Add IsAllocator and IsDeallocator API functions by keremsahn · Pull Request #1020 · compiler-research/CppInterOp

keremsahn · 2026-06-02T02:38:14Z

Adds enum classes for allocation and deallocation types, creates API functions for queries.

keremsahn · 2026-06-02T02:40:20Z

Adds enum classes for allocation and deallocation types, creates API functions for queries.

@Vipul-Cariappa this PR is a prototype, I just searched for most obvious attributes there are more attributes should be added, please provide me feedback. Also it seems not to be possible to detect allocation-deallocation type via attributes. So it may not to be a good idea to build query functions based on attribute searching

codecov · 2026-06-02T02:42:23Z

Codecov Report

❌ Patch coverage is 93.54839% with 2 lines in your changes missing coverage. Please review.
✅ Project coverage is 86.71%. Comparing base (91a0c1e) to head (a93b347).

Files with missing lines	Patch %	Lines
lib/CppInterOp/CppInterOp.cpp	93.54%	2 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1020      +/-   ##
==========================================
+ Coverage   86.62%   86.71%   +0.08%     
==========================================
  Files          23       23              
  Lines        5946     5977      +31     
==========================================
+ Hits         5151     5183      +32     
+ Misses        795      794       -1

Files with missing lines	Coverage Δ
lib/CppInterOp/CppInterOp.cpp	`89.16% <93.54%> (+0.04%)`	⬆️

... and 1 file with indirect coverage changes

Files with missing lines	Coverage Δ
lib/CppInterOp/CppInterOp.cpp	`89.16% <93.54%> (+0.04%)`	⬆️

... and 1 file with indirect coverage changes

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

Vipul-Cariappa

This looks good. Nice job. Some minor comments.

Vipul-Cariappa · 2026-06-02T09:14:37Z

+  assert(Fn && "Not func");
+  auto* D = static_cast<const Decl*>(Fn);
+  assert(D && "Not func");
+  auto* FD = dyn_cast<FunctionDecl>(D);
+  assert(FD && "Not func");


We don't want to crash the application if Fn = nullptr, instead if return a sensible default. In this case, false. You should remove the the asserts.
Asserts are used to detect bugs within the CppInterOp codebase itself. In this case, it is the user providing a null value.

Vipul-Cariappa · 2026-06-02T09:38:31Z

+  assert(Fn && "Not func");
+  auto* D = static_cast<const Decl*>(Fn);
+  assert(D && "Not func");
+  auto* FD = dyn_cast<FunctionDecl>(D);
+  assert(FD && "Not func");


Please remove the asserts. As explained above.

Vipul-Cariappa · 2026-06-02T09:44:13Z

+enum class AllocType : unsigned char {
+  New,
+  NewArr,
+  Malloc,
+  None,
+  Unknown // Could not detect type, for now
+};
+
+enum class DeallocType : unsigned char {
+  Delete,
+  DeleteArr,
+  Free,
+  None,
+  Unknown // Could not detect type, for now
+};


These are not yet used. So let's remove them from this PR. But this looks good. We should model based on this. Are there attributes that differentiate between malloc and new?
Also looking at https://clang.llvm.org/docs/AttributeReference.html#ownership-holds-ownership-returns-ownership-takes-clang-static-analyzer. There is ownership_holds. I believe the DeallocType should also express that.

Yeah this is a problem we can not differentiate allocation-deallocation types from attributes. I will change the PR by your suggestions next week.

These are not yet used. So let's remove them from this PR. But this looks good. We should model based on this. Are there attributes that differentiate between malloc and new? Also looking at https://clang.llvm.org/docs/AttributeReference.html#ownership-holds-ownership-returns-ownership-takes-clang-static-analyzer. There is ownership_holds. I believe the DeallocType should also express that.

I did not handle OwnershipHolds in current commit but this can be another ownership category, probably cppyy will handle this same as a OwnershipTakes but it may be better to distinguish them. What do you think

Vipul-Cariappa · 2026-06-02T09:47:39Z

Please consider rebasing the PR before continuing the work.

vgvassilev · 2026-06-02T09:55:55Z

+  return INTEROP_RETURN(false);
+}
+
+bool IsDeallocator(TCppConstFunction_t Fn) {


Don't forget the builtins: https://github.com/vgvassilev/clad/blob/b10626e37d7d831ac94ca8a559ce9d2c93fa6ed4/lib/Differentiator/CladUtils.cpp#L865

I added for malloc and free, we may extend it for realloc etc. later

github-actions

clang-tidy made some suggestions

github-actions · 2026-06-11T20:13:35Z

+  if (Fn) {
+    if (auto* D = unwrap<clang::Decl>(Fn)) {
+      if (auto* FD = dyn_cast<FunctionDecl>(D)) {
+        if (FD->getBuiltinID() == Builtin::ID::BImalloc)


warning: no header providing "clang::Builtin::ID" is directly included [misc-include-cleaner]

lib/CppInterOp/CppInterOp.cpp:15:

+ #include <clang/Basic/Builtins.h>

github-actions · 2026-06-16T16:37:22Z

clang-tidy review says "All clean, LGTM! 👍"

keremsahn · 2026-06-16T22:02:01Z

@vgvassilev @aaronj0 @Vipul-Cariappa Hello all, can you review the PR, it is getting larger and harder to debug for me

github-actions

clang-tidy made some suggestions

github-actions

clang-tidy made some suggestions

github-actions · 2026-06-16T23:11:34Z

@@ -0,0 +1,7 @@
+#ifndef UNITTESTS_CPPINTEROP_APINOTES_TESTHEADER_H
+#define UNITTESTS_CPPINTEROP_APINOTES_TESTHEADER_H


warning: header guard does not follow preferred style [llvm-header-guard]

Suggested change

#define UNITTESTS_CPPINTEROP_APINOTES_TESTHEADER_H

#ifndef GITHUB_WORKSPACE_UNITTESTS_CPPINTEROP_APINOTES_TESTHEADER_H

#define GITHUB_WORKSPACE_UNITTESTS_CPPINTEROP_APINOTES_TESTHEADER_H

Vipul-Cariappa

Minor comments:

Vipul-Cariappa · 2026-06-24T12:52:51Z

+        if (FDA->getSemanticSpelling() != RestrictAttr::Declspec_restrict)
+          return INTEROP_RETURN(true);


Suggested change

if (FDA->getSemanticSpelling() != RestrictAttr::Declspec_restrict)

return INTEROP_RETURN(true);

return INTEROP_RETURN(true);

Shouldn't this be the following?
I am looking at the definition of RestrictAttr::getSemanticSpelling.

https://learn.microsoft.com/en-us/cpp/cpp/restrict?view=msvc-170
The definition of declspec(restrict) is kinda different from gnu's restrict. It implies pointer returned by the function has no alias, so it is not a neccesity for function to allocate memory.

Vipul-Cariappa

LGTM! Just a small comment. You also need to rebase, so that I can merge.

Vipul-Cariappa · 2026-07-03T04:48:25Z

+    // FIXME: These attributes are not currently compatible with our memory
+    // representation, they are added for APINotes' test


Suggested change

// FIXME: These attributes are not currently compatible with our memory

// representation, they are added for APINotes' test

Note required here. This FIXME should be part of the other PR.

keremsahn force-pushed the is-allocator branch 2 times, most recently from 3b0f128 to 4b90462 Compare June 2, 2026 02:44

Vipul-Cariappa self-requested a review June 2, 2026 09:10

Vipul-Cariappa requested changes Jun 2, 2026

View reviewed changes

vgvassilev reviewed Jun 2, 2026

View reviewed changes

keremsahn force-pushed the is-allocator branch 2 times, most recently from 34e247f to e11ab32 Compare June 11, 2026 20:00

github-actions Bot reviewed Jun 11, 2026

View reviewed changes

keremsahn force-pushed the is-allocator branch from e11ab32 to 11a9c7e Compare June 11, 2026 23:03

keremsahn requested review from Vipul-Cariappa and vgvassilev June 11, 2026 23:07

keremsahn marked this pull request as draft June 16, 2026 16:00

keremsahn force-pushed the is-allocator branch 2 times, most recently from 6968612 to b4e1fd8 Compare June 16, 2026 16:23

keremsahn force-pushed the is-allocator branch from b4e1fd8 to 6047259 Compare June 16, 2026 21:59

keremsahn marked this pull request as ready for review June 16, 2026 22:02

github-actions Bot reviewed Jun 16, 2026

View reviewed changes

Comment thread unittests/CppInterOp/APINotes/TestHeader.h

keremsahn marked this pull request as draft June 16, 2026 22:23

keremsahn force-pushed the is-allocator branch from 6047259 to e0201e5 Compare June 16, 2026 23:00

github-actions Bot reviewed Jun 16, 2026

View reviewed changes

keremsahn force-pushed the is-allocator branch from e0201e5 to 5622a42 Compare June 16, 2026 23:36

keremsahn marked this pull request as ready for review June 17, 2026 09:48

keremsahn force-pushed the is-allocator branch from 5622a42 to 63a6484 Compare June 17, 2026 12:27

Vipul-Cariappa reviewed Jun 25, 2026

View reviewed changes

keremsahn force-pushed the is-allocator branch from fe9d821 to b01bb46 Compare June 25, 2026 18:50

keremsahn requested a review from Vipul-Cariappa June 26, 2026 08:27

Vipul-Cariappa approved these changes Jul 3, 2026

View reviewed changes

Little Fix

6937192

keremsahn force-pushed the is-allocator branch from b01bb46 to 9b441a7 Compare July 3, 2026 12:04

Fix: Early return

a60399f

keremsahn force-pushed the is-allocator branch from 9b441a7 to a60399f Compare July 3, 2026 16:40

Merge branch 'main' into is-allocator

a93b347

		@@ -0,0 +1,7 @@
		#ifndef UNITTESTS_CPPINTEROP_APINOTES_TESTHEADER_H
		#define UNITTESTS_CPPINTEROP_APINOTES_TESTHEADER_H

	#define UNITTESTS_CPPINTEROP_APINOTES_TESTHEADER_H
	#ifndef GITHUB_WORKSPACE_UNITTESTS_CPPINTEROP_APINOTES_TESTHEADER_H
	#define GITHUB_WORKSPACE_UNITTESTS_CPPINTEROP_APINOTES_TESTHEADER_H

		if (FDA->getSemanticSpelling() != RestrictAttr::Declspec_restrict)
		return INTEROP_RETURN(true);

		// FIXME: These attributes are not currently compatible with our memory
		// representation, they are added for APINotes' test

Uh oh!

Conversation

keremsahn commented Jun 2, 2026

Uh oh!

keremsahn commented Jun 2, 2026

Uh oh!

codecov Bot commented Jun 2, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

Vipul-Cariappa left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

keremsahn Jun 11, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Vipul-Cariappa commented Jun 2, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

github-actions Bot left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

github-actions Bot Jun 11, 2026

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

github-actions Bot commented Jun 16, 2026

Uh oh!

keremsahn commented Jun 16, 2026

Uh oh!

github-actions Bot left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

github-actions Bot left a comment

Choose a reason for hiding this comment

Uh oh!

github-actions Bot Jun 16, 2026

Choose a reason for hiding this comment

Uh oh!

Vipul-Cariappa left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Vipul-Cariappa left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

codecov Bot commented Jun 2, 2026 •

edited

Loading

keremsahn Jun 11, 2026 •

edited

Loading