Wenyong Huang

Could you check whether the pointers of param are the same to ensure that the pointer is correctly passed into NativeApi_Print? e.g. in ThreadLoadWasm, before calling wasm function, and in...

@arvid-norlander For interpreter, could you try changing `WASM_CPU_SUPPORTS_UNALIGNED_ADDR_ACCESS` to 0 in core/config.h? Or add `add_definitions(-DWASM_CPU_SUPPORTS_UNALIGNED_ADDR_ACCESS=0)` in CMakeLists.txt. For fast-jit, it really needs effort to modify the code, since we assume...

Hi, since x86-64/x86-32 supports memory access on unaligned address, currently we just do in that way in many places of the runtime implementation, it should be a normal behavior, there...

The C/C++ compiler also finally compiles the source code into assembly, AFAIK, most of the x86 instructions support unaligned memory access except few vectorize related instructions, so there seems only...

Yes, maybe we can define some common functions in `core/shared/utils/bh_common.h` by using memcpy: ```C static inline uint16 bh_get_uint16(const uint8 *buf) { uint16 ret; bh_memcpy_s(&ret, sizeof(uint16), buf, sizeof(uint16)); return ret; }...

> > ```c > > bh_set_int16 > > ``` > > why is only bh_set_int16 signed? otherwise sounds reasonable to me. > Sorry for typo, it should be bh_set_uint16. >...

Normally the code in interpreter or AOT code is performance critical, I think the several functions you modified in PR #3436 should have little impact on performance, and as @yamt...

Please try #3447.

@bashor Welcome. Could you upload the wast file also? It is a little complex to analyze the wasm file without wast info, I used the tool compiled from wasm GC...

@bashor I tried to debug it but since I am not familiar with the exception handling, I am not sure whether it is caused by the case itself, or by...