Wouldn't it be cleaner to declare lc_id as u_char *lc_id ? Also, why the type is u_char ?

Why +1 ?

Well, the structure is allocated large enough to
store the actual "id", which is an 8bit byte "string"
that the server handles opaquely. (ie. Just
bcmp()'s == or !=). It was made lc_id[1], so I
didn't need to "+ 1" for the null termination byte
when I allocated it. (This code was written long
ago. If the diff was against -head, this wouldn't
have been a change.)

If you think lc_id[0] is clearer, that's easy to change. All that is needed is a "+ 1" on the size
allocated via malloc().

This is normally small, so I don't think a 3rd
malloc() is justified, but I can do it that way.
(If it is defined as *lc_id, I think I need to do a
third malloc?)

As for u_char, it's just 8bit arbitrary bytes, so
that seemed more appropriate to char, but
I don't think anything in the code cares what
the type is, so long as it stores 8bits.

It's "+ i", where "i" is the number of non-null
bytes stored in lc_id. If it is declared lc_id[0],
this needs to change to "+ i + 1".
(And the font seems to make it look like a "1"
for me too. I love web based stuff;-).)