Re: EVC++ 3.0 to EVC++ 4.0...

r_z_aret_at_pen_fact.com
Date: 02/24/05 

Date: Thu, 24 Feb 2005 14:16:18 -0500

On Thu, 24 Feb 2005 13:26:00 +1300, "Doug Forster"
<doug_ZAPTHIS_AT_ZAPTHIS_TONIQ_DOT_CO_DOT_NZ> wrote:

>Hi James,
>
>>>...except one customer claims that ever since the upgrade the devices can
>>>spontaneously hang, and a hard reset must be done.
>
>Did you really mean hard reset? In my experience that usually happens only
>with faulty hardware. I have had no problems with any transition from

I first noticed a problem with my app when I was testing on a device
that had no separate reset button. One keypad combination was for soft
reset and one was for hard reset. When my app hung, the soft reset
combination had no effect, so I had to do a hard reset each time. I'm
inclined to say the fact that the soft reset was disabled is a
hardware problem, but not quite in the usual sense.

My applications almost always call my wrapper function, rather than
call SetWindowText directly. The symptoms were quite clearly tied to a
call to SetWindowText in that function. Applying the workaround
eliminated the symptoms. The symptoms came back as soon as I disabled
the workaround. And the workaround was to pass an empty string rather
than NULL whenever the caller "asked" to send a NULL. I had worked all
this out before I found the note on Microsoft's site.

>PPC2002 to PPC2003 or EVC3 to EVC4 and I am sure my apps call SetWindowText
>both directly and indirectly via MFC many times over. We have also tested
>one app on a Symbol PPC2003 scanner also with no problems, though I don't
>know if any of our customers use this device. I've no idea what the
>SetWindowText bug really is, but my instinct is still that the bug is most
>likely yours.
>
>BTW our customers routinely complain that new software releases cause all
>their problems. Sometimes they are right but usually not!
>
>Cheers
>
>Doug Forster
>
>"James Hunter Ross" <james.ross@oneilsoft.com> wrote in message
>news:OPxbcTHGFHA.3376@TK2MSFTNGP12.phx.gbl...
>> Interesting, thanks! I certainly call SetWindowText(), who doesn't? ;-)
>>
>> I don't think I ever pass NULL, but I didnt notice any reference to NULL
>> in the MS problem description; it was very generic: "The SetWindowText
>> function can overflow an internal buffer that is used by the Edit control"
>>
>> Wow. Ok. If the problem exists for EVC++ 3.0 and EVC++ 4.0, perhaps we
>> were just lucky with our memory alignments and stuff. I can't believe
>> they don't provide more; we can't eliminate SetWindowText() calls, even
>> though most are for label (STATIC) controls. All calls we do are through
>> MFC CStatic, CEdit, CButton, etc.
>>
>> We have no control over the platform either; we use Symbo devices, so I
>> guess Symbol would need to release a newer version incorporating all the
>> latest MS changes.
>>
>> We are doomed, whether or not this is the problem...
>>
>

-----------------------------------------
To reply to me, remove the underscores (_) from my email address (and please indicate which newsgroup and message).

Robert E. Zaret, eMVP
PenFact, Inc.
500 Harrison Ave., Suite 3R
Boston, MA 02118
www.penfact.com

Quantcast