[sheepdog] [PATCH V4 2/2] collie: optimize 'collie vdi check' command

[Liu Yuan]

On 09/03/2012 07:12 PM, Yunkai Zhang wrote:
> On Mon, Sep 3, 2012 at 7:03 PM, Liu Yuan <namei.unix at gmail.com> wrote:
>> On 09/03/2012 06:54 PM, Yunkai Zhang wrote:
>>> We can compare this result with comand line tools: sha1sum, which can
>>> also help us to verify program's correction I think.
>>
>> Could you specify a use case for this? I have no idea how end users can
>> make use of sha1 string. For a crashed cluster, I think want users need
>> to do is, run check to restore cluster consistency by one command like fsck.
> 
> Our root problem is just why we need to move sha1_to_hex from
> farm/sha1_file.c to lib/sha1.c, does this change bother you?
> 

Of course, it is ugly. Sha1 is only used in Farm, moving it out of Farm implementation
is nonsense to me.

> Show sha1 string will make the output more meaningful for us, why need
> to a use case? It's a so small/simple thing, do you like argue
> everything?
> 

1) it is my duty to review patches and whoever tries to move sha1_to_hex() out of
sha1_file.c, I will ask why he needs it and verify if it is a reasonable movement.
2) it is you who always take patch comments as malicious argument which is meant to
be against you but in fact only meaningful to patch itself.
3) I don't like arguing at all. If I am not interested in it, I most of time simply ignore it.

I think I'd better run away from this patch set. It waste of my time to reply emails like this.

Thanks,
Yuan