[vtkusers] kernel32.dll error with VTK 5.0 and 5.0.2 in vtkInformationStringKey::Set

[Andres Barrera]

Hello Obada,

    Thank you very much for your e-mail. Today I got everything working! I 
downloaded the latest version of CMake and re-compile VTK 5.0.2 in 'Debug' 
mode. Although the VTK headers were OK in my application, I edited my Win 
Environmental Variables (I had declared the dir where my VTK 4.2 libraries 
live at 'PATH').  I don't know what made it, but it is working now :-)  I 
know you are not supposed to change so many things at the same time if you 
want to catch where the problem is... I probably combined all your 
suggestions together!

   Thanks again

          Cheers

                Andres



>From: "Obada Mahdi" <omahdi at gmx.de>
>To: "Andres Barrera" <andresba at hotmail.com>
>CC: lintworm2 at yahoo.co.uk, vtkusers at vtk.org
>Subject: Re: [vtkusers] kernel32.dll error with VTK 5.0 and 5.0.2 in 
>vtkInformationStringKey::Set
>Date: Mon, 23 Oct 2006 22:24:56 +0200

>
>Hi Andres!
>
>On 10/23/06, Andres Barrera <andresba at hotmail.com> wrote:
>>         Thank you very much for your help. Regarding to the error, I'll 
>>try
>>modifying the original vtk implementation for vtkInformationStringValue, 
>>but
>>I don't keep too much hope on it.... The construction looked weird to me
>>too, but this class is being used by many people... Therefore somebody 
>>else
>>should have fixed it before, but I didn't find anything on the users 
>>list...
>>I addition, it is crashing 2 lines before it reaches 'v->Delete();'...
>
>Just to be sure:
>- Did you build VTK in "Debug" mode, or as "Release" or
>"RelWithDebInfo" (something that involves optimization flags)?  WIth
>compiler optimizations, references to lines in the original source
>code can be inaccurate.
>- Is the value of "v" non-NULL just before executing the offending line?
>
>Regarding Han's comment:
>> >vtkInformationStringValue* v = new vtkInformationStringValue;
>[...]
>> >Shouldn't this be:
>> >
>> >vtkInformationStringValue* v = vtkInformationStringValue::New();
>
>vtkInformationStringValue is defined only locally inside
>"vtkInformationStringKey.cxx" and does not adapt the construction
>scheme that involves protected constructors and a static New() method,
>yet it supports the reference counting mechanism inherited from
>vtkObjectBase, so calling Delete() should work just fine.
>
>> >I am not sure, i don't have documentation on me right now, so if i'm 
>>saying
>> >stupid things: Sorry for that... :)
>
>I don't think it is a stupid thing to mention at all.  I suppose that
>mixing up construction conventions for custom VTK classes could indeed
>cause trouble :-)
>
>> > >  Any ideas/suggestions?
>
>Headers and libraries: Any chance that the VTK headers used for
>building the example do not match the version of the libraries used
>for linking?  This can cause weird problems, for example if you
>compile using the "right" headers, but the linker uses libraries from
>a different VTK build.
>
>CMake version: You seem to use a relatively old version of CMake--try
>using a more recent version.
>
>I am not familiar with "C++ Builder", but you can check the dashboard
>(go to www.vtk.org and select "Quality Dashboard") if the combination
>of your system/compiler is actively being tested--if it is, it should
>give you an idea whether it should work or not ;)
>
>
>Regards
>
>Obada

_________________________________________________________________
Try the next generation of search with Windows Live Search today!  
http://imagine-windowslive.com/minisites/searchlaunch/?locale=en-us&source=hmtagline