Re: proc_file_read bug?

[Thomas Hood]

On Fri, 2002-01-25 at 09:29, Andreas Schwab wrote:
> |>                 /* This is a hack to allow mangling of file pos independent
> |>                  * of actual bytes read.  Simply place the data at page,
> |>                  * return the bytes, and set `start' to the desired offset
> |>                  * as an unsigned int. - [e-mail]
> |>                  */

> It is documented, RTFC.

Comment or Code?  The comment is somewhat ambiguous and incorrect.

Reading the code, I take it that "start" is either a pointer into
the buffer where the string of n data bytes starts, or else
(when it is assigned a value less than the beginning of the buffer)
it is a special value by which the file offset is to be adjusted,
instead of n.  Thus the comment might be clarified:
    /* 
     * This is a hack to allow adjusting the file offset
     * by a number different from the number of bytes read.
     * Simply place the data at page, return the number of
     * bytes read, and set "start" to the (signed long) amount
     * by which the file offset is to be increased or
     * decreased
     */

My question then is: why would one want to adjust the file
offset other than by +n?

--
Thomas