Jeff King <[email protected]> writes:
> @@ -116,19 +116,11 @@ unsigned int memihash_cont(unsigned int hash_seed,
> const void *buf, size_t len);
> * Converts a cryptographic hash (e.g. SHA-1) into an int-sized hash code
> * for use in hash tables. Cryptographic hashes are supposed to have
> * uniform distribution, so in contrast to `memhash()`, this just copies
> - * the first `sizeof(int)` bytes without shuffling any bits. Note that
> - * the results will be different on big-endian and little-endian
> - * platforms, so they should not be stored or transferred over the net.
;-)
> + * the first `sizeof(int)` bytes without shuffling any bits.
> */
> static inline unsigned int oidhash(const struct object_id *oid)
> {
> - /*
> - * Equivalent to 'return *(unsigned int *)oid->hash;', but safe on
> - * platforms that don't support unaligned reads.
> - */
> - unsigned int hash;
> - memcpy(&hash, oid->hash, sizeof(hash));
> - return hash;
> + return get_be32(oid->hash);
> }
>
> /*
> diff --git a/t/t0016-oidmap.sh b/t/t0016-oidmap.sh
> index bbe719e950..6656db9d69 100755
> --- a/t/t0016-oidmap.sh
> +++ b/t/t0016-oidmap.sh
> @@ -93,9 +93,9 @@ put three 3
> iterate" "NULL
> NULL
> NULL
> +$(git rev-parse three) 3
> $(git rev-parse two) 2
> -$(git rev-parse one) 1
> -$(git rev-parse three) 3"
> +$(git rev-parse one) 1"
>
> '
>
> which not only fixes this test but any other hash-based oddities. I
> wonder if it's appreciably less efficient. I'll bet I could nerd-snipe
> René into doing a bunch of measurements and explorations of the
> disassembled code. ;)
I'd rather see us go in the direction of discouraging people from
relying on the hash order.
