Re: Permanent I/O Error occured to DSPF

["Charles Wilt" <charles.wilt@xxxxxxxxx>]

On Thu, Oct 9, 2008 at 5:18 PM, Simon Coulter <shc@xxxxxxxxxxxxxxxxx> wrote:
> On 10/10/2008, at 2:47 AM, Charles Wilt wrote:
>
> > Now interestingly enough, I get the error when using IBM's Client
> > Access; however if I use Powerterm, I don't get the error.  The device
> > description is the same in either case.
>
> The Query5250 Response allows the controller to specify the maximum
> number of input fields allowed. Normally this is set to X'100' (i.e.,
> 256). I guess it's possible for an emulator to increase this value
> but I haven't tried that. Perhaps Powerterm is doing this which might
> explain your observed behaviour? 256 was certainly the last published
> maximum but perhaps it's been increased (540 shows 256 and 610 is off
> the air at present).

I don't have 256 fields on the screen....Although, there's lots of
hidden and program fields.  Are those counted toward the limit?

Are we talking 256 fields total, or is this a per record format limit?

> You said previously that this error occurs on the READ. I would re-
> examine that because it doesn't make sense to me. READ causes the
> emulator to format a 5250 response and send it back to the host.
> Negative response codes generally indicate problems on a WRITE of
> data. I accept that DFTWRT(*YES) would make it seem to occur on the
> READ but since you say you disabled that and it still happens on the
> READ I think something else is occurring.

Ok, I set DFRWTR(*NO) again, having lost it when I recompiled, and
stepped through the writes.  I'm getting the error on a WRITE of the
format that contains the function key text.  The only fields are 4
program-to-system fields for the DSPATR of the text. The value of all
4 fields is x'3A'.

> You could try tracing the 5250 data stream. Gene Gaunt published a
> TRC5250 command some time ago (I think for iSeries News) that
> provides a nicely formatted dump. It's not complete and has a few
> defects but for the most part it works OK. You could use this to trap
> the 5250 data stream and see what it shows.

I found the article on mcpressonline.com.  Looks like it was published
in the old mag, and I haven't been able to find the code to download.
Maybe somebody here will have a copy.

Otherwise I found a reference to PC5250 offering a way to trace the
data stream, maybe I'll try that.

Thanks for your help!
Charles