Re: Proposal: block-based vector allocator

[Stefan Monnier]

> I don't understand this.  Any block-allocated vector P0 has the size
> less than or equal to VECTOR_BLOCK_BYTES.  So, mem_find (P0) will
> always return mem_node M0 where M0->end - M0->start ==
> VECTOR_BLOCK_BYTES (IOW, such a mem_node always corresponds to
> a vector block).  If P1's size is greater than VECTOR_BLOCK_BYTES,
> including the case when it's equal to sizeof (vector_block), mem_find
> (P1) will always return it's own corresponding node M1 where P1 is
> equal to M1->start and M1->end - M1->start is never equal to
> VECTOR_BLOCK_BYTES.

BTW, if we change the vector allocation so it only uses vector-blocks
for vectors smaller than "vector-block-size / 2", as I think we should,
then I think your reasoning breaks down.


        Stefan