Add interface to GSL interpolation functionality

[Jari Häkkinen]

​http://www.gnu.org/software/gsl/manual/html_node/Interpolation.html

The GSLInterpolation class should throw GSL_error not gsl_status.

[Jari Häkkinen]

(In [1643]) Addresses #466 and #425. Added cubic spline interpolation (not complete interface to GSL interpolation stuff).

[Jari Häkkinen]

(In [1644]) Addresses #466. Changed to VectorBase? passing. Added documentation.

[Jari Häkkinen]

(In [1645]) Addresses #466. Changed name of CSpline to CSplineInterpolation.

[Jari Häkkinen]

(In [1646]) Addresses #466 added linear interpolation classes.

[Jari Häkkinen]

(In [1647]) Addresses #466. Added more interpolation algorithms. The full set of GSL 1.11 interpolation methods is now interfaced.

[Jari Häkkinen]

(In [1648]) Addresses #466. Added more GSL functionality, added constructor tests.

[Jari Häkkinen]

More tests are needed before closing this ticket.

[Peter]

Replying to jari:

(In [1647]) Addresses #466. Added more interpolation algorithms. The full set of GSL 1.11 interpolation methods is now interfaced.

Hm, this comment was a bit misleading. Does yat require GSL 1.11? No, I successfully compiled against GSL 1.10. So what is really required? Is GSL 1.8 still sufficient or do we need to update configure and documentation?

[Jari Häkkinen]

Replying to peter:

Replying to jari:

(In [1647]) Addresses #466. Added more interpolation algorithms. The full set of GSL 1.11 interpolation methods is now interfaced.

Hm, this comment was a bit misleading. Does yat require GSL 1.11? No, I successfully compiled against GSL 1.10. So what is really required? Is GSL 1.8 still sufficient or do we need to update configure and documentation?

It may be misleading but I do not state that 1.11 is required, and this was by purpose. It was meant to say that if you use 1.11 you will not have any problems. I based my coding on GSL 1.11, and you have now tested 1.10. Being lazy and I prefer not to dig up 1.8 docs to verify that 1.8 has the same interpolation interface as 1.10 the easy option is to require 1.10.

[Peter]

Replying to jari:

Replying to peter:

Replying to jari:

(In [1647]) Addresses #466. Added more interpolation algorithms. The full set of GSL 1.11 interpolation methods is now interfaced.

Hm, this comment was a bit misleading. Does yat require GSL 1.11? No, I successfully compiled against GSL 1.10. So what is really required? Is GSL 1.8 still sufficient or do we need to update configure and documentation?

It may be misleading but I do not state that 1.11 is required, and this was by purpose. It was meant to say that if you use 1.11 you will not have any problems. I based my coding on GSL 1.11, and you have now tested 1.10. Being lazy and I prefer not to dig up 1.8 docs to verify that 1.8 has the same interpolation interface as 1.10 the easy option is to require 1.10.

Yes, your statement is, of course, correct. I checked the NEWS file in GSL and AFAICS there is no change in interpolation functions (except some bug fixes) after GSL 1.1, so I think we could stay with requirement GSL 1.8.

[Jari Häkkinen]

(In [1655]) Addresses #466. Removed bsearch from the GSL wrapper.

[Peter]

(In [1661]) refs #466: moved destructor to be (pure) virtual and public.

[Peter]

Replying to peter:

(In [1661]) refs #466: moved destructor to be (pure) virtual and public.

Shall we move constructor as well?

[Jari Häkkinen]

Replying to peter:

Replying to peter:

(In [1661]) refs #466: moved destructor to be (pure) virtual and public.

Shall we move constructor as well?

Sure.

[Peter]

(In [1663]) typos in Makefile.am - refs #466

[Jari Häkkinen]

(In [1722]) Addresses #466. Moved default contructor to be public.

[Jari Häkkinen]

(In [1724]) Fixes #466. GSL_error is now thrown on GSL errors. Last thing in this ticket.