Re: Run time error in TN5250.EXE version 0.17.4 for Windows

Hi Bill
There is a patch for winterm.c in the repository that fixes some bugs,
maybe that one, but it is applied after 0.17.4.  You can get mingw and
compile it yourself.

It was supposed to fix the crash in XP when the system goes into lock
mode, but it doesn't if there is any screen activity after it goes into
lock mode.

The patch in the repository doesn't match the patch in the archives of
this list.  The patch in the archives removes line 2154:
                TN5250_ASSERT(0);
I recompiled it without this line, and it no longer locks up.  I don't
know if this will fix your problem or not.  I don't even know enough to
know if this might cause problems elsewhere, but it's worked ok on 4
systems for the last two weeks.
John Chambers

-----Original Message-----
From: linux5250-bounces-Zwy7GipZuJhWk0Htik3J/[email protected]
[mailto:linux5250-bounces-Zwy7GipZuJhWk0Htik3J/[email protected]] On Behalf Of Bill Evans
Sent: Thursday, December 17, 2009 3:56 AM
To: linux5250-Zwy7GipZuJhWk0Htik3J/[email protected]
Subject: [LINUX5250] Run time error in TN5250.EXE version 0.17.4 for
Windows

   Sorry if this gets posted twice.  I wasn't subscribed to the mailing
list
   when I posted the first time.

   -----------------------------------------------------------

   I was previously using version 0.17.3 on Windows XP SP3 but there is
a bug
   in that version with copying to the clipboard.  Basically, the text
   highlighted and copied from the 5250 display via Edit / Copy isn't
pasted
   correctly to a target application.

   So, I tried version 0.17.4 and that problem appears to have been
fixed.
   However, there's a bug in 0.17.4 that causes a run time error when
   clicking on the last line of the 5250 window.  The error occurs even
   before signing on to the host.

   I generated a trace file and the run time error occurs here:

   Assertion (This)->cy < (This)->h failed at ../lib5250/dbuffer.c, line
412.
   I looked at the code in dbuffer.c on the website.  Routine
   tn5250_dbuffer_cursor_set in module dbuffer.c triggers the
ASSERT_VALID
   macro.  If line 24 is clicked on, the ASSERT_VALID macro will fail
since
   at that point (This)->cy < (This)->h cannot be true since line 24
cannot
   be less than 24, the screen height.

   Can't this be easily resolved by recompling with NDBUG defined on the
   command line with the -D option to disable the ASSERT macros?
   Unfortunately, I don't have the facilities to perform a
re-compilation.

   I previously privately emailed this info to Scott Klement (I found
him in
   the AUTHORS.TXT file in the distribution and I've read some of his
   columns, etc elsewhere) but I haven't heard from him thus far.  I
   understand ... it's the holidays.

   Anyway, I see that this bug was also reported to the website on
9/16/2009
   by jmarkmurphy but there hasn't been any response.

   Could someone help with this?  I need the clipboard copy/paste fixes
in
   this version but I cannot cope with it terminating when I
accidentally
   click on line 24.

   Thanks in advance.

   Bill Evans
-- 
This is the Linux 5250 Development Project (LINUX5250) mailing list
To post a message email: LINUX5250-Zwy7GipZuJhWk0Htik3J/[email protected]
To subscribe, unsubscribe, or change list options,
visit: http://lists.midrange.com/mailman/listinfo/linux5250
or email: LINUX5250-request-Zwy7GipZuJhWk0Htik3J/[email protected]
Before posting, please take a moment to review the archives
at http://archive.midrange.com/linux5250.

-- 
This is the Linux 5250 Development Project (LINUX5250) mailing list
To post a message email: LINUX5250-Zwy7GipZuJhWk0Htik3J/[email protected]
To subscribe, unsubscribe, or change list options,
visit: http://lists.midrange.com/mailman/listinfo/linux5250
or email: LINUX5250-request-Zwy7GipZuJhWk0Htik3J/[email protected]
Before posting, please take a moment to review the archives
at http://archive.midrange.com/linux5250.