dd displaying stats in a human readable form

[Olivier Delhomme]

Hi,

As Paul Eggert said, we could display really readable stats for dd while
still using the mechanism of SIGUSR1 :

>in the normal case where bytes-read equals bytes-written (in this
>case, 2.7 GB) and there are no partial blocks.  The less common cases
>could use more-unusual formats, e.g.:
>
>    2.7G copied (1 partial record)
>    2.7G copied (3 partial input records, 1 partial output record)
>    2.7G read (5 partial records), 2.8G written
>    2.7G read (5 partial records), 2.8G written (3 partial records)
>
>This would be more readable and understandable, which I think is
>largely the point here.

Jim meyering proposed a conv=human option and a
display-block-size='block-size-todeal-with' option. 

I think we could merge the two options (and forgot the confusing
display-block-size='block-size-todeal-with') and add this feature to a
new human option (that will act as conv) : --human='block-size-todeal-with'

What do you think about this idea ? 

I can understand that the 'dd' has to be light weighted but in some cases,
having a simple way to know how much of data has been processed is
important.

Regards,

Olivier Delhomme.

--