On Mon, 2014-07-28 at 19:03 +0700, Nguyễn Thái Ngọc Duy wrote:
> +# Define HAVE_SHM if you platform support shm_* functions in librt.
s/you/your/
> +static void free_index_shm(struct index_shm *is)
Does not actually free its argument; should be release_index_shm.
--
To unsubscribe from this
On Wed, Jul 30, 2014 at 3:08 PM, Eric Sunshine wrote:
>> +static void share_index(struct index_state *istate, struct index_shm *is)
>> +{
>> + void *new_mmap;
>> + if (istate->mmap_size <= 20 ||
>> + hashcmp(istate->sha1,
>> + (unsigned char *)istate->mmap +
On Monday, July 28, 2014, Nguyễn Thái Ngọc Duy wrote:
> Instead of reading the index from disk and worrying about disk
> corruption, the index is cached in memory (memory bit-flips happen
> too, but hopefully less often). The result is faster read.
>
> The biggest gain is not having to verify the
Instead of reading the index from disk and worrying about disk
corruption, the index is cached in memory (memory bit-flips happen
too, but hopefully less often). The result is faster read.
The biggest gain is not having to verify the trailing SHA-1, which
takes lots of time especially on large ind
4 matches
Mail list logo
