {

ValueDecl *valueDecl = DR->getDecl();

VarDecl *varDecl = dyn_cast<VarDecl>(valueDecl);

if (varDecl) {

Expr *initializer = varDecl->getInit();

if (initializer) {

if (const IntegerLiteral *intLit =

dyn_cast<IntegerLiteral>(initializer)) {

unsigned int rayFlagValue = intLit->getValue().getZExtValue();

if (rayFlagValue & (unsigned int)DXIL::RayFlag::ForceOMM2State) {

IsRayFlagForceOMM2State = true;

}

}

}

}

} else if (const IntegerLiteral *intLit =

dyn_cast<IntegerLiteral>(RayFlagArg->getSubExpr())) {

unsigned int rayFlagValue = intLit->getValue().getZExtValue();

if (rayFlagValue & (unsigned int)DXIL::RayFlag::ForceOMM2State) {

IsRayFlagForceOMM2State = true;

}

}

}

Expr *Expr1 = CE->getArg(1);

llvm::APSInt constantResult;

bool IsRayFlagForceOMM2State =

Expr1->isIntegerConstantExpr(constantResult, sema->getASTContext()) &&

(constantResult.getLimitedValue() &

(uint64_t)DXIL::RayFlag::ForceOMM2State) != 0;

tex3d · 2025-03-04T03:18:58Z

tools/clang/lib/Sema/SemaHLSLDiagnoseTU.cpp

+      auto *attr = methodDecl->getAttr<HLSLIntrinsicAttr>();
+      if (!attr)
+        return true;


Need to check group before assuming you can cast and compare the opcode.

Suggested change

auto *attr = methodDecl->getAttr<HLSLIntrinsicAttr>();

if (!attr)

return true;

auto *attr = methodDecl->getAttr<HLSLIntrinsicAttr>();

if (!attr)

return true;

StringRef OpcodeGroup = GetHLOpcodeGroupName(HLOpcodeGroup::HLIntrinsic);

if (attr->getGroup() != OpcodeGroup)

return true;

tex3d · 2025-03-04T03:24:36Z

tools/clang/lib/Sema/SemaHLSLDiagnoseTU.cpp

+    if (IsRayFlagForceOMM2State) {
+      const DeclRefExpr *DRE = dyn_cast<DeclRefExpr>(ME->getBase());
+      assert(DRE);
+      const VarDecl *VD = cast<VarDecl>(DRE->getDecl());


I don't think this is the right approach. I think you want to call getImplicitObjectArgument() on the CXXMemberCallExpr instead. Then use the expression's type to drill into the template argument.

…d declrefexpr

damyanp · 2025-03-07T22:39:42Z

tools/clang/lib/Sema/SemaHLSLDiagnoseTU.cpp

+          dyn_cast<DeclRefExpr>(callExpr->getImplicitObjectArgument());
+      assert(DRE);
+      QualType QT = DRE->getType();
+      auto typeRecordDecl = QT->getAsCXXRecordDecl();


Suggested change

auto typeRecordDecl = QT->getAsCXXRecordDecl();

auto *typeRecordDecl = QT->getAsCXXRecordDecl();

damyanp

Approved, assuming the test failures are fixed and Tex is happy you've addressed his feedback.

…errors

bob80905 requested a review from a team as a code owner February 24, 2025 21:52

bob80905 requested review from tex3d and removed request for a team February 24, 2025 21:52

bob80905 force-pushed the OMM01 branch from 2112a74 to 087ee90 Compare February 26, 2025 19:57

bob80905 added 4 commits February 26, 2025 13:30

draft

6728f10

add diagnostics

f34857a

finish tests, implement tracerayinline check

be104ce

remove unused vars

2b12ea4

bob80905 force-pushed the OMM01 branch from 087ee90 to 2b12ea4 Compare February 26, 2025 21:40

bob80905 and others added 2 commits February 26, 2025 13:41

Merge branch 'main' into OMM01

918fef1

chore: autopublish 2025-02-26T21:52:47Z

48b8420

damyanp reviewed Feb 27, 2025

View reviewed changes

tools/clang/lib/Sema/SemaHLSL.cpp Outdated Show resolved Hide resolved

tools/clang/lib/Sema/SemaHLSLDiagnoseTU.cpp Outdated Show resolved Hide resolved

refactor to reduce indentation

f24d511

damyanp reviewed Feb 28, 2025

View reviewed changes

bob80905 and others added 7 commits February 28, 2025 10:41

address Damyan, do some refactoring

6c6a306

chore: autopublish 2025-02-28T18:47:22Z

632169a

fix nullptr being passed

eecca92

Merge branch 'main' into OMM01

e620046

remove string comparison

6f11089

Merge branch 'OMM01' of https://github.com/bob80905/DirectXShaderComp…

0c06197

…iler into OMM01

fix refactor, fix typo

3d32cb8

tex3d reviewed Mar 4, 2025

View reviewed changes

bob80905 added 4 commits March 5, 2025 17:18

address all but 2 of tex's comments, need avail attr detection + avoi…

22054e5

…d declrefexpr

use avail attrs

82f377d

use getimplicitobjectargument()

1ed850d

self review

ccd51b9

damyanp reviewed Mar 7, 2025

View reviewed changes

damyanp approved these changes Mar 7, 2025

View reviewed changes

damyanp assigned bob80905 Mar 8, 2025

bob80905 added 3 commits March 7, 2025 16:59

reintroduce nullptr safety check

2fc68db

address Damyan

9b7a357

move test file and undo changes to prevent codegenhashstability test …

71bb274

…errors

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[OMM] Implement front end diagnostics for OMM, including on TraceRayInline #7156

[OMM] Implement front end diagnostics for OMM, including on TraceRayInline #7156

bob80905 commented Feb 24, 2025 •

edited

Loading

github-actions bot commented Feb 26, 2025 •

edited

Loading

damyanp left a comment

damyanp left a comment

damyanp Feb 28, 2025

tex3d left a comment

tex3d Mar 4, 2025

bob80905 Mar 5, 2025

bob80905 Mar 8, 2025

bob80905 Mar 8, 2025

tex3d Mar 4, 2025

tex3d Mar 4, 2025

tex3d Mar 4, 2025

bob80905 Mar 5, 2025

tex3d Mar 4, 2025

tex3d Mar 4, 2025

tex3d Mar 4, 2025

bob80905 Mar 6, 2025

tex3d Mar 4, 2025

tex3d Mar 4, 2025

tex3d Mar 4, 2025

tex3d Mar 4, 2025

damyanp Mar 7, 2025

damyanp left a comment

	bool DiagnoseRayQueryObject(QualType qt, Declarator &D, Sema &S) {
	static bool DiagnoseRayQueryObject(QualType qt, Declarator &D, Sema &S) {

	def warn_hlsl_unexpected_value_for_ray_query_flags_template_arg : Warning<
	def warn_hlsl_rayquery_flags_disallowed: Warning<

	def warn_hlsl_unexpected_rayquery_flag : Warning<
	def warn_hlsl_rayquery_flags_conflict : Warning<

	auto typeRecordDecl = QT->getAsCXXRecordDecl();
	auto *typeRecordDecl = QT->getAsCXXRecordDecl();

[OMM] Implement front end diagnostics for OMM, including on TraceRayInline #7156

Are you sure you want to change the base?

[OMM] Implement front end diagnostics for OMM, including on TraceRayInline #7156

Conversation

bob80905 commented Feb 24, 2025 • edited Loading

github-actions bot commented Feb 26, 2025 • edited Loading

damyanp left a comment

Choose a reason for hiding this comment

damyanp left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

tex3d left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

damyanp left a comment

Choose a reason for hiding this comment

bob80905 commented Feb 24, 2025 •

edited

Loading

github-actions bot commented Feb 26, 2025 •

edited

Loading