Hi guys,

could someone please try this out:

On the wince 5.0 display based platform which has
KernelDebugger/KITL/CESH/VMini enabled, boot in debug
mode (with KernelDebugger/KITL/CESH/VMini on) and then
try to debug a GUI based app from eVC4 SP4 using KITL transport
and CESH startup server. Just leave it to run and see if it will close
without exceptions.

I've tried that on three different platforms (ARMV4I and x86) using
different templates and corresponding Platform SDKs for eVC. The result
is always the same, when the GUI app is closing (WinMain execution finishes
and control is retured to the caller) this cause first-chance exeption in PB
debugger and eVC debugger shows that app jumped to some absolutely
crazy address (like 0xCCCCC on X86). Once You try to "Pass exception
to debugger" eVC will hang completely and die with "Program performed
an illegal operation and will be closed".

The problem seems to be connected with GUI apps only (both plain API
and MFC), for example apps generated by the eVC app wizard show
this behaviour verry well.

Another thing, the same app that caused exception during closing
in debug mode can be launched and closed without any exception
in non-debug mode (means normal execution).

Absolutely no problems with non-GUI apps, they close just nice
in both debug and non-debug modes.

Is this a known issue for wince 5.0? As far as I remember there
were no such problems in 4.2 and earalier versions.

Btw, there are no problems with debugging GUI apps
from BP itself, only eVC shows this behaviour. And it seems
the problem is KernelDebugger or KITL related.

Re: A little experiment by Voidcoder

Voidcoder
Tue Aug 30 13:04:02 CDT 2005

Anyone please!

The most of You guys use kernel debugger and KITL enabled. This
is just a "1 minute" test. Please help me to understand if it is possible
at all.




"Voidcoder" <voidcoder@yahoo.com> wrote in message news:uCItG8GrFHA.1204@TK2MSFTNGP15.phx.gbl...
> Hi guys,
>
> could someone please try this out:
>
> On the wince 5.0 display based platform which has
> KernelDebugger/KITL/CESH/VMini enabled, boot in debug
> mode (with KernelDebugger/KITL/CESH/VMini on) and then
> try to debug a GUI based app from eVC4 SP4 using KITL transport
> and CESH startup server. Just leave it to run and see if it will close
> without exceptions.
>
> I've tried that on three different platforms (ARMV4I and x86) using
> different templates and corresponding Platform SDKs for eVC. The result
> is always the same, when the GUI app is closing (WinMain execution finishes
> and control is retured to the caller) this cause first-chance exeption in PB
> debugger and eVC debugger shows that app jumped to some absolutely
> crazy address (like 0xCCCCC on X86). Once You try to "Pass exception
> to debugger" eVC will hang completely and die with "Program performed
> an illegal operation and will be closed".
>
> The problem seems to be connected with GUI apps only (both plain API
> and MFC), for example apps generated by the eVC app wizard show
> this behaviour verry well.
>
> Another thing, the same app that caused exception during closing
> in debug mode can be launched and closed without any exception
> in non-debug mode (means normal execution).
>
> Absolutely no problems with non-GUI apps, they close just nice
> in both debug and non-debug modes.
>
> Is this a known issue for wince 5.0? As far as I remember there
> were no such problems in 4.2 and earalier versions.
>
> Btw, there are no problems with debugging GUI apps
> from BP itself, only eVC shows this behaviour. And it seems
> the problem is KernelDebugger or KITL related.
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>



Re: A little experiment by davidka

davidka
Thu Sep 01 13:40:06 CDT 2005

Perhaps a DLL got unloaded too early during process termination. In debug,
all freed memory gets initialized to 0xCC in order to catch bugs related to
referencing freed memory as early as possible. In retail, you don't see the
problem because although the referenced memory is freed, it is still mapped
into the process's address space and it still retains the same contents as
before being freed.

David Kanz [MSFT]
This posting is provided "AS IS" with no warranties, and confers no rights.

--------------------
| From: "Voidcoder" <voidcoder@yahoo.com>
| References: <uCItG8GrFHA.1204@TK2MSFTNGP15.phx.gbl>
| Subject: Re: A little experiment
| Date: Tue, 30 Aug 2005 11:04:02 -0700
| Lines: 64
| X-Priority: 3
| X-MSMail-Priority: Normal
| X-Newsreader: Microsoft Outlook Express 6.00.2900.2180
| X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.2180
| X-RFC2646: Format=Flowed; Response
| Message-ID: <ej8HcHUrFHA.1168@TK2MSFTNGP10.phx.gbl>
| Newsgroups: microsoft.public.windowsce.platbuilder
| NNTP-Posting-Host: mail.karo-electronics.de 213.146.116.110
| Path: TK2MSFTNGXA01.phx.gbl!TK2MSFTNGP08.phx.gbl!TK2MSFTNGP10.phx.gbl
| Xref: TK2MSFTNGXA01.phx.gbl microsoft.public.windowsce.platbuilder:23874
| X-Tomcat-NG: microsoft.public.windowsce.platbuilder
|
| Anyone please!
|
| The most of You guys use kernel debugger and KITL enabled. This
| is just a "1 minute" test. Please help me to understand if it is possible
| at all.
|
|
|
|
| "Voidcoder" <voidcoder@yahoo.com> wrote in message
news:uCItG8GrFHA.1204@TK2MSFTNGP15.phx.gbl...
| > Hi guys,
| >
| > could someone please try this out:
| >
| > On the wince 5.0 display based platform which has
| > KernelDebugger/KITL/CESH/VMini enabled, boot in debug
| > mode (with KernelDebugger/KITL/CESH/VMini on) and then
| > try to debug a GUI based app from eVC4 SP4 using KITL transport
| > and CESH startup server. Just leave it to run and see if it will close
| > without exceptions.
| >
| > I've tried that on three different platforms (ARMV4I and x86) using
| > different templates and corresponding Platform SDKs for eVC. The result
| > is always the same, when the GUI app is closing (WinMain execution
finishes
| > and control is retured to the caller) this cause first-chance exeption
in PB
| > debugger and eVC debugger shows that app jumped to some absolutely
| > crazy address (like 0xCCCCC on X86). Once You try to "Pass exception
| > to debugger" eVC will hang completely and die with "Program performed
| > an illegal operation and will be closed".
| >
| > The problem seems to be connected with GUI apps only (both plain API
| > and MFC), for example apps generated by the eVC app wizard show
| > this behaviour verry well.
| >
| > Another thing, the same app that caused exception during closing
| > in debug mode can be launched and closed without any exception
| > in non-debug mode (means normal execution).
| >
| > Absolutely no problems with non-GUI apps, they close just nice
| > in both debug and non-debug modes.
| >
| > Is this a known issue for wince 5.0? As far as I remember there
| > were no such problems in 4.2 and earalier versions.
| >
| > Btw, there are no problems with debugging GUI apps
| > from BP itself, only eVC shows this behaviour. And it seems
| > the problem is KernelDebugger or KITL related.
| >
| >
| >
| >
| >
| >
| >
| >
| >
| >
| >
| >
| >
| >
| >
|
|
|


Re: A little experiment by voidcoder

voidcoder
Fri Sep 02 18:10:58 CDT 2005

Hi David.

Thnaks for the reply. Do You know any solution or wokaround
for this issue? Personally I plan to switch to PB debugger
(thanks to Steve) for both drivers and eVC project debugging,
but I also received an email from the guy asking if I found a
solution. Hi has the same problem on MIPS platform.


""David Kanz [MSFT]"" <davidka@online.microsoft.com> wrote in message
news:IB8AsSyrFHA.1204@TK2MSFTNGXA01.phx.gbl...
> Perhaps a DLL got unloaded too early during process termination. In debug,
> all freed memory gets initialized to 0xCC in order to catch bugs related
> to
> referencing freed memory as early as possible. In retail, you don't see
> the
> problem because although the referenced memory is freed, it is still
> mapped
> into the process's address space and it still retains the same contents as
> before being freed.
>
> David Kanz [MSFT]
> This posting is provided "AS IS" with no warranties, and confers no
> rights.
>
> --------------------
> | From: "Voidcoder" <voidcoder@yahoo.com>
> | References: <uCItG8GrFHA.1204@TK2MSFTNGP15.phx.gbl>
> | Subject: Re: A little experiment
> | Date: Tue, 30 Aug 2005 11:04:02 -0700
> | Lines: 64
> | X-Priority: 3
> | X-MSMail-Priority: Normal
> | X-Newsreader: Microsoft Outlook Express 6.00.2900.2180
> | X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.2180
> | X-RFC2646: Format=Flowed; Response
> | Message-ID: <ej8HcHUrFHA.1168@TK2MSFTNGP10.phx.gbl>
> | Newsgroups: microsoft.public.windowsce.platbuilder
> | NNTP-Posting-Host: mail.karo-electronics.de 213.146.116.110
> | Path: TK2MSFTNGXA01.phx.gbl!TK2MSFTNGP08.phx.gbl!TK2MSFTNGP10.phx.gbl
> | Xref: TK2MSFTNGXA01.phx.gbl microsoft.public.windowsce.platbuilder:23874
> | X-Tomcat-NG: microsoft.public.windowsce.platbuilder
> |
> | Anyone please!
> |
> | The most of You guys use kernel debugger and KITL enabled. This
> | is just a "1 minute" test. Please help me to understand if it is
> possible
> | at all.
> |
> |
> |
> |
> | "Voidcoder" <voidcoder@yahoo.com> wrote in message
> news:uCItG8GrFHA.1204@TK2MSFTNGP15.phx.gbl...
> | > Hi guys,
> | >
> | > could someone please try this out:
> | >
> | > On the wince 5.0 display based platform which has
> | > KernelDebugger/KITL/CESH/VMini enabled, boot in debug
> | > mode (with KernelDebugger/KITL/CESH/VMini on) and then
> | > try to debug a GUI based app from eVC4 SP4 using KITL transport
> | > and CESH startup server. Just leave it to run and see if it will close
> | > without exceptions.
> | >
> | > I've tried that on three different platforms (ARMV4I and x86) using
> | > different templates and corresponding Platform SDKs for eVC. The
> result
> | > is always the same, when the GUI app is closing (WinMain execution
> finishes
> | > and control is retured to the caller) this cause first-chance exeption
> in PB
> | > debugger and eVC debugger shows that app jumped to some absolutely
> | > crazy address (like 0xCCCCC on X86). Once You try to "Pass exception
> | > to debugger" eVC will hang completely and die with "Program performed
> | > an illegal operation and will be closed".
> | >
> | > The problem seems to be connected with GUI apps only (both plain API
> | > and MFC), for example apps generated by the eVC app wizard show
> | > this behaviour verry well.
> | >
> | > Another thing, the same app that caused exception during closing
> | > in debug mode can be launched and closed without any exception
> | > in non-debug mode (means normal execution).
> | >
> | > Absolutely no problems with non-GUI apps, they close just nice
> | > in both debug and non-debug modes.
> | >
> | > Is this a known issue for wince 5.0? As far as I remember there
> | > were no such problems in 4.2 and earalier versions.
> | >
> | > Btw, there are no problems with debugging GUI apps
> | > from BP itself, only eVC shows this behaviour. And it seems
> | > the problem is KernelDebugger or KITL related.
> | >
> | >
> | >
> | >
> | >
> | >
> | >
> | >
> | >
> | >
> | >
> | >
> | >
> | >
> | >
> |
> |
> |
>



Re: A little experiment by davidka

davidka
Wed Sep 07 19:45:16 CDT 2005

No, I don't know a solution or workaround. I was guessing at what is
happening based on what I know of debug CE and your description of the
problem. I have seen similar instances like this (calling DLL code that is
no longer present) but it was not related to eVC.

David Kanz [MSFT]
This posting is provided "AS IS" with no warranties, and confers no rights.

--------------------
| From: "voidcoder" <voidcoder@yahoo.com>
| References: <uCItG8GrFHA.1204@TK2MSFTNGP15.phx.gbl>
<ej8HcHUrFHA.1168@TK2MSFTNGP10.phx.gbl>
<IB8AsSyrFHA.1204@TK2MSFTNGXA01.phx.gbl>
| Subject: Re: A little experiment
| Date: Sat, 3 Sep 2005 01:10:58 +0200
| Lines: 116
| X-Priority: 3
| X-MSMail-Priority: Normal
| X-Newsreader: Microsoft Outlook Express 6.00.2900.2180
| X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.2180
| X-RFC2646: Format=Flowed; Original
| Message-ID: <u6PoPOBsFHA.3720@TK2MSFTNGP14.phx.gbl>
| Newsgroups: microsoft.public.windowsce.platbuilder
| NNTP-Posting-Host: dsl-082-083-241-191.arcor-ip.net 82.83.241.191
| Path: TK2MSFTNGXA01.phx.gbl!TK2MSFTNGP08.phx.gbl!TK2MSFTNGP14.phx.gbl
| Xref: TK2MSFTNGXA01.phx.gbl microsoft.public.windowsce.platbuilder:24116
| X-Tomcat-NG: microsoft.public.windowsce.platbuilder
|
| Hi David.
|
| Thnaks for the reply. Do You know any solution or wokaround
| for this issue? Personally I plan to switch to PB debugger
| (thanks to Steve) for both drivers and eVC project debugging,
| but I also received an email from the guy asking if I found a
| solution. Hi has the same problem on MIPS platform.
|
|
| ""David Kanz [MSFT]"" <davidka@online.microsoft.com> wrote in message
| news:IB8AsSyrFHA.1204@TK2MSFTNGXA01.phx.gbl...
| > Perhaps a DLL got unloaded too early during process termination. In
debug,
| > all freed memory gets initialized to 0xCC in order to catch bugs
related
| > to
| > referencing freed memory as early as possible. In retail, you don't see
| > the
| > problem because although the referenced memory is freed, it is still
| > mapped
| > into the process's address space and it still retains the same contents
as
| > before being freed.
| >
| > David Kanz [MSFT]
| > This posting is provided "AS IS" with no warranties, and confers no
| > rights.
| >
| > --------------------
| > | From: "Voidcoder" <voidcoder@yahoo.com>
| > | References: <uCItG8GrFHA.1204@TK2MSFTNGP15.phx.gbl>
| > | Subject: Re: A little experiment
| > | Date: Tue, 30 Aug 2005 11:04:02 -0700
| > | Lines: 64
| > | X-Priority: 3
| > | X-MSMail-Priority: Normal
| > | X-Newsreader: Microsoft Outlook Express 6.00.2900.2180
| > | X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.2180
| > | X-RFC2646: Format=Flowed; Response
| > | Message-ID: <ej8HcHUrFHA.1168@TK2MSFTNGP10.phx.gbl>
| > | Newsgroups: microsoft.public.windowsce.platbuilder
| > | NNTP-Posting-Host: mail.karo-electronics.de 213.146.116.110
| > | Path: TK2MSFTNGXA01.phx.gbl!TK2MSFTNGP08.phx.gbl!TK2MSFTNGP10.phx.gbl
| > | Xref: TK2MSFTNGXA01.phx.gbl
microsoft.public.windowsce.platbuilder:23874
| > | X-Tomcat-NG: microsoft.public.windowsce.platbuilder
| > |
| > | Anyone please!
| > |
| > | The most of You guys use kernel debugger and KITL enabled. This
| > | is just a "1 minute" test. Please help me to understand if it is
| > possible
| > | at all.
| > |
| > |
| > |
| > |
| > | "Voidcoder" <voidcoder@yahoo.com> wrote in message
| > news:uCItG8GrFHA.1204@TK2MSFTNGP15.phx.gbl...
| > | > Hi guys,
| > | >
| > | > could someone please try this out:
| > | >
| > | > On the wince 5.0 display based platform which has
| > | > KernelDebugger/KITL/CESH/VMini enabled, boot in debug
| > | > mode (with KernelDebugger/KITL/CESH/VMini on) and then
| > | > try to debug a GUI based app from eVC4 SP4 using KITL transport
| > | > and CESH startup server. Just leave it to run and see if it will
close
| > | > without exceptions.
| > | >
| > | > I've tried that on three different platforms (ARMV4I and x86) using
| > | > different templates and corresponding Platform SDKs for eVC. The
| > result
| > | > is always the same, when the GUI app is closing (WinMain execution
| > finishes
| > | > and control is retured to the caller) this cause first-chance
exeption
| > in PB
| > | > debugger and eVC debugger shows that app jumped to some absolutely
| > | > crazy address (like 0xCCCCC on X86). Once You try to "Pass
exception
| > | > to debugger" eVC will hang completely and die with "Program
performed
| > | > an illegal operation and will be closed".
| > | >
| > | > The problem seems to be connected with GUI apps only (both plain API
| > | > and MFC), for example apps generated by the eVC app wizard show
| > | > this behaviour verry well.
| > | >
| > | > Another thing, the same app that caused exception during closing
| > | > in debug mode can be launched and closed without any exception
| > | > in non-debug mode (means normal execution).
| > | >
| > | > Absolutely no problems with non-GUI apps, they close just nice
| > | > in both debug and non-debug modes.
| > | >
| > | > Is this a known issue for wince 5.0? As far as I remember there
| > | > were no such problems in 4.2 and earalier versions.
| > | >
| > | > Btw, there are no problems with debugging GUI apps
| > | > from BP itself, only eVC shows this behaviour. And it seems
| > | > the problem is KernelDebugger or KITL related.
| > | >
| > | >
| > | >
| > | >
| > | >
| > | >
| > | >
| > | >
| > | >
| > | >
| > | >
| > | >
| > | >
| > | >
| > | >
| > |
| > |
| > |
| >
|
|
|