Pip Cet wrote:

> I notice that there are a few places where you use XFIXNUMPTR
> directly, rather than make_mint_ptr. Is that intentional? I think now
> would be a good time to fix it and make XFIXNUMPTR internal to lisp.h.

It's intentional since those few places don't need the full power of 
make_mint_ptr (often the pointers are already aligned) and there might be 
trouble if storage allocation fails if the pointer is not aligned. Perhaps 
you're right and this is overkill; it is confusing at any rate. In the meantime 
I installed the attached first patch to fix a glitch I saw in this area while 
looking at the current XFIXNUMPTR uses.

> Also, the doc comments for most_negative_fixnum and
> most_positive_fixnum in data.c still need updating.

Thanks, done in the second attached patch.