info-inventor-dev
[Top] [All Lists]

Re: OpenInventor on SUN, X error

To: connolsv@xxxxxxxx (Connolly, Steven V.)
Subject: Re: OpenInventor on SUN, X error
From: jlim@xxxxxxxxxxxxxxxxxxxx (Jonathan Lim)
Date: Thu, 8 Feb 2001 14:02:54 -0800 (PST)
Cc: info-inventor-dev@xxxxxxxxxxx ('info-inventor-dev@xxxxxxxxxxx')
In-reply-to: <BB93374116D9D41199200002A51341CD200F71@xxxxxxxxxxxxxxxxxxxx> from "Connolly, Steven V." at Feb 08, 2001 04:13:26 PM
Sender: owner-info-inventor-dev@xxxxxxxxxxx
On Thu Feb  8 13:13:26 2001, connolsv@xxxxxxxx wrote:
> 
>     - the SUN C++ compiler does not like the ".c++" suffix so I had to
>       rename such files to end in ".cc".

How about making symbolic links instead? It's too much of a pain to rename all
the files.

>     - I had to "gut" all the methods in "SoTexture2.c++" since it included
>       "jpeglib.h" which I did not find. (This means SoTexture2 is not
>       usable but links ok).

You should be able to download the JPEG library and build it for Solaris.

>     - I had to add "#include <sunmath.h>" to appropriate src files that call
>       "floorf()", "cosf()", "atanf()", etc. If I get this running on the
>       SUN and it is desirable to migrate these changes back to mother
>       version, I can get a list of files.
>       (We could use "#ifdef SomeSunThing".)

Try including that header in lib/database/include/Inventor/SbBasic.h.

>     - "Too few arguments" in "inventor/lib/interaction/src/draggers"
>       (as well as other places):
>  
>       "SoTranslate2Dragger.cc", line 90: Warning: Too few arguments
>         in macro SO__QUOTE.
>  
>       "Line 90" in the code (notice the missing arg!):
>  
>       "SO_KIT_ADD_CATALOG_ENTRY(translatorSwitch, SoSwitch, TRUE,
>                                 geomSeparator, ,FALSE);
>  
>       I made no changes here. just crossed my fingers.

Picky preprocessor ... see database/include/Inventor/misc/SoBasic.h.

>     - warnings about inexact declarations for callback functions.
>  
>         "SoShape.cc", line 881: Warning (Anachronism): Formal argument
>           CallBackFunc of type extern "C" void(*)() in call to
>           gluTessCallback(GLUtesselator*, unsigned, extern "C" void(*)()) is
>           being passed void(*)().

So it's just saying that it's passing in a void(*)() instead of extern "C"
void(*)()?

>     - Many warnings of the kind:
>  
>           "SoIndexedFaceSet.cc", line 345: Warning: materialBinding hides
>                                        SoIndexedShape::materialBinding.
> 
>        I assume we will have to live with these for a very long time but
>        you would think the compiler takes the appropriate and obvious
> action.

I believe the local variable takes precedence over the one defined in the
parent class, so it shouldn't be a problem.


              Jonathan Lim  _  Silicon Graphics  _  Mountain View
              GPS Graphics     Computer Systems        CA, USA

<Prev in Thread] Current Thread [Next in Thread>