Re: incorrect cursor position -- LINUX5250


>This is Kenneth from the tn5250j project.  I would be happy to talk shop
>about the cursor positioning.  Well that is in about a week's time that is

>as tomorrow I will be leaving on vacation for a week and will not be back
>until the beginning of June. ...

Sounds good.

>It is possible though to get it almost 100% even on restoring the windows
>from savescreen and restorescreen functions.  The first thing that will
have
>to be done is get the enhanced functions turned on and working, then there

>is the exact flags in the query response back to the controller to return
>the IC and MC commands in the correct order.

If by enhanced you mean the "Write To Display: Structured Field Order" I'm
not sure I agree with you %100 there.
Correct me if I'm wrong, but don't we know if we pretend to be an older
terminal in the Write Structured Field response, that the AS400 will never
send WTD SFOs.
These features can even be turned off in Client Access by editing your .ws
session file. (I think it's something like ENPTUI=N (Enhanced
Non-Programmable Terminal User Interface) in the [5250] section, but don't
quote me on that.)
Needless to say, _regardless_ of what type of terminal our emulators are
emulating, the cursor should still end up in the right position.

Back to what I said earlier about SaveScreen and RestoreScreen ignoring
cursor position. It looks right now like I'm 100% wrong.
The documentation actually says the exact opposite. I could have sworn that
I had read that a RestoreScreen is always followed by a WTD to actually set
the cursor position. But I can't seem to find where I read it.

>Let me tell you it was almost
>a solid three weeks flipping bits and options and then interpreting them.

I believe it.

>The tn5250j is still not correct until I sit down and start putting it
into
>there.
>I hope this helps and again would be glad to talk to someone on this to
get
>another persons view.  What do you say?  You want to put this crusade
>together and all three projects benefit?  Yes I know the code is a mess
but
>it has been progressing since the last time you looked at it.

Yes. My blood is boiling once again over this little cursor.
I'll take a look at your code later (sf is down right now) and start
getting back into the "IC/MC/WTD" state of mind and warm up ethereal.

I think what we could do to start is jointly create some sort of state
diagram about behavior we know to be true, and behavior that is ambiguous
in the documentation. Then we can we can begin to clarify exactly how many
combinations of behavior we have to implement/test in order to cover all
our bases with the ambiguous cases.
This should also include any bit flags (WTD CCs, etc) we are suspicious of.
We can then start to record what succeeds and fails with each test.

-Mark
email:
markb
cms400
com