nan values in a different case other than clp

Dear Developers,

Basic information:

CLASS version 3.2.0
Ubuntu mate 18.04
gcc version 7.5.0
What I did: I am using a modified version of class by implementing a new case other than CLP in w_fld. I did something like this:
```
case PDE:
*w_fld = -1.-(a*(2.*pba->alpha*(a-pba->a_m)+3.*(a-pba->a_m)*(a-pba->a_m))/(1.+pba->alpha*(a-pba->a_m)*(a-pba->a_m)+pba->beta*(a-pba->a_m)*(a-pba->a_m)*(a-pba->a_m))/3.);
printf("w_fld=%g\n",w_fld);
```
and the corresponding modifications according to it. All the compilation worked well , all thanks to well written interactive code of yours.
Problem I am facing: When I run the class from the .ini file including the extra parameters that I included. It was giving me a segmentation fault. So, as suggested earlier by you guys in my previous queries to print all the modifications I made. I did that . Then , I found out that It was giving nan values in some sections. The output was like this:

Reading input parameters -> matched budget equations by adjusting Omega_fld = 0.690026 Running CLASS version v3.2.0 Computing background w_fld=0 dw_over_da_fld=0 integral_fld=0 rho_fld_today=3.5303e-08 w_fld=6.95276e-310 dw_over_da_fld=6.95276e-310 integral_fld=6.95276e-310 rho_fld_ini=nan w_fld=6.95276e-310 dw_over_da_fld=6.95276e-310 integral_fld=6.95276e-310 rho_total=nan p_tot=nan Chose ndf15 as generic_evolver w_fld=0 dw_over_da_fld=0 integral_fld=0 rho_total=nan p_tot=nan Segmentation fault (core dumped) .

I don't know why it is giving nan values. Please suggest me what could be wrong in here. Your help would mean a lot to me.

Thanks in advance, shan.

lesgourg / class_public

nan values in a different case other than clp #476