On 08/07/2013 05:48 PM, Junio C Hamano wrote:
> Matthieu Moy <Matthieu.Moy@grenoble-inp.fr> writes:
>>
>> seems overkill to me: why don't you just let cmd_repack call
>> update_server_info(0)?
> 
> My feeling exactly.  I would rather see a patch that does not touch
> pack-objects at all, and use run_command() interface to spawn it.
> Once we do have to pack, the necessary processing cycle will dwarf
> the fork/exec latency anyway, no?
> 

Thanks for clarification. That was my initial idea as well, 
to not touch the pack-objects logic. 

However Duy send that patch (basically as is, 
I just made it apply again), 
and I guessed that I'd get to results
faster with an already existing approach.

I will reconsider and try to remove the additional logic from 
pack-objects again (so it will not get touched) and move it to the
repack itself. It is just a way to understand, 'what needs to be done'.

Stefan