bug in doc pvalue

src/juobs_tools.jl

@@ -1196,8 +1196,7 @@ The p-value for a given `\chi^2` is the probability of, given the data you have,
 have the data well described by the fit function. `nmc` is the number of MC samples used to estimate the p-value integral, default is 5000.
 By now it only works with a vector for weights (containing the diagonal of W)
 
-Example:
-
+```@example
 function fit_defs(f::Function,x,W)
     chisq(p,d)=sum((d-f(x,p)).^2 .*W)
     return chisq
@@ -1213,7 +1212,7 @@ wpm[1] = [-1.0,-1.0,4-0,-1.0]
 Q = pvalue(chisq, chi2, value.(up), y, wpm; W = 1.0 ./ err.(y) .^ 2, nmc=10000)
 #Q = pvalue(chisq, chi2, value.(up), y; W = 1.0 ./ err.(y) .^ 2, nmc=10000)
 #Q = pvalue(chisq, chi2, value.(up), y)
-
+```
 """
 function pvalue(chisq::Function,
     chi2::Float64,