Re: dd tool: program vs. manual contradiction

[Jim Meyering]

Thank you for the report.
GNU dd has worked that way since at least 1992,
but I see nothing in the POSIX spec

  http://www.opengroup.org/onlinepubs/007904975/utilities/dd.html

that suggests such a restriction is necessary,
so I've just changed the code to allow one to combine the unblock
and sync conversions.

Now, dd works like this:

  $ printf 010203x | ./dd cbs=2 ibs=2 conv=unblock,sync 2>/dev/null |cat -A
  01$
  02$
  03$
  x$

At first (reading only the description of sync), I thought the last line
should be space-padded.  But the description of unblock says this:

    `unblock'
          Replace trailing spaces in each `cbs'-sized input block with a
          newline.


Volker Paul <address@hidden> wrote:
> I found what I see as a contradiction between
> the dd program error output and its manual.
>
> info dd extract:
>     `sync'
>     Pad every input block to size of `ibs' with trailing zero
>     bytes.  When used with `block' or `unblock', pad with spaces
>     instead of zero bytes.
>
> I read this as: sync can be used with unblock.
> But I get the error message:
>
> echo 000100020003 | dd cbs=4 conv=sync,unblock
> dd: only one conv in {ascii,ebcdic,ibm}, {lcase,ucase}, {block,unblock},
> {unblock,sync}
>
> which I read as: unblock and sync can't be used together.
>
> What is true?
>
> (What I wanted to do is:
> Convert a file with fixed block size, here 4 bytes,
> to one with every block ended with \n:
> 0001
> 0002
> 0003
> )