Testing Graphics in a Testing Notebook

Question

Testing Notebooks report the result of Plot as a Failure, when in fact it should be a Success.

Is there a way to fix this?

Chris K · Accepted Answer · 2018-02-13 15:56:29Z

3

This is a similar issue to Testing Grids in a Testing Notebook and has the same solution: add a {SameTest -> (Rasterize[#1] == Rasterize[#2] &)}.

Seems related to the issue raised in Testing equality of graphics.

answered Feb 13, 2018 at 15:56

Chris K

17.6k3 gold badges34 silver badges69 bronze badges

1

$\begingroup$ +1. It does seem inconvenient to have to do this. (Note that rasterizing on tests approximate equality only. E.g. Rasterize[Plot[Sin[x] + 10^-5, {x, 0, 10}]] == Rasterize[Plot[Sin[x], {x, 0, 10}]].) $\endgroup$
– Michael E2
Feb 13, 2018 at 16:41
$\begingroup$ I wonder why it fails. CellChangeTimes? Or something silly like that? Imo a bug. $\endgroup$
– ktm
Feb 13, 2018 at 16:47

Add a comment |

egwene sedai · Accepted Answer · 2018-02-13 19:00:58Z

3

Digging around using FullForm[]shows that the difference is a string-valued private tag

`Charting`Private`Tag$[random number]

Therefore, we may manually set that tag number to, say, 1 via a rule

nullRule = x_String /; StringMatchQ[x, "Charting`Private`Tag$*"] -> 
   "Charting`Private`Tag$1";

then the test runs okay:

answered Feb 13, 2018 at 17:43

egwene sedai

2,29514 silver badges23 bronze badges

$\begingroup$ also: does not work with pasted-in figure as expected output; also strongly suggest compare the data using, e.g. mathematica.stackexchange.com/questions/125222 $\endgroup$
– egwene sedai
Feb 13, 2018 at 19:11
$\begingroup$ See mathematica.stackexchange.com/questions/165652/… for the same solution and others. $\endgroup$
– Michael E2
Feb 13, 2018 at 22:23

Add a comment |

2 Answers 2