[gme-users] Linking problems with STLPort

Jacques Kerner jacques.kerner at oktal.fr
Fri Jul 9 19:00:10 CDT 2004


Thanks,

Somehow we were able to correct this problem by using the stlport 
installation of the first machine, and using it to replace the one on 
the second machine (where it did not compile). And the error disappears. 
So it must be a different way of compiling the stlport on both machines 
that have lead to the linking problem (as I understand it). I could not 
locate the cause more precisely. As you mention, there must be something 
different in the setting that where used at one time or another.

Sorry for the disturbance, and thank you for your efforts. In case the 
problem reappears, there will be at least some precedent.

Jacques Kerner.

Akos Ledeczi a écrit :

> Are you sure that the global include directory settings in Visual 
> Studio are the same on both machines? Because if they are not, the 
> order it looks for stl stuff in the directories could differ and it 
> could find another stl on the machine you get the link errors. Akos
>
>
> Jacques Kerner wrote:
>
>> Hi,
>>
>> I am having problems linking interpreters for GME. Sometimes it 
>> works, but sometimes, I get the following kind of errors :
>>
>> INIWriter.obj : error LNK2001: external symbol "__declspec(dllimport)
>> public: static bool __cdecl _STL::
>> __char_traits_base<char,int>::eq_int_type(int const &,int const &)"
>> (__imp_?eq_int_type@?$__char_traits_base at DH@_STL@@SA_NABH0 at Z) unresolved
>>
>>
>> (and a lot of them). This code compiles and links perfectly well on a 
>> machine, but installed on another machine, it doesn't link, in spite 
>> of the fact that the installation of STLPort was conducted in the 
>> same fashion, and the include libraries are put in the exact same order.
>>
>> I think it is actually a problem of namespace naming and confusion 
>> between "_STL" and "std". Despite my efforts, I was unable to find a 
>> solution to this problem. How can I prevent this from happening ? Any 
>> hints on this problem are deeply welcome. Thanks in advance!
>>
>> Jacques Kerner.
>>
>> Jeff Parsons a écrit :
>>
>>> Hi,
>>>
>>>  
>>>
>>>> -----Original Message-----
>>>> From: Peter Volgyesi [mailto:peter.volgyesi at vanderbilt.edu] Sent: 
>>>> Thursday, July 08, 2004 5:18 AM
>>>> To: gme-users
>>>> Subject: RE: [gme-users] connections to ports w/ decorator
>>>>
>>>>
>>>> Hi,
>>>>
>>>> Hotspots are not controlled by the decorator. The decorator is able 
>>>> (should) specify only the ports and their locations. The two 
>>>> relevant methods (COM
>>>> level):
>>>>
>>>> HRESULT GetPortLocation([in] IMgaFCO *fco, [out] long *sx, [out] 
>>>> long *sy, [out] long *ex, [out] long *ey); HRESULT GetPorts([out, 
>>>> retval] IMgaFCOs **portFCOs);
>>>> The methods are implemented in the BoxDecorator as
>>>>
>>>> STDMETHODIMP CBoxDecorator::GetPortLocation( IMgaFCO *pFCO, long 
>>>> *sx, long *sy, long *ex, long *ey ); STDMETHODIMP 
>>>> CBoxDecorator::GetPorts( IMgaFCOs **portFCOs );
>>>>
>>>> First, GME calls GetPorts, where you can specify a collection of 
>>>> child objects (visible as ports). Next, GME calls the 
>>>> GetPortLocation() method repeatedly for each port object in that 
>>>> collection.
>>>>
>>>> Ports are automatically handled as hotspots (connectable points) by 
>>>> the GUI along with the primeter of the bounding object.
>>>>
>>>> Hope, this helps.
>>>>   
>>>
>>>
>>>
>>> Yes, it does. I guess I figured that, since my decorator is calling
>>> elsewhere to gets its own port information anyway, and not creating
>>> or destroying ports, that it holds nothing that GME can't get
>>> somewhere else. But even if that were true, GME would still need
>>> the port location info, since a decorator could draw the port
>>> somewhere other than where the default decorator would draw it
>>> (which my decorator does). So, duh, once I implemented those
>>> 2 callback methods in my decorator, it worked fine. Thanks
>>> for the explanation.
>>>
>>> Jeff
>>>
>>>  
>>>
>>>> -- 
>>>> peter
>>>>
>>>>
>>>>  
>>>>
>>>>> -----Original Message-----
>>>>> From: gme-users-bounces at list.isis.vanderbilt.edu
>>>>> [mailto:gme-users-bounces at list.isis.vanderbilt.edu] On Behalf Of 
>>>>> Jeff Parsons
>>>>> Sent: Wednesday, July 07, 2004 7:09 PM
>>>>> To: gme-users
>>>>> Subject: [gme-users] connections to ports w/ decorator
>>>>>
>>>>> Hi,
>>>>>
>>>>> I have written a decorator that takes over the drawing of
>>>>> just a few things in a model, the relevant one being one certain 
>>>>> kind of model and its ports. What the decorator does is choose 
>>>>> between mirror image icons depending on which side of the model 
>>>>> the port is drawn on. These ports can have connections to other 
>>>>> ports in the same kind of model. It seems that, using the 
>>>>> decorator, connections don't appear in the model, whereas without 
>>>>> it they do.
>>>>> I've borrowed some code from GME source
>>>>> code for this stuff, leaving out some things since my use case is 
>>>>> not as general as GME's default decorator. I've looked for 
>>>>> something I left out relating to hotspots but all that stuff seems 
>>>>> to be outside the decorator part of GEM and I don't see how I 
>>>>> could change my decorator code to make this work. Can anyone help 
>>>>> me out with info about what code in ModelDecorator or 
>>>>> PortDecorator controls hotspots, if any?
>>>>>
>>>>> thanks,
>>>>>
>>>>> Jeff
>>>>> _______________________________________________
>>>>> gme-users mailing list
>>>>> gme-users at list.isis.vanderbilt.edu
>>>>> http://list.isis.vanderbilt.edu/mailman/listinfo/gme-users
>>>>>     
>>>>
>>>>
>>>> _______________________________________________
>>>> gme-users mailing list
>>>> gme-users at list.isis.vanderbilt.edu
>>>> http://list.isis.vanderbilt.edu/mailman/listinfo/gme-users
>>>>
>>>>   
>>>
>>>
>>> _______________________________________________
>>> gme-users mailing list
>>> gme-users at list.isis.vanderbilt.edu
>>> http://list.isis.vanderbilt.edu/mailman/listinfo/gme-users
>>>
>>>  
>>>
>> _______________________________________________
>> gme-users mailing list
>> gme-users at list.isis.vanderbilt.edu
>> http://list.isis.vanderbilt.edu/mailman/listinfo/gme-users
>
> _______________________________________________
> gme-users mailing list
> gme-users at list.isis.vanderbilt.edu
> http://list.isis.vanderbilt.edu/mailman/listinfo/gme-users
>


More information about the gme-users mailing list