Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 79.74%. Comparing base (41f26a3) to head (334b41d).

@@            Coverage Diff             @@
##             main     #678      +/-   ##
==========================================
+ Coverage   79.68%   79.74%   +0.06%     
==========================================
  Files           9        9              
  Lines        3942     3955      +13     
==========================================
+ Hits         3141     3154      +13     
  Misses        801      801              

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

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

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

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

Hi @anutosh491, @mcbarton,
Could you please look at this. Why is this failing on Emscripten?
The only Emscripten relevant change is the I->declare(...) in Cpp::CreateInterpreter.

Weird ..... it says the exception handling test fails. I can recreate the exact same example in the lite deployment

Weird ..... it says the exception handling test fails. I can recreate the exact same example in the lite deployment

CppInterOp/unittests/CppInterOp/InterpreterTest.cpp

Lines 155 to 179 in 515c1c8

	TEST(InterpreterTest, EmscriptenExceptionHandling) {
	#ifndef EMSCRIPTEN
	GTEST_SKIP() << "This test is intended to check exception handling for Emscripten builds.";
	#endif
	std::vector<const char*> Args = {
	"-std=c++20",
	"-v",
	"-fexceptions",
	"-fcxx-exceptions",
	"-mllvm", "-enable-emscripten-cxx-exceptions",
	"-mllvm", "-enable-emscripten-sjlj"
	};
	Cpp::CreateInterpreter(Args);
	const char* tryCatchCode = R"(
	try {
	throw 1;
	} catch (...) {
	0;
	}
	)";
	EXPECT_TRUE(Cpp::Process(tryCatchCode) == 0);

@anutosh491 could this have anything to do with the fact that exception handling in CppInterOps deployment doesn't currently work (I'm assuming your example comes from xeus-cpps deployment)? That requires this PR to go in #686 (patch taken from Emscripten forge).

No it shouldn't matter (unless anything in createinterpreter or process changed)

Hi @anutosh491, @mcbarton, Could you please look at this. Why is this failing on Emscripten? The only Emscripten relevant change is the I->declare(...) in Cpp::CreateInterpreter.

I think this can be ifdef-ed if it's too hard to fix for emscripten...

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

@vgvassilev @Vipul-Cariappa if you look at this job https://github.com/compiler-research/CppInterOp/actions/runs/16648661110/job/47122629847?pr=678 it passes the test other Emscripten jobs fail for. The only difference between this job and the other llvm 20 Emscripten jobs is that this patch has been applied on main https://github.com/compiler-research/CppInterOp/blob/main/patches/llvm/emscripten-clang20-3-enable_exception_handling.patch . I hadn't updated all the llvm 20 caches yet, since it quite disruptive and I planned to do it this evening when things are quiet. Once I have updated the llvm 20 Emscripten cache on main, then all the llvm 20 jobs should then pass. To get all the jobs to pass you would need to backport this patch to llvm 19.

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

@Vipul-Cariappa @vgvassilev The llvm 19 Emscripten jobs are now passing the test they were previously failing for, now the exception handling patch has been backported. The ci should hopefully pass now.

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

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

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

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