check null pointer before using memcpy()

[ygj6]

Avoid passing null pointer to std::memcpy().
Fixed the following error mentioned by @jamessan in issue #881:

runtime error: null pointer passed as argument 2, which is declared to never be null

[jamessan]

Should this only be inside the if?

[redboltz]

same as #890 (comment)

[redboltz]

@ygj6 , could you apply the same fix as #890 ?

[codecov-commenter]

Codecov Report

Merging #891 into cpp_master will decrease coverage by 0.07%.
The diff coverage is 100.00%.

@@              Coverage Diff               @@
##           cpp_master     #891      +/-   ##
==============================================
- Coverage       86.51%   86.44%   -0.08%     
==============================================
  Files              72       72              
  Lines            5071     5089      +18     
==============================================
+ Hits             4387     4399      +12     
- Misses            684      690       +6     

[redboltz]

It seems that CI errors are reported.

I think that visit_str, visit_bin, and visit_ext should be as follows:

    bool visit_str(const char* v, uint32_t size) {
        assert(v || size == 0);
        if (size > m_limit.str()) throw msgpack::str_size_overflow("str size overflow");
        msgpack::object* obj = m_stack.back();
        obj->type = msgpack::type::STR;
        if (m_func && m_func(obj->type, size, m_user_data)) {
            obj->via.str.ptr = v;
            obj->via.str.size = size;
            set_referenced(true);
        }
        else {
            if (v) {
                char* tmp = static_cast<char*>(zone().allocate_align(size, MSGPACK_ZONE_ALIGNOF(char)));
                std::memcpy(tmp, v, size);
                obj->via.str.ptr = tmp;
                obj->via.str.size = size;
            }
            else {
                obj->via.str.ptr = MSGPACK_NULLPTR;
                obj->via.str.size = 0;
            }
        }
        return true;
    }

[redboltz]

I'm not sure #891 (comment) solves the CI errors.
I tried to reproduce the error on my environment but I couldn't, so far.

[redboltz]

I'm not sure #891 (comment) solves the CI errors.
I tried to reproduce the error on my environment but I couldn't, so far.

I've tested my update and it works well.
So I apply the update and merge the PR.