Weird behaviour since ROOT 6.18

[RENATO_QUAGLIANI]

Dear all,
I have a fitting code which produce a linearY scale plot just fine
but in logY scale things don’t make much sense
The funny thing is that in ROOT 6.16 all was working smoothly.
Did something has changed on RooCurve ?

Thanks in advance
Renato

---

_ROOT Version: 6.18
_Platform: (tested MacOS, Centos7-gcc8)

---

[RENATO_QUAGLIANI]

To add a reproducer:

rf107_plotstyles.C

If you run this macro in the tutorials and you take the plots produced and setLogY() you will see someething like this ( only for the FilledCurves ) .

[RENATO_QUAGLIANI]

TO add again ,
Adding VLines() to thee Drawing i get :

I don’t think this is expected.

[couet]

I’ll check the rf example you pointed.

[RENATO_QUAGLIANI]

Just to add, if interactively I click thee plot before setting LogY, i get the plot correct. Is there a way to “emulate” a click in a macro?

[couet]

Something looks wrong … let me check.

[couet]

I managed to reproduce the effect with the following simple macro:

{
   TCanvas *C = new TCanvas("C", "C",0,45,800,800);
   C->SetLogy();
   Double_t X[5] = { -11.0, -10.5, -10.0, -9.5, -9.0, };
   Double_t Y[5] = { 0.0000, 0.0080, 0.0085, 0.0090, 0.0095, };
   TGraph *graph = new TGraph(5,X,Y);
   graph->Draw("al*");
}

Looking at it…

[couet]

I found the problem. It is fixed in the ROOT master now.
Thanks to have seen it.

[pamputt]

@couet, since the bug has been introduced by this commit, does it change something about the line clipping improvement introduced by this commit?

[couet]

@pamputt :
Yes, I saw that commit was the culprit.
It is ok. I ran the macro provided in that commit.
The result is still good.

[RENATO_QUAGLIANI]

Is there any workaround for ROOT 6.18 that can be used?
Thanks for finding the commit which was messing up the plot!
I would like to avoid dropping before the next release (we are using ROOT 6.18 on CVMFS for our analysis) the Filled Style plotting.

[couet]

It will be in 6.18/04