Powerbuilder 3.0a GPF's

Powerbuilder 3.0a GPF's

Post by mdchac » Mon, 12 Sep 1994 01:19:47




>>We're just getting started with PB3.0a and are running into GPFs too
>>numerous to describe.  We've done alot of client/server development in
>>Visual Basic without being interrupted by continual GPFs.  I disagree
>>with the notion that these are due to Windows' instability.  If that's
>>the case, why don't we see so many GPFs with VB or Access?  In fact, we
>>just finished a sizeable development effort with VB and SQL Server and
>>you could count the GPFs during a two month effort in one hand!

While it may be urban legend, I believe that Microsoft has a vested
interest in making sure that their products run better than their
competitors.  While they may not actively try to put bugs in their
competitors products it doesn't take much imagination to think that
VB and other Microsoft developers have access to information (e.g.
"undocumented features" of windows) that non-Microsoft developers don't.

Mike
--

  ^0 0^  __  
  ( v )      Woof.    
    U    (or is it Moo?)          BRAIN MODERNITY

 
 
 

Powerbuilder 3.0a GPF's

Post by Steve Ya » Sun, 11 Sep 1994 07:12:31



>We're just getting started with PB3.0a and are running into GPFs too
>numerous to describe.  We've done alot of client/server development in
>Visual Basic without being interrupted by continual GPFs.  I disagree
>with the notion that these are due to Windows' instability.  If that's
>the case, why don't we see so many GPFs with VB or Access?  In fact, we
>just finished a sizeable development effort with VB and SQL Server and
>you could count the GPFs during a two month effort in one hand!

>PB is an unstable product.  We're sticking with it because we believe it
>has potential.  We especially like the repository and integration with
>ERwin.  If VB had a good repository and links to case tools, and a viable
>data control, it would be a far superior tool.

>Jim Laurel
>Aspect Software Engineering

It has been my experience that many of the GPF's are caused by "bad coding".
As you gain more experience with the product, the number of GPF's that you
encounter will decrease accordingly.  That is not to say that the errors that
caused the GPF's originally couldn't have been handled more gracefully by
PB.  With so much flack about the instability of 3.0, I hope the Powersoft
people learned their lesson and put out a decent 4.0 product.

 
 
 

Powerbuilder 3.0a GPF's

Post by Michael Parks Swa » Mon, 12 Sep 1994 05:35:31



>It has been my experience that many of the GPF's are caused by "bad coding".

  I once wrote an application with three lines of code (one "open" and two
"opensheets"), and was able to get it to GPF reliably. Wasting several weeks
to find out what sets of legal conditions PB freaks out on is not my idea
of productivity. ( PB seems more stable on my machine, but I've been doing
most of my coding lately in C. Other machines still exhibit odd behavior
regularly.)
--
Mike "Agent of Chaos" Swaim

  If you know why a gay person is like a Visa Gold Card,
PLEASE TELL ME.
 
 
 

Powerbuilder 3.0a GPF's

Post by William Harr » Sun, 11 Sep 1994 13:05:34


: It has been my experience that many of the GPF's are caused by "bad coding".

<snort> It is not bad coding that causes the development environment to GPF
when opening a datawindow.  In working with PB, I count it a triumph to get
through any four-hour session without a GPF.  There were days when I had a
dozen GPF's, none traceable to a programming error.  

The development environment itself should never GPF.  The applications
might do all kinds of strange things, but the environment itself should be
free of it.  PB is not.

: That is not to say that the errors that
: caused the GPF's originally couldn't have been handled more gracefully by
: PB.  With so much flack about the instability of 3.0, I hope the Powersoft
: people learned their lesson and put out a decent 4.0 product.

Like, say, not GPF'ing when you try to SetItem into a Querymode datawindow?  
(I'm at home and don't have the app in front of me, but I believe I'm
describing the case correctly.) GPF'ing on that is ...  well, bad.  And
rather darn hard to figure out when they don't say you can't do that.

I'd be harder on Powersoft, only we never could get our 30-day evaluation
copy of SQLWindows to run against Informix, and Knowledgeware's product
(while it worked fine, as far as I could tell) had an amazingly kludgy
link with its basic-like programming language.  Powerbuilder was, when
we were evaluating it, the best of the lot.

--
William Harris                                  (Standard disclaimers)
CARS Information Systems
4000 Executive Park Drive
Cincinnati, OH  45241

 
 
 

Powerbuilder 3.0a GPF's

Post by Jan C. Zawadz » Wed, 14 Sep 1994 12:50:50



>(stuff deleted)
>While it may be urban legend, I believe that Microsoft has a vested
>interest in making sure that their products run better than their
>competitors.  While they may not actively try to put bugs in their
>competitors products it doesn't take much imagination to think that
>VB and other Microsoft developers have access to information (e.g.
>"undocumented features" of windows) that non-Microsoft developers don't.

Oh, give it a break...  While I have little doubt that their OS side trades
info with the app developers (during the company picnic, or whatever), do you
REALLY think that Microsoft is giving special treatment to VB developers?  
If you compare what PB does for its developers, well, yes - MS does have a
superior (IMHO) developer program.  This simply means that this is yet another
facet of PB that sucks, not that MS is giving away the family jewels to anyone
who decides to test drive VB.  Even if they did, that in my opinion would be
just the example (kick?) software companies need...  Ice PB - that should
convince you that majority of the problems are NOT with the MS side of the
system.

Quote:>Mike

Jan
 
 
 

Powerbuilder 3.0a GPF's

Post by Jon Kvis » Wed, 14 Sep 1994 23:21:03


I totally agree that PB 3.0a is an unstable product. I have been struggeling
with a (random) GPF which occurs now and then when opening new MDI-childs. The
application is developed for a customer, and they (and we) have put a lot of
effort and hours into testing our application and trying to find the problem.

Last week our customer was able to get out new DLLs from PowerSoft (Sorry, I
dont know the exact version number). After installing the new DLLs and
recompiling our original source, the GPFs has wanished. This leads me to
conclude that PowerSoft should concentrate on making a stable PB3.x version
(which they might have done now), before rushing out a 4.0 version.

As mentioned by others, we have also chosen PB because we think it has a great
potential. I really hope that we haven't chosen a developing tool that has
great functionallity but newer become stable!


 
 
 

Powerbuilder 3.0a GPF's

Post by Craig Wagn » Wed, 14 Sep 1994 23:19:00



Quote:>Oh, give it a break...  While I have little doubt that their OS side trades
>info with the app developers (during the company picnic, or whatever), do you
>REALLY think that Microsoft is giving special treatment to VB developers?

I don't know.  Do you know they don't?  It would seem in Microsoft's best
interests to ensure their application products work better than anyone else's
with Windows, given the small margins on software these days.

Quote:>If you compare what PB does for its developers, well, yes - MS does have a
>superior (IMHO) developer program.  This simply means that this is yet another
>facet of PB that sucks

Ah, now I understand, you've actually worked for both companies.  You must
have, in order to make the above statement.

Quote:>Ice PB - that should
>convince you that majority of the problems are NOT with the MS side of the
>system.

I thought this was the PowerBuilder newsgroup, not the Microsoft or Visual
Basic advocacy group.  My newsreader must be messed up.
--


Calgary, AB
 
 
 

Powerbuilder 3.0a GPF's

Post by Corwi » Fri, 16 Sep 1994 21:51:05



writes:

I agree with the majority that most GPF's are caused by bad or improper
coding techniques.  When I began working with the product some 6 months
ago I encountered many GPF'S.  However, after spending time on the phone
with the support staff at Powersoft I was able to eliminate the problem by
correcting errors in my code or by downloading a bugfix from the BBS.
Powerbuilder is now in release 3.0a03.  The 30a03 disks are available on
the BBS for download.  Once these fixes were installed all of the problems
seemed to have gone away.  Since powerbuilder provides flexible links to
the WindowsAPI via external function's, it is very easy to cause problems.
 It is not Powersoft's responsibility to ensure that its users can't write
invalid calls to the API.  If you don't know what you are doing READ.  

Corwinm

 
 
 

Powerbuilder 3.0a GPF's

Post by Michael Parks Swa » Sat, 17 Sep 1994 00:20:48




>>Oh, give it a break...  While I have little doubt that their OS side trades
>>info with the app developers (during the company picnic, or whatever), do you
>>REALLY think that Microsoft is giving special treatment to VB developers?

>I don't know.  Do you know they don't?  It would seem in Microsoft's best
>interests to ensure their application products work better than anyone else's
>with Windows, given the small margins on software these days.

  And they'd be shooting themselves in the foot. When their products start
using undocumented functions, they're going to have to support those functions
for a long time.

Quote:>>Ice PB - that should
>>convince you that majority of the problems are NOT with the MS side of the
>>system.

>I thought this was the PowerBuilder newsgroup, not the Microsoft or Visual
>Basic advocacy group.  My newsreader must be messed up.

  When people start making random statements about Microsoft sabatoging
other development platforms, I think that rebuttals are in order. In
particular, if ICE sees PB doing bad things in Windows, it is sloppy
coding on PowerSoft's part, not MicroSoft employees sneaking in late at
night and modifying code.
--
Mike "Agent of Chaos" Swaim

  If you know why a gay person is like a Visa Gold Card,
PLEASE TELL ME.
 
 
 

Powerbuilder 3.0a GPF's

Post by Trent Andre » Sat, 17 Sep 1994 03:56:52



>Path:
>uuneo.neosoft.com!news.sesqui.net!darwin.sura.net!howland.reston.ans.net!europa.
>eng.gtefsd.com!MathWorks.Com!yeshua.marcam.com!usc!cs.utexas.edu!uunet!newstf01.
>cr1.aol.com!search01.news.aol.com!not-for-mail

>Newsgroups: comp.soft-sys.powerbuilder
>Subject: Re: Powerbuilder 3.0a GPF's
>Date: 15 Sep 1994 08:51:05 -0400
>Organization: America Online, Inc. (1-800-827-6364)
>Lines: 16



>NNTP-Posting-Host: search01.news.aol.com

>writes:
>I agree with the majority that most GPF's are caused by bad or improper
>coding techniques.  When I began working with the product some 6 months
>ago I encountered many GPF'S.  However, after spending time on the phone
>with the support staff at Powersoft I was able to eliminate the problem by
>correcting errors in my code or by downloading a bugfix from the BBS.
>Powerbuilder is now in release 3.0a03.  The 30a03 disks are available on
>the BBS for download.  Once these fixes were installed all of the problems
>seemed to have gone away.  Since powerbuilder provides flexible links to
>the WindowsAPI via external function's, it is very easy to cause problems.
> It is not Powersoft's responsibility to ensure that its users can't write
>invalid calls to the API.  If you don't know what you are doing READ.  
>Corwinm

I agree.  Although sometimes thing don't work the way you think they should.

A have discovered that many problems are caused by the interaction of
PB, Windows and a funky video, network or database driver.  Unfourtunately
the a client/server application can be screwed up by any bug in the whole
system (front end to back end).  It is tough to find the guilty party.

 
 
 

Powerbuilder 3.0a GPF's

Post by Craig Wagn » Sat, 17 Sep 1994 23:36:23



Quote:>A have discovered that many problems are caused by the interaction of
>PB, Windows and a funky video, network or database driver.  Unfourtunately
>the a client/server application can be screwed up by any bug in the whole
>system (front end to back end).  It is tough to find the guilty party.

And typically, because PB is the piece you actually see, it gets the blame.
--


 
 
 

Powerbuilder 3.0a GPF's

Post by Anthony Stee » Sun, 18 Sep 1994 06:37:31




>Subject: Re: Powerbuilder 3.0a GPF's
>Date: 15 Sep 1994 08:51:05 -0400

>writes:
>I agree with the majority that most GPF's are caused by bad or improper
>coding techniques.  

Can someone please be specific about what "bad or improper coding techinques"
means? This is not a language like C  where you can write past the end of an
array or pretend that a char is a float. Heck, there aren't even any
pointers. Is there some part of the PowerScript language that we should avoid
or watch out for?
GPFs at runtime are a rather unfriendly form of compiler error message, if
indeed there is something wrong with the code.

Quote:> It is not Powersoft's ,responsibility to ensure that its
> users can't write invalid calls to the API.  

I agree. This is the only way that I can think of right now that a GPF could
result from programmer error. It always pays to take extra care with
interfaces between software written in different languages.
 
 
 

Powerbuilder 3.0a GPF's

Post by Ebbe Jonss » Tue, 20 Sep 1994 22:33:34




> writes:
> I agree with the majority that most GPF's are caused by bad or improper
> coding techniques.  When I began working with the product some 6 months

The majority, maybe, but there are several which do not seem to have anything
to do with coding techniques; one bug in particular that crops up every
once in a while when working with MDI child windows and menus that have
multiple inheritance levels.

Quote:>  It is not Powersoft's responsibility to ensure that its users can't write
> invalid calls to the API.  If you don't know what you are doing READ.  

But it is Powersoft's responsibility to ensure that code written within
the guidelines set forth by them does not GPF (randomly), especially when
the code does nothing more than a simple OpenSheet.

--

DBM Database Marketing Oy, Helsinki, FINLAND    
Tel: +358 0 605 900, Fax: +358 0 605 991                 " Why me? "    

 
 
 

Powerbuilder 3.0a GPF's

Post by Anthony Stee » Fri, 23 Sep 1994 10:14:03




>Subject: Re: Powerbuilder 3.0a GPF's
>Date: 21 Sep 1994 10:26:02 -0400

>Steele) writes:
>> Can someone please be specific about what "bad or improper coding
>> techinques" means? This is not a language like C  where you can write
>One that took our group for a loop when we first started with PB (and this
>was also our first Windows development) was putting something in the
>ItemChanged event that caused another ItemChanged event.  It doesn't take
>long for the GPF to happen (and causes chaos with the de*, too).  The
>warnings are there in the manual, so we have only ourselves to blame for
>not RTFM.  

This is worth taking note of. We have had this one before, but it does not
qualify for the GPF's that we now experience on 2 counts.
1) As noted in another post, the message is "stack overflow" not GPF. This is
an sure indication that the program has gone recursive and run out of stack
space.
2) This one occurs pretty repeatably. The problem we have can manifest
eratically, and is not dependant on the user's actions.

Let me rephrase the original question. We can distinguish two kinds of "Bad or
improper coding techinques":

1) Things that are legal at compile time, but cause (usually) repeatable
problems at runtime. These are understandable in terms of the
powerbuilder environment.These include:

 - Performing an action in event A that triggers event A (instant stack
overflow)
- Posting to event A from the closequery event. The event script for A will
run after the window & instances are destoyed. If the script for event A
references instance vars/etc of the window, they will not be valid when the
script for event A fires.

2) Things that put PB into "random GPF mode". These are things that
SHOULD work, but do not
These include:

- Not regenerating and optimising before  building the EXE
- In one case, extensive use of dwModify
- perhaps our current problem

When I hear the word gun I reach for my art

Anthony Steele

===x===

 
 
 

Powerbuilder 3.0a GPF's

Post by Michael Parks Swa » Sun, 25 Sep 1994 06:23:01



>- Posting to event A from the closequery event. The event script for A will
>run after the window & instances are destoyed. If the script for event A
>references instance vars/etc of the window, they will not be valid when the
>script for event A fires.

  This should not cause a GPF. I have a window with a long closeQuery event
(it involves closing a dde connection with another app that our app abused
greatly), and if an event's executing after the window closes, we just get
a trappable runtime error.
--
Mike "Agent of Chaos" Swaim

  If you know why a gay person is like a Visa Gold Card,
PLEASE TELL ME.
 
 
 

1. Gpf In Powerbuilder V3.0a


 BN> Dear Netters:
 BN>
 BN> Has anybody else come across this problem?  When I change the
 BN> location
 BN> of the powerbar/toolbars with my mouse it causes powerbuilder to
 BN> have a GPF (General Protection Fault) and shuts down.  I have
 BN> version
 BN> 3.0a.  I just bought it yesterday.  I am running it on a Gateway
 BN> 486/25sx,
 BN> SVGA, 8 meg of ram, windows v3.11.
I have a similar machine spec, still a Gateway but a DX and also have 8 megs
of ram.  I haven't run into this problem. We also intalled another 8 users and
no reports there either.  It's more likely a windows setup problem.  having
said that we're using msdos 6.1 if that's any differnt to you?  BN>

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

--- VIPER v2.0

2. Verizon news

3. GPF in PowerBuilder v3.0a

4. Problem Disk?

5. GPF's after upgrade to 3.0a

6. Question regarding emacs 18.59

7. Powerbuilder newbie fears GPF's

8. anyone know hot to scan for names if you only know the IP adres?

9. GPF's in MDI PowerBuilder Application LIbrary application

10. GPF, GPF, GPF

11. HELP: GPF, GPF, GPF,...

12. GPF in PB3.0a

13. HELP!!! GPF in PB3.0a