Re: Moving data into dynamic storage

[Rich Duzenbury <rduz@xxxxxxxxxxxxxxxxxxx>]

At 01:52 PM 11/22/02, you wrote:

> Clearly, MOVEL is inappropriate since you don't often really know
> how much storage is actually allocated.
Not so clearly to me.  I ran a test and set the length of 'entry' to be one
byte, thinking that the movel might overrun the buffer, but it does
not.  Only the first byte of storage got loaded, according to debug.


> First, as someone else mentioned, declare your based variable with
> some large length, such as 65535. Since no static storage is
> allocated, the length doesn't really matter. Then, *only* work with
> the data using %SUBST. If the length is maintained in variable
> ENTRY_LEN, then use the based variable as %SUBST(ENTRY:1:ENTRY_LEN).
> This works on both left-hand side and right-hand side of an EVAL
> statement.
I am trying to avoid any length restriction, at least up to the limits of
the ALLOC opcode.


> Of course, you're still limited to a maximum of 65535 characters. If
> you want to work with more data than that, then you need to manage
> the bits and bytes yourself at an even lower level.
The data is of variable length.  I don't know how long it will be.  I do
know that I am restricted to 16M on the alloc, according to the RPGLE
reference manual.  I would like to know whatever the correct solution is to
allow me to reference the entire buffer.

Hmmm.  Can I prototype and call memcpy?  I don't see it in the information
center, but I guess I'll see if I can whip up a test.

Regards,
Rich