Unexplained code in unicode_to_ansi_alloc()

[pkopylov]

Can anyone explain why a Unicode is expected to be put in the buffer there?

Nonetheless, 4*len+1 bytes are allocated in the code mentioned above, why are just 4*len bytes passed as the target buffer length to the function filling the aptr buffer there ?

When len is zero, it leads to allocating exactly one byte of the memory, but the buffer pointed by aptr will not be filled at all in the called function unicode_to_ansi_copy() if HAVE_ICONV is undefined or connection is not given, right? See: this loop has no iteration and this condition is false.

[v-chojas]

Why do you ask? Are you trying to track down a bug?

[pkopylov]

Why do you ask? Are you trying to track down a bug?

I am studying CVE-2018-7409. It seems that mentioned functions are a scope of this CVE.