|
From: | David Bateman |
Subject: | Re: automated (or otherwise) tests for graphics code? |
Date: | Thu, 16 Oct 2008 20:54:12 +0100 |
User-agent: | Mozilla-Thunderbird 2.0.0.16 (X11/20080724) |
John W. Eaton wrote:
We seem to be having a lot of regressions in the graphics code lately, so maybe it is time to start thinking of some way to providetests for these functions.
Hey progress is making a different mistake -)
Unfortunately, I don't have any good way to do automatic tests for most of the graphics functions since the results are visual. But at least we could have an informal agreement that no changes should be pushed unless a make check works and also all the graphics demos run successfully. I know that running all the demos could be tedious, but it is also troublesome to have to keep chasing bugs after nearly every change in the plotting functions... To make it easier to run demos, I added a new function called rundemos that can be used this way: rundemos plot to run all the demos in all the files in a given directory in the path (in this case, the plot subdirectory). Without any arguments, rundemos runs all demos in all .m files in the load path. Comments?
I don't think there is any good way to do this automatically.. Someone has to sit down and look at the output of the demos to see if they are correct.. The problem with contourf that Ben fixed this morning was purely visual, and so there is no way it could have been found with an automated test.
Wasn't there already a "make icheck" target that runs all the demos? Yes it makes sense to limit that to the plotting code as it is there that there are really issues, so I support the proposed change..
D.
jwe
-- David Bateman address@hidden 35 rue Gambetta +33 1 46 04 02 18 (Home) 92100 Boulogne-Billancourt FRANCE +33 6 72 01 06 33 (Mob)
[Prev in Thread] | Current Thread | [Next in Thread] |