Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > comp.lang.clipper.visual-objects > #4952 > unrolled thread

Error handling seems replaced by quit without warning

Started byD.J.W. van Kooten <public@ic2remove.this.com>
First post2013-03-22 20:08 +0100
Last post2013-03-28 17:31 -0500
Articles 10 — 4 participants

Back to article view | Back to comp.lang.clipper.visual-objects


Contents

  Error handling seems replaced by quit without warning D.J.W. van Kooten <public@ic2remove.this.com> - 2013-03-22 20:08 +0100
    Re: Error handling seems replaced by quit without warning "Jamal" <jamal@examples.net> - 2013-03-25 15:26 -0400
      Re: Error handling seems replaced by quit without warning D.J.W. van Kooten <public@ic2remove.this.com> - 2013-03-27 15:21 +0100
        Re: Error handling seems replaced by quit without warning "Jamal" <jamal@examples.net> - 2013-03-27 10:58 -0400
          Re: Error handling seems replaced by quit without warning D.J.W. van Kooten <public@ic2remove.this.com> - 2013-03-27 21:01 +0100
            Re: Error handling seems replaced by quit without warning "Jamal" <jamal@examples.net> - 2013-03-27 18:34 -0400
              Re: Error handling seems replaced by quit without warning D.J.W. van Kooten <public@ic2remove.this.com> - 2013-03-28 06:22 +0100
        Re: Error handling seems replaced by quit without warning "Wolfgang Riedmann" <wriedmann@gmail.com> - 2013-03-27 16:12 +0100
    Re: Error handling seems replaced by quit without warning D.J.W. van Kooten <public@ic2remove.this.com> - 2013-03-28 21:13 +0100
      Re: Error handling seems replaced by quit without warning "Willie Moore" <williem@wmconsulting.com> - 2013-03-28 17:31 -0500

#4952 — Error handling seems replaced by quit without warning

FromD.J.W. van Kooten <public@ic2remove.this.com>
Date2013-03-22 20:08 +0100
SubjectError handling seems replaced by quit without warning
Message-ID<vtapk8l4llqu0hlqru4up765fq1a0nr06s@4ax.com>
This I also posted in the vps NG, maybe someone reading this NG
recognizes this:

Lately I noticed that I hardly (or never?) see my error handler if
something goes wrong. Instead the program
directly quits, e.g. when I subtract 2 dwords to a <0 value, or run a
non present method of a class because I accidentaly defined the wrong
class.

I am not aware if I changed something. Nor my adapted error handler,
nor the default VO error handler (if I remove the lib) ever comes up
anymore. Compile settings have been unchanged.

Is there anything I can check/set to get my normal error handling
back?

Dick

[toc] | [next] | [standalone]


#4961

From"Jamal" <jamal@examples.net>
Date2013-03-25 15:26 -0400
Message-ID<kiq8da$tu3$1@speranza.aioe.org>
In reply to#4952
1. What compiler options do you have for your application?  Do you have high 
optimization checked? If so, set it to None.
2. Turn off any run time checks.

HTH,
Jamal

"D.J.W. van Kooten"  wrote in message 
news:vtapk8l4llqu0hlqru4up765fq1a0nr06s@4ax.com...

This I also posted in the vps NG, maybe someone reading this NG
recognizes this:

Lately I noticed that I hardly (or never?) see my error handler if
something goes wrong. Instead the program
directly quits, e.g. when I subtract 2 dwords to a <0 value, or run a
non present method of a class because I accidentaly defined the wrong
class.

I am not aware if I changed something. Nor my adapted error handler,
nor the default VO error handler (if I remove the lib) ever comes up
anymore. Compile settings have been unchanged.

Is there anything I can check/set to get my normal error handling
back?

Dick 

[toc] | [prev] | [next] | [standalone]


#4973

FromD.J.W. van Kooten <public@ic2remove.this.com>
Date2013-03-27 15:21 +0100
Message-ID<nhv5l81pke68u3s7vcsi2emtsp5j4eq3vq@4ax.com>
In reply to#4961
On Mon, 25 Mar 2013 15:26:05 -0400, "Jamal" <jamal@examples.net>
wrote:

>1. What compiler options do you have for your application?  Do you have high 
>optimization checked? If so, set it to None.
>2. Turn off any run time checks.
>

Hello Jamal,

Currently I am totally lost. Yes, your seggestion worked for the
latest location where the program did quit. This are the facts:

The "offending" code is:
cField:=SELF:oTPRELATIEVENSTER:odcmNaam:Value

cField is a string, SELF:oTPRELATIEVENSTER:odcmNaam:Value is a string
from an exported class variable. It quits when run with high
optimization.

But it works 100% fine (reading the value in the variable, so no run
time error ) when I:

-put opimization to none
-check debug on for the method in which it went wrong

_Debout shows ERROR IN NOIVARGET VALUE (I think that's something in
the VO libs) when it quits (and not when I set optimization to none or
debug on)....

I've even rebuild my whole repo, exporting all aef's and importing
them in a new fresh repo (but haven't had any incident with the repo
since 2.8). 

Then I added a textbox in the same method showing 1/n where n = 0. The
program again quit without error....

Does this make any sense to someone?

Dick

[toc] | [prev] | [next] | [standalone]


#4974

From"Jamal" <jamal@examples.net>
Date2013-03-27 10:58 -0400
Message-ID<kiv1em$7p9$1@speranza.aioe.org>
In reply to#4973
Hi Dick,

The original VO compiler developers wanted to make us feel good about having 
the high optimization features, sadly these feature cause unpredictable 
runtime results as the compiler wants to take short cuts when generating 
code.

Can you duplicate the error handler issue in a small test AEF?

Jamal

"D.J.W. van Kooten"  wrote in message 
news:nhv5l81pke68u3s7vcsi2emtsp5j4eq3vq@4ax.com...

On Mon, 25 Mar 2013 15:26:05 -0400, "Jamal" <jamal@examples.net>
wrote:

>1. What compiler options do you have for your application?  Do you have 
>high
>optimization checked? If so, set it to None.
>2. Turn off any run time checks.
>

Hello Jamal,

Currently I am totally lost. Yes, your seggestion worked for the
latest location where the program did quit. This are the facts:

The "offending" code is:
cField:=SELF:oTPRELATIEVENSTER:odcmNaam:Value

cField is a string, SELF:oTPRELATIEVENSTER:odcmNaam:Value is a string
from an exported class variable. It quits when run with high
optimization.

But it works 100% fine (reading the value in the variable, so no run
time error ) when I:

-put opimization to none
-check debug on for the method in which it went wrong

_Debout shows ERROR IN NOIVARGET VALUE (I think that's something in
the VO libs) when it quits (and not when I set optimization to none or
debug on)....

I've even rebuild my whole repo, exporting all aef's and importing
them in a new fresh repo (but haven't had any incident with the repo
since 2.8).

Then I added a textbox in the same method showing 1/n where n = 0. The
program again quit without error....

Does this make any sense to someone?

Dick 

[toc] | [prev] | [next] | [standalone]


#4979

FromD.J.W. van Kooten <public@ic2remove.this.com>
Date2013-03-27 21:01 +0100
Message-ID<3nj6l8pma17c7c3n02js7c3tig30vvf3fn@4ax.com>
In reply to#4974
On Wed, 27 Mar 2013 10:58:03 -0400, "Jamal" <jamal@examples.net>
wrote:

Hello Jamal, Wolfgang,

Thanks for thinking with me.
>
>The original VO compiler developers wanted to make us feel good about having 
>the high optimization features, sadly these feature cause unpredictable 
>runtime results as the compiler wants to take short cuts when generating 
>code.
>
>Can you duplicate the error handler issue in a small test AEF?
>

Not yet. The big problem is that the "errorspots" do not make sense
anyway so if I would try to move the 'offending' code to another place
it may work perfectly well there. E.g. the division by 0 quits on my
main library (in this case with and without debug on), makes VO hang
when moved to a basic test AEF and gave an error (not a runtime error
but a VO library error) when the program was imported on my employees
PC in basically the repo which he takes over from me regularely.

When this was VO 2.6 I would think a repo problem, but 2.8 SP4, and
after a full reimport? I still have no clue. 

Dick

[toc] | [prev] | [next] | [standalone]


#4980

From"Jamal" <jamal@examples.net>
Date2013-03-27 18:34 -0400
Message-ID<kivs6m$o6n$1@speranza.aioe.org>
In reply to#4979
Dick,

Why not trap the division by zero?

May be with something like this in your error handler:

IF oError:GenCode == EG_ZERODIV
    RETURN 0
ENDIF

Jamal

"D.J.W. van Kooten"  wrote in message 
news:3nj6l8pma17c7c3n02js7c3tig30vvf3fn@4ax.com...

On Wed, 27 Mar 2013 10:58:03 -0400, "Jamal" <jamal@examples.net>
wrote:

Hello Jamal, Wolfgang,

Thanks for thinking with me.
>
>The original VO compiler developers wanted to make us feel good about 
>having
>the high optimization features, sadly these feature cause unpredictable
>runtime results as the compiler wants to take short cuts when generating
>code.
>
>Can you duplicate the error handler issue in a small test AEF?
>

Not yet. The big problem is that the "errorspots" do not make sense
anyway so if I would try to move the 'offending' code to another place
it may work perfectly well there. E.g. the division by 0 quits on my
main library (in this case with and without debug on), makes VO hang
when moved to a basic test AEF and gave an error (not a runtime error
but a VO library error) when the program was imported on my employees
PC in basically the repo which he takes over from me regularely.

When this was VO 2.6 I would think a repo problem, but 2.8 SP4, and
after a full reimport? I still have no clue.

Dick 

[toc] | [prev] | [next] | [standalone]


#4981

FromD.J.W. van Kooten <public@ic2remove.this.com>
Date2013-03-28 06:22 +0100
Message-ID<n7k7l859dcp2d1v1077li9bq4amggg5pd5@4ax.com>
In reply to#4980
On Wed, 27 Mar 2013 18:34:37 -0400, "Jamal" <jamal@examples.net>
wrote:

Hello Jamal,
>
>Why not trap the division by zero?
>
The division by zero was not the problem. I just added it to force  a
run time error.

Dick

[toc] | [prev] | [next] | [standalone]


#4975

From"Wolfgang Riedmann" <wriedmann@gmail.com>
Date2013-03-27 16:12 +0100
Message-ID<arggi7FmtbnU1@mid.individual.net>
In reply to#4973
Hi Dick,

> I've even rebuild my whole repo, exporting all aef's and importing
> them in a new fresh repo (but haven't had any incident with the repo
> since 2.8). 
> 
> Then I added a textbox in the same method showing 1/n where n = 0. The
> program again quit without error....
> 
> Does this make any sense to someone?

do you have tried to run it with the debug flag set?

Maybe you are doing something that corrupts the runtime so much, that
even the error handler crashes.

Wolfgang


-- 

[toc] | [prev] | [next] | [standalone]


#4985

FromD.J.W. van Kooten <public@ic2remove.this.com>
Date2013-03-28 21:13 +0100
Message-ID<3099l8t4ba41neuh2i2f6dud58bch08i4k@4ax.com>
In reply to#4952
On Fri, 22 Mar 2013 20:08:35 +0100, D.J.W. van Kooten
<public@ic2remove.this.com> wrote:

I've found it, and the answer is quite obvious I must admit. Months
ago, I added an error handler somewhere in a shell subclass to deal
with a specific error. What it did however was "eat all my errors". 

Removed the error handler and the problem is gone....

Dick

[toc] | [prev] | [next] | [standalone]


#4986

From"Willie Moore" <williem@wmconsulting.com>
Date2013-03-28 17:31 -0500
Message-ID<kj2gce$mtf$1@speranza.aioe.org>
In reply to#4985
Dick,

I am glad that you got it solved.

Regards,
Willie
Vulcan VIP

"D.J.W. van Kooten"  wrote in message 
news:3099l8t4ba41neuh2i2f6dud58bch08i4k@4ax.com...

On Fri, 22 Mar 2013 20:08:35 +0100, D.J.W. van Kooten
<public@ic2remove.this.com> wrote:

I've found it, and the answer is quite obvious I must admit. Months
ago, I added an error handler somewhere in a shell subclass to deal
with a specific error. What it did however was "eat all my errors".

Removed the error handler and the problem is gone....

Dick 

[toc] | [prev] | [standalone]


Back to top | Article view | comp.lang.clipper.visual-objects


csiph-web