RE: ITL 4.0.0-1 with MTL 2.1.2-21

Thanks for your quick reply. I tried the function you provided but it
appears that a bunch more fixing would be required to get it working.

Since I am out of my depth with this code, I have fixed it by going back to
code that used to compile. 

I kept the new version of mtl_amend.h that fixes the linalg_traits of
std::vector. I reverted back to revision 1.3 of mtl_classical_gram_schmidt.h
and then put in an #include for mtl_amend.h

This seems to compile OK.

Do you know if there were any important revisions to
mtl_classical_gram_schmidt.h between version 1.3 and 1.8 before the
itl_to_BLAS mixup? I would like to include them in my current code. 


Regards,

Bryan Leppard
Neotechnology Consultants Ltd.


-----Original Message-----
From: Lie-Quan Lee [mailto:[email protected]] 
Sent: June 13, 2003 12:12 PM
To: General Matrix Template Library (MTL) list
Subject: Re: MTL: ITL 4.0.0-1 with MTL 2.1.2-21

You are right -- Somehow the change made in the classical gram schmidt
class was accidentally mixed with function get_data() required in
itl_to_BLAS interface.

Well, one quick way to fix is have following function:

template <class Vector, class VecRef, class Indexer>
typename Vector::value_type* 
get_data(const mtl::oned_part<Vector,VecRef,Indexer>& v) 
{
  typedef mtl::oned_part<Vector,VecRef,Indexer> V;
  V& vref = const_cast<V& >(v);
  return &(vref.get_ref(0));
}



On Fri, 2003-06-13 at 10:28, Bryan Leppard wrote:
> I am trying to use ITL 4.0.0-1 with MTL 2.1.2-21. It looks like
> changes made in ´itl\interface\detail\mtl_classical_gram_schmidt.h¡
> are not compiling.
> 
>  
> 
> =========================================================
> 
> .\My Documents\Visual Studio
>
Projects\ITL-4.0.0-1\itl\interface\detail\mtl_classical_gram_schmidt.h(146)
: error C2039: 'data' : is not a member of
'mtl::oned_part<Vector,VecRef,Indexer>'
> 
>         with
> 
>         [
> 
>            
>
Vector=mtl::matrix_implementation<mtl::generate_rect<itl::classical_gram_sch
midt<DENSE_VECTOR>::value_type,mtl::column_major,mtl::dense<>,0,0>::storage_
t,mtl::gen_rect_indexer<mtl::column_orien,0,0,mtl::generate_rect<itl::classi
cal_gram_schmidt<DENSE_VECTOR>::value_type,mtl::column_major,mtl::dense<>,0,
0>::size_type>>::OldOneD,
> 
>            
>
VecRef=mtl::matrix_implementation<mtl::generate_rect<itl::classical_gram_sch
midt<DENSE_VECTOR>::value_type,mtl::column_major,mtl::dense<>,0,0>::storage_
t,mtl::gen_rect_indexer<mtl::column_orien,0,0,mtl::generate_rect<itl::classi
cal_gram_schmidt<DENSE_VECTOR>::value_type,mtl::column_major,mtl::dense<>,0,
0>::size_type>>::OldOneD,
> 
>            
>
Indexer=mtl::rect_indexer<mtl::generate_rect<itl::classical_gram_schmidt<DEN
SE_VECTOR>::value_type,mtl::column_major,mtl::dense<>,0,0>::size_type,mtl::c
olumn_orien,0,0>::OneDIndexer
> 
>         ]
> 
> ===================================================================
> 
>  
> 
> To me, it looks like line 146 is using function get_data() to try and
> get a pointer to the internal data of an mtl::one_part<> object. This
> doesnÿt compile because mtl::one_part<> does not have a data() member
> function. In this case, is doesnÿt seem appropriate for the function
> to try and access the internal data. Can you confirm this? 
> 
>  
> 
> I am not quite sure how to fix this problem. Any suggestions?
> 
>  
> 
> Regards,
> 
>  
> 
> Bryan Leppard
> 
> Neotechnology Consultants Ltd.
> 
>  
> 
> 
> 
> ______________________________________________________________________
> _______________________________________________
> This list is archived at http://www.osl.iu.edu/MailArchives/mtl-devel/
-- 
Lie-Quan Lee  (AKA: Rich Lee)          Tel:   650-926-5356
Advanced Computations Department       Fax:   650-926-4603
Stanford Linear Accelerator Center     Email: [email protected]



_______________________________________________
This list is archived at http://www.osl.iu.edu/MailArchives/mtl-devel/