Re: install problems - ../../mtl/mtl.h:236: error: dependent-name `mtl::dim_n<Vector>::RET' is parsed as a non-type, but instantiation yields atype...

Thanks for the help.  I just joined the list last week and could not
search the archives, just browse it (is this correct or was I just
exceptionally stupid on Friday?) so I didn't find the relevant past
postings on this issue.

Jeff

On 3/26/07, Peter Gottschling <[email protected]> wrote:
> Dear Jeff,
>
> As Gunter said this are known problems.  More precisely, old compilers
> tolerated such potential ambiguities (and usually generated the right code
> ;-), anyway) where newer compilers are stricter.  I should have fixed this
> since a long time and in fact, I sorted out a whole lot of such compiler
> complaints months ago.  On one day this week, I will go through the sources
> and look what I find and apply all the patches sent to the mailing list by
> nice people.  Hopefully, these (simple but annoying) problems will be gone
> then.  I will post a new distribution then and mail it to the list.
>
> Best,
> Peter
>
> P.S.: I'm pleased to learn that MTL is used at UofC.  I lived for 3 months
> in the iHouse in Hyde park.
>
>
> On Mar 25, 2007, at 12:36 AM, Jeff Hammond wrote:
>
> Dear MTL Friends,
>
> I attempted to install MTL on my Xeon EM64T machine with no
> special options:
>
>
> > tar -xzvf mtl-2.1.2-22.tar.gz
> > ./configure --prefix=/usr/local/mtl
> > make
>
> and got this error:
>
> make[2]: Entering directory
> `/usr/local/mtl/mtl-2.1.2-22/contrib/examples'
> g++ -DHAVE_CONFIG_H -I. -I. -I../../mtl   -I../..
> -ftemplate-depth-30 -Wall -c abs_sum.cc
> ../../mtl/mtl.h: In function `typename
> mtl::linalg_traits<Linalg>::magnitude_type
> mtl::one_norm(const
> Vector&, mtl::oned_tag) [with Vector =
> mtl::external_vec<double, 4, unsigned int>]':
> ../../mtl/mtl.h:576:   instantiated from `typename
> mtl::linalg_traits<Linalg>::magnitude_type
> mtl::one_norm(const
> LinalgObj&) [with LinalgObj = mtl::external_vec<double, 4,
> unsigned int>]'
> abs_sum.cc:20:   instantiated from here
> ../../mtl/mtl.h:236: error: dependent-name `
> mtl::dim_n<Vector>::RET' is parsed as a non-type, but
> instantiation yields atype
> ../../mtl/mtl.h:236: note: say `typename
> mtl::dim_n<Vector>::RET' if a type is meant
> make[2]: *** [abs_sum.o] Error 1
> make[2]: Leaving directory
> `/usr/local/mtl/mtl-2.1.2-22/contrib/examples'
> make[1]: *** [all-recursive] Error 1
> make[1]: Leaving directory `/usr/local/mtl/mtl-2.1.2-22/contrib'
> make: *** [all-recursive] Error 1
>
> Neither I, a C++ amateur, nor my office mate, who knows his
> stuff in this regard, have a clue of how to resolve this problem.
>
> This is my machine:  > uname -a
> Linux coulson 2.6.9-42.0.10.ELsmp #1 SMP Fri Feb 16 17:13:42
> EST 2007 x86_64 x86_64 x86_64 GNU/Linux
>
> I attached a TGZ file with the results of 'env', 'configure',
> 'make' and 'g++ -v'.
>
> I also tried and failed repeatedly to use the Intel C++
> compiler instead but changing every single reference to 'g++'
> in every file failed to do the trick and I became confused and
> frustrated at this.
>
> Thanks for any help you can provide.
>
> Jeff Hammond
> The University of Chicago
> <output.tgz>_______________________________________________
> This list is archived at
> http://www.osl.iu.edu/MailArchives/mtl-devel/
>
>
> ------------
>
> Peter Gottschling
>
> Research Associate
>
> Open Systems Laboratory
>
> Indiana University
>
> 135 Lindley Hall
>
> Bloomington, IN 47405
>
> Tel.: +1-812-855-3608   Fax: +1-812-856-0853
>
> http://www.osl.iu.edu/~pgottsch
>
>
> _______________________________________________
> This list is archived at
> http://www.osl.iu.edu/MailArchives/mtl-devel/
>


-- 
Jeff Hammond
The University of Chicago
_______________________________________________
This list is archived at http://www.osl.iu.edu/MailArchives/mtl-devel/