octave-interval @0.1.3_0: update to 0.1.4

[Schamschula (Marius Schamschula)]

octave-interval has been updated to version 0.1.4:

0.1.4 2015-03-15
================

 * New interval constructors: hull, midrad

 * New interval arithmetic functions: cbrt, cot, coth, csc, csch, dilog, ei, 
   erf, erfc, gamma, gammaln, psi, rsqrt, sec, sech
 
 * mtimes: Support for fast, less accurate evaluation using BLAS routines
 
 * mldivide, mrdivide, inv: Improved performance by using faster mtimes
 
 * infsup, infsupdec: Enabled broadcasting of scalar boundaries

 * rad: May compute mid and rad simultaneously
 
 * subsref: Access to interval properties using field syntax: x.inf and x.sup

[michaelld (Michael Dickens)]

I'm working on it. It's a little more complicated than just bumping the version and checksums, because they moved to using c++11 -- for which I'm struggling to find a good way to provide consistent support. I"m emailing the MP dev list to see what they think.

[michaelld (Michael Dickens)]

Can you test out the attached patch & see how/if it works for you? I think this will work reasonably generically across OS versions.

[Schamschula (Marius Schamschula)]

Replying to michaelld@…:

Can you test out the attached patch & see how/if it works for you? I think this will work reasonably generically across OS versions.

No problems here. I (re)built it against octave @3.8.2_7+atlas+gcc49+glgui+gui+metis+qtgui.

Note: It built correctly before w/o any of these patches.

[michaelld (Michael Dickens)]

If you're using stdlib=libc++ (which is the default on 10.9+), then your patch should work except in very odd circumstances (e.g., when the user has set "configure.compiler=llvm-gcc-4.2"). Your patch doesn't work for me on 10.8, because I'm using stdlib=libstdc++ (which is the default on 10.8-). So, I need to come up with a standard way to handle using c++11, which is (I think) what my patch does. I've queried the MP dev list & we'll see what comes of it.

[michaelld (Michael Dickens)]

I committed these changes in r134244. I think the code to force use of a C++11 compiler is pretty robust; it probably won't work for 100% of users, but likely the vast vast majority of them.