unable to install Unicorn-0.1.2

Asked by Rishik Ranjan on 2011-06-03

My dolfin version is 0.9.10
     ffc version is 0.9.9
     ufc 2.0.0

Following the instruction on the README file I did the following
1. ./configure
2. make (Compile without running any tests)

But I got the following error message after executing "make"

In file included from AdaptiveRefinementProjectScalar.cpp:1:
../src/unicorn/AdaptiveRefinementProjectScalar.h:398: error: expected ')' before '&' token
../src/unicorn/AdaptiveRefinementProjectScalar.h:410: error: ISO C++ forbids declaration of 'Array' with no type
../src/unicorn/AdaptiveRefinementProjectScalar.h:410: error: invalid use of '::'
../src/unicorn/AdaptiveRefinementProjectScalar.h:410: error: expected ';' before '<' token
../src/unicorn/AdaptiveRefinementProjectScalar.h:415: error: expected ';' before 'private'
../src/unicorn/AdaptiveRefinementProjectScalar.h:421: error: ISO C++ forbids declaration of 'Array' with no type
../src/unicorn/AdaptiveRefinementProjectScalar.h:421: error: invalid use of '::'
../src/unicorn/AdaptiveRefinementProjectScalar.h:421: error: expected ';' before '<' token
AdaptiveRefinementProjectScalar.cpp: In member function 'virtual ufc::dof_map* UFC_AdaptiveRefinementProjectScalarLinearForm_dof_map_0::create_sub_dof_map(unsigned int) const':
AdaptiveRefinementProjectScalar.cpp:1211: error: cannot allocate an object of abstract type 'UFC_AdaptiveRefinementProjectScalarLinearForm_dof_map_0'
../src/unicorn/AdaptiveRefinementProjectScalar.h:167: note: because the following virtual functions are pure within 'UFC_AdaptiveRefinementProjectScalarLinearForm_dof_map_0':
/usr/include/ufc.h:203: note: virtual unsigned int ufc::dof_map::local_dimension(const ufc::cell&) const
/usr/include/ufc.h:206: note: virtual unsigned int ufc::dof_map::max_local_dimension() const
AdaptiveRefinementProjectScalar.cpp: In member function 'virtual ufc::dof_map* UFC_AdaptiveRefinementProjectScalarLinearForm_dof_map_1::create_sub_dof_map(unsigned int) const':
AdaptiveRefinementProjectScalar.cpp:1379: error: cannot allocate an object of abstract type 'UFC_AdaptiveRefinementProjectScalarLinearForm_dof_map_1'
../src/unicorn/AdaptiveRefinementProjectScalar.h:240: note: because the following virtual functions are pure within 'UFC_AdaptiveRefinementProjectScalarLinearForm_dof_map_1':
/usr/include/ufc.h:203: note: virtual unsigned int ufc::dof_map::local_dimension(const ufc::cell&) const
/usr/include/ufc.h:206: note: virtual unsigned int ufc::dof_map::max_local_dimension() const
AdaptiveRefinementProjectScalar.cpp: In member function 'virtual ufc::dof_map* UFC_AdaptiveRefinementProjectScalarLinearForm::create_dof_map(unsigned int) const':
AdaptiveRefinementProjectScalar.cpp:1520: error: cannot allocate an object of abstract type 'UFC_AdaptiveRefinementProjectScalarLinearForm_dof_map_0'
../src/unicorn/AdaptiveRefinementProjectScalar.h:167: note: since type 'UFC_AdaptiveRefinementProjectScalarLinearForm_dof_map_0' has pure virtual functions
AdaptiveRefinementProjectScalar.cpp:1523: error: cannot allocate an object of abstract type 'UFC_AdaptiveRefinementProjectScalarLinearForm_dof_map_1'
../src/unicorn/AdaptiveRefinementProjectScalar.h:240: note: since type 'UFC_AdaptiveRefinementProjectScalarLinearForm_dof_map_1' has pure virtual functions
make[1]: *** [AdaptiveRefinementProjectScalar.lo] Error 1
make: *** [all-recursive] Error 1

Please tell me how to fix this.

Question information

Language:
English Edit question
Status:
Solved
For:
Unicorn Edit question
Assignee:
No assignee Edit question
Solved by:
Niclas Jansson
Solved:
2011-06-07
Last query:
2011-06-07
Last reply:
2011-06-06
Niclas Jansson (njansson) said : #1

Unicorn depends on an older FFC/UFC release and a special branch of DOLFIN. The current dependencies are:

    * ufc (1.1)
    * ffc (0.5.1)
    * dolfin (0.8.1-hpc)
    * BLAS with a working c-interface (cblas, sunperf, atlas, gsl)

Rishik Ranjan (rik-mad) said : #2

Thank you.

I was able to obtain 0.8.1-hpc but I am unable to locate the required ufc and ffc versions online.
Where can I find them?

Rishik Ranjan (rik-mad) said : #4

Thanks Niclas Jansson, that solved my question.