Re: tn5250-0.15.2 is realeased. -- LINUX5250

Okay, yesterday I went in and tested all this stuff (finally), and, uh, fixed
it again.  Well, mostly ;)

I blackmailed one of my co-workers into giving me a crash course in SDA/PDM/
DDS, so I hope to soon have a test regression suite (which I hope can be
distributed with the emulator).  During the course of this crash course, I
got very strange looks from said co-worker as I managed to get the emulator
to do some pretty twisted things, such as place letters in random places,
shift fields one cell to the left but expect entry in the original places,
not accept numeric input, flicker weirdly, and core dump when a trace file
was enabled.  I think I had a bad compile or something, as I was not able
to duplicate some of the problems.  Others, however...

There are updates in CVS.

On Mon, Nov 22, 1999 at 09:06:27AM -0500, Sean Porterfield wrote:
> > > I saw some strange colors and data flash up on my screen then disappear.
> >
> > Fixed, I think.
> 
> Different, but still there.  My whole screen turned blue with tiny horizontal
> black lines (like they were blue cursors and the black was space between 
>lines.)
> There were some other colors and symbols, but it flashes off too fast to 
>really
> see.

Hmm, I caught a very strange issue where random keystrokes would be 'handled'
for no apparent reason when the emulator first starts up - I've added a work-
around, but I should really address this and figure out why later.  On my
machine, it puts some strange escape characters into the display and causes
the first few fields of the signon to turn inverse purple.  This may or may
not be the same issue.

> 
> > > When I tried to browse a source member in SEU, it crashed:
> > >     Assertion (This)->cx < (This)->w failed at dbuffer.c, line 114.
> >
> > Hmmm, try it out.  I don't think so, but...
> 
> Still crashes, haven't checked for a specific error.

Hmm, I've done this a few times but don't seem to have any problems.  If you're
still getting this, send me a trace file.

> 
> > > I can't enter anything in my numeric field anymore and was going to look
> > > up what data type it was.  It has an S for keyboard shift of Signed
> > > Numeric.
> >
> > Fixed.
> 
> Not on mine!  :)  The whole session locked up on that one and still doesn't 
>allow
> anything in the Y data type.

Uh, my fixing actually broke more than it fixed ;)  The fix was fixed and the
fix fix is now in CVS ;)  I'm not making a release just yet because I'm having
a problem during testing and I'm not sure if it has to do with the emulator
or my horrific first attempts at AS/400 CL.  The scoop is that the fields may
be transmitting to the 400 improperly, but entering data in Numeric Only and
Signed Numeric fields should be completely to spec now, including fully proper
Field Exit/ Field- / Field+ handling.

> 
> > None of this has been tested real well (if at all), but I must sleep ...
> 
> Hope you got enough sleep!
> 
> One of the programs I use on my system has a numeric field (type Y).  I went 
>in
> and tried to key a number, got the XII (input inhibit) then reset.  When
> I pressed F3 to exit that program, my previous screen wasn't properly aligned.
> My menu went across the screen - everything was in the wrong place, and I 
>didn't
> have the input line even though I could type in a field.

Hmm, I've encountered this a few times, but couldn't duplicate it once I had
the time to debug.  It seems to happen somewhat at random, at least for me.
Tracefile, please!

> 
> It screws up even if I don't key in any field - just go in then F3 back out.  
>My
> screen has all the text in the first few lines, then a purple patch, then a 
>lot
> of red lines.  Very odd!
> 
> If that makes any sense at all...  Would a trace file help, or perhaps are you
> aware of this already?

Yes, I am already aware of the fact that a trace file would help ;)

> 
> I have another screen that is entirely black except for the 5250 in the lower
> left, and my cursor.  It wasn't black a minute ago, but I ran a program then
> exited out.  It didn't restore the screen that was there.

Hmm, I widened the field that keeps track of saved format tables and saved
display formats, thereby allowing me to remove some strange, slow code that
I had kludged in as an intermediate fix.  That may be the problem.  I didn't
seem to have any problems there; however.

> 
> Let me know if trace files will help for any of this.

A trace file would help.  Thanks for the testing, and certainly for the
patience.

> 
> Sean
> 

-Jay 'Eraserhead' Felice

P.S.  There are updates in CVS for the Gnome-based front end 
(cvs checkout gnome-5250), and I've put James' DDS conversion tools in the
repository (cvs checkout dds-tools).  There's still some weird keyboard
handling issues with gnome-5250 ... it's extremely beta.  dds-tools, on the
other hand, is very featureless at this point, but very simple, useable, and
most importantly, useful.  It contains the utilities dds2sql and dds2curses.

Busy weekend ;)
+---
| 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
+---