Re: 0.14.0 released! [BUGS] -- LINUX5250

On Tue, Nov 02, 1999 at 09:17:15AM -0500, Sean Porterfield wrote:
> > - Field Minus sequence (Esc+M) added.
> 
> > - Signed numeric fields will let you type a digit in the sign position.
> 
> I've only been able to get a negative number in a signed numeric field (data 
>type
> S in DDS) but not in Numeric (data type Y in DDS).  Our order entry program 
>uses
> the Y; CA allows negative numbers there.

Huh?  That doesn't make sense.  I beleive you, but that doesn't make sense ;)

> 
> When a negative number is in the signed numeric field, the negative sign will 
>not
> go away when keying a positive number or just field exit with the field blank.
> Then the program crashes because of the - in the field.

Yes this is part of the group of problems I hope to attack in the 0.15.x
series.  It has to do with the silly way in which tn5250 is currently handling
field data ... basically this is a sync problem between the display buffer
and the field data.  In 0.15.1, there is only one copy of the data (in the
display buffer) so that this _can't_ happen.  To this point, we've been relying
on the programmer remembering at every point that the display is modified to
modify the format table as well, which is both more difficult and more code.

The second part of this bug is that Field Exit for a numeric or signed numeric
field doesn't make the value of the field positive ... I realized this while
hacking out 0.15.1 - it's just plain unimplemented.

> 
> >From debug (STRISDB) pressing F9 retrieves the previous command.  If I make a
> change to that command and press enter, tn5250 seems to only send back what
> I typed rather than the whole command that is displayed on my screen.
> Example: testing an RPG program that has *in90 and *in91.  Type "w *in90" 
>without
> quotes, of course; press enter.  Press F9; change the 90 to 91; press enter
> again.  ISDB reports "Statement 1 not found." indicating it only got the 1
> instead of the whole "w *in91".  (It says "Statement 2 not found." if I 
>change it
> to *in92, etc.)

This is the same basic problem as above - a sync problem between format data
and display data.

Can you send me a tracefile demonstrating as many of the above problems as
possible?

Fixing these problems in 0.14.x would be a whole lot of work, why don't we
postpone the other additional features scheduled in 0.15.x and stabilize
0.15.x instead?  Looking at it now, stabilizing 0.15.x doesn't seem as much
work as I had anticipated.

I'm saying this because I expect a few more of these bugs to pop up once I fix
these, and a few more after that, and I wouldn't expect to get much done on
0.15 otherwise.

-Jay 'Eraserhead' Felice

P.S. Scott, can you send me a tracefile of your crash?  I don't seem to
have your email here anywhere.
+---
| This is the LINUX5250 Mailing List!
| To submit a new message, send your mail to LINUX5250@midrange.com.
| To subscribe to this list send email to LINUX5250-SUB@midrange.com.
| To unsubscribe from this list send email to LINUX5250-UNSUB@midrange.com.
| Questions should be directed to the list owner/operator: david@midrange.com
+---