On 20/09/2006, at 11:58, Davi Arnaut wrote:
On 20/09/2006, at 11:00, Brian Akins wrote:
Davi Arnaut wrote:
On 20/09/2006, at 10:16, Brian Akins wrote:
> Davi Arnaut wrote:
>> Simplify the array and table serialization code, separating
it from
>> the underlying I/O operations.
>
> Probably faster to just put every thing in an iovec (think
writev).
Probably no, apr_brigade_writev does (quite) the same.
Doesn't mean apr_brigade_writev does it "fast" either...
If the serialization simply returned an iovec, mod_mem_cache could
use apr_pstrcatv and mod_disk_cache could use apr_file_writev.
There are no brigades _yet_, also pay attention to the comment:
/* XXX this is a temporary function, it will be removed later */
And yes, once we have brigades between then it is much better/faster
to use the writev
functions.
--
Davi Arnaut
