Julien Schueller

hello @mleair, would it be possible to post the is_empty_section function code ? I would like to try to upstream a patch

seems good

a rebase is needed

hello @mbaudin47, this seems ready but needs a rebase, can we still target 1.22 ?

+1 but needs a rebase

@mbaudin47 what do we do for i=0 ? -0.5/size can yield inf there too

ok, but it will need a careful review in #2555, as we do not use the sorted sample but just use its size for the grid and then compute empirical...

not pleased with this patch: [qqplot.patch.txt](https://github.com/openturns/openturns/files/14662893/qqplot.patch.txt)

the range for plotting continuous distributions is different than the result of getRange used for numerical methods it uses instead QMin/QMax RM quantiles (to add more margins maybe ?) see...

for an unknown reason the error is noisy, with a spike at the 2nd iteration, the selection stops there:  we can override the default convergence parameters to go further...