On 12/27/12 2:09 PM, Jack Howarth wrote:
> Andrew,
> Attached is alternative afni packaging for the 10.7 tree that disables the
> use
> of omp (since this hasn't been implemented for clang yet) and fixes the
> remaining language compatibility issues that causes clang errors. These are...
>
> --- AFNI_2010_10_19_1028/mri_warp.c.orig 2012-12-27 15:25:38.000000000
> -0500
> +++ AFNI_2010_10_19_1028/mri_warp.c 2012-12-27 15:32:03.000000000 -0500
> @@ -23,7 +23,7 @@
>
> static float sx_scale , sy_scale ; /* global scaler data */
>
> -INLINE void xxMRI_scaler( float xpr, float ypr, float *xx , float *yy )
> +static inline void xxMRI_scaler( float xpr, float ypr, float *xx , float *yy
> )
> {
> *xx = sx_scale * xpr ;
> *yy = sy_scale * ypr ;
> @@ -338,7 +338,7 @@
>
> static float rot_dx , rot_dy , rot_cph , rot_sph ; /* global rotfunc data
> */
>
> -INLINE void xxMRI_rotfunc( float xpr , float ypr , float *xx , float *yy )
> +static inline void xxMRI_rotfunc( float xpr , float ypr , float *xx , float
> *yy )
> {
> *xx = rot_cph * xpr + rot_sph * ypr + rot_dx ;
> *yy = -rot_sph * xpr + rot_cph * ypr + rot_dy ;
> --- AFNI_2010_10_19_1028/svm/3dsvm_common.c.orig 2012-12-27
> 15:48:28.000000000 -0500
> +++ AFNI_2010_10_19_1028/svm/3dsvm_common.c 2012-12-27 15:49:46.000000000
> -0500
> @@ -1566,7 +1566,7 @@
> kernelName);
> }
>
> - EXRETURN;
> + RETURN(0);
> }
>
> /* --- allocate memory ---*/
> @@ -1577,7 +1577,7 @@
> Clear2c(maps->names, maps->nmaps);
> Clear2d(maps->data, maps->nmaps, maps->nvox);
>
> - EXRETURN;
> + RETURN(0);
> }
>
> freeModelMaps(MODEL_MAPS *maps)
> @@ -1587,7 +1587,7 @@
> free2d(maps->data, maps->nmaps);
> free2c(maps->names, maps->nmaps);
>
> - EXRETURN;
> + RETURN(0);
> }
>
> void addToModelMap_bucket ( MODEL_MAPS *maps, AFNI_MODEL *afni_model,
>
> The first is http://clang.llvm.org/compatibility.html#inline for C99 inline
> functions causing undefined symbols on linkage and is fixed as recommended by
> making these static. The second is the usage of the wrong return macro in
> 3dsvm_common.c. The attached packaging builds fine with 'fink -m' on
> x86_64-apple-darwin12
> under Xcode 4.5.2.
> Jack
>
>
>
Out of curiosity, is there a noticeable performance hit in not using omp?
--
Alexander Hansen, Ph.D.
Fink User Liaison
My package updates: http://finkakh.wordpress.com/
------------------------------------------------------------------------------
Master HTML5, CSS3, ASP.NET, MVC, AJAX, Knockout.js, Web API and
much more. Get web development skills now with LearnDevNow -
350+ hours of step-by-step video tutorials by Microsoft MVPs and experts.
SALE $99.99 this month only -- learn more at:
http://p.sf.net/sfu/learnmore_122812
_______________________________________________
Fink-devel mailing list
[email protected]
List archive:
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel