Doctest fix for: Sage can't make sense of a Maxima result

[kcrisman]

Mpmath has an inverse erf. Maxima now has an inverse erf. We need it too.

It should be both symbolic and evaluated. !Function_Erf could be used as a model. Make sure to add it to the conversion list in sage.symbolic.pynac! (Which should be elsewhere in any case...)

See this ask.sagemath question for background.

Defect because Sage can't make sense of a Maxima result:

sage: var('z,t')
(z, t)
sage: PDF = exp(-x^2 /2)/sqrt(2*pi)
sage: integralExpr = integrate(PDF,x,z,oo).subs(z==log(t))
sage: y = solve(integralExpr==z,t)[0].rhs().subs(z==1/4)    
sage: y
e^(sqrt(2)*inverse_erf(1/2))
sage: inverse_erf?
Object `inverse_erf` not found.

Depends on #21819

Component: symbolics

Keywords: erf inverse erfinv inverf

Author: Ralf Stephan

Branch/Commit: 2901da8

Reviewer: Travis Scrimshaw

Issue created by migration from https://trac.sagemath.org/ticket/11349

[rwst]

Description changed:

--- 
+++ 
@@ -3,3 +3,18 @@
 It should be both symbolic and evaluated.  !Function_Erf could be used as a model.  Make sure to add it to the conversion list in sage.symbolic.pynac!  (Which should be elsewhere in any case...)
 
 See [this ask.sagemath question](http://ask.sagemath.org/question/549/evaluating-inverse-erf) for background.  
+
+Defect because Sage can't make sense of a Maxima result:
+
+```
+sage: var('z,t')
+(z, t)
+sage: PDF = exp(-x^2 /2)/sqrt(2*pi)
+sage: integralExpr = integrate(PDF,x,z,oo).subs(z==log(t))
+sage: y = solve(integralExpr==z,t)[0].rhs().subs(z==1/4)    
+sage: y
+e^(sqrt(2)*inverse_erf(1/2))
+sage: inverse_erf?
+Object `inverse_erf` not found.
+```
+