Hi, Looking through the slice allocator, I see that it requests blocks from posix_memalign that are aligned to page boundaries. Can I then assume the following? For a type T, if T is required to be aligned to A=2^i byte boundaries, and is of size S=j*A, then g_slice_new(T) will always returned correctly aligned memory (i.e, g_slice_new(T)^(A-1) == 0) A and S are less than a page size, of course. I'm also aware that large memory requests are forwarded to malloc rather than posix_memalign, so alignment is not guaranteed there. But I only need to allocate individual blocks. Specifically, I need 16 byte alignment. I think this assumption is valid in a basic fixed-size pool allocator, so long as the pool itself is aligned, but I'm not sure about all this magazine stuff in the slice allocator. Thanks, Scott _______________________________________________ gtk-list mailing list gtk-list@xxxxxxxxx http://mail.gnome.org/mailman/listinfo/gtk-list
