problem with surfaces on tkz-fct

Question

i am using this wonderful tool that tkz-fct is for creating nice documents but I have a weird result from recently. This is my document :

\documentclass[10pt,a4paper]{article}
\usepackage[utf8]{inputenc}
\usepackage[T1]{fontenc}
\usepackage{calc}
\usepackage{color}
\usepackage{pifont}                     %pour utiliser la police Ding
\usepackage[left=1.5cm,right=1.5cm,top=2cm,bottom=2cm]{geometry} %marges du document
\usepackage[french]{babel}
\usepackage[usenames,dvipsnames]{xcolor}
\usepackage[L]{thmbox}
\usepackage{eurosym}
\usepackage{enumitem}               %pour modifer itemize et enumerate
\usepackage{amsmath,amsfonts,amssymb}
\usepackage{mathrsfs}               %pour utiliser mathscr
\usepackage{array,multirow,makecell}
\usepackage{fancyhdr}               %package pour modifier les en-têtes
\usepackage{fancybox}               %pour boîtes spéciales
\usepackage{tikz,tkz-tab,tkz-fct}
\usepackage{tikz-dependency}        %permet de créer des schemas
\usepackage{sectsty}                %permet de modifier les styles des sections (voir \sectionfont l.21)
\usepackage{fourier-orns}
\usepackage{shadethm}               %pour theoremes en gris
\setcellgapes{1pt}
\makegapedcells
\newcolumntype{R}[1]{>{\raggedleft\arraybackslash }b{#1}}   %pour redéfinir les colonnes dans tabular
\newcolumntype{L}[1]{>{\raggedright\arraybackslash }b{#1}}
\newcolumntype{C}[1]{>{\centering\arraybackslash }b{#1}}
\renewcommand{\thesection}{\Roman{section}} %permet de changer la numérotation en chiffres romains des sections
\newshadetheorem{defi}{Définition}  %crée un nouveau type de théorème
\newshadetheorem{thm}{Théorème}
\newshadetheorem{ppte}{Propriété}
\sectionfont{\sectionrule{3ex}{3pt}{-1.5ex}{1pt}} %mettre un trait de séparation dans les sections (package sectsty)
\pagestyle{fancy} %style en-têtes
\lhead{} %en-tête gauche
\chead{Lois à densité} %en-tête centrale
\rhead{T\up{ale}} %en-tête droite
\begin{document}
\begin{tikzpicture}
     \tkzInit[xmin=-4,xmax=4,ymax=1]
     \tkzDrawX[label={},noticks]
     %\tkzDrawY
     \tkzFct[domain=-2:2]{exp(-x**2)}
     \tkzDrawArea[color=blue, domain = -0.5:1]
     \tkzText(-1,0.75){$\mathscr{C}_{f}$}
\end{tikzpicture}
\end{document}

But from now, it doesn't produce the result I would like as the surface start from the point (-0.5,f(-0.5)) and not as an integral from -0.5 to 1. This is a picture of the result :

Which is obviously not what I would like and this extends now to all my "old" docs which were "normal" before. I don't think I have done anything special except updating my Tex distribution ... If someone has an idea ??

Answer 1

\documentclass[10pt,a4paper]{article}
\usepackage{tkz-fct}
\begin{document}
\begin{tikzpicture}
     \tkzInit[xmin=-4,xmax=4,ymax=1]
     \tkzDrawX[label={},noticks]
     \tkzFct[domain=-2:2]{exp(-x**2)}
     \tkzDrawArea[color=blue, domain = -0.5:1]
\end{tikzpicture}
\end{document}

The problem appears with pgf 3.0. The result is correct with pgf 2.1. I need to make some investigations to find the bug

Here a solution but I don't know why the result is different with pgf 3.0

The problem comes from the code in tkz-fct.sty:

        (\tkz@ba,0)--plot [id = \tkz@fct@id]%
       function{(\tkzFctgnuLast)/\tkz@init@ystep}--(\tkz@bb,0);

you need to remplace this code by

    plot [id = \tkz@fct@id]%
    function{(\tkzFctgnuLast)/\tkz@init@ystep} -- (\tkz@bb,0) -- (\tkz@ba,0) ;

or

     plot [id = \tkz@fct@id]%
     function{(\tkzFctgnuLast)/\tkz@init@ystep} -- (\tkz@bb,0) -- (\tkz@ba,0) -- cycle;

Possible is to to find tkz-fct 1.16 c and to replace the code or you can download the version 1.18 c from my site:

tkz-fct 1.18 c

Now I need to understand why ?