[GRASS-dev] Test report: different message breaks the test

[Vaclav Petras]

On Fri, Oct 16, 2015 at 4:33 PM, Pietro <peter.zamb at gmail.com> wrote:

>
> Il 16 ott 2015 5:38 PM, "Vaclav Petras" <wenzeslaus at gmail.com> ha scritto:
> >
> > Hi,
> >
> > it seems that some change in error message broke the test. I'm not sure
> what should the the general approach in these cases. The test should be
> changed with the message that's clear but the question is how precise we
> want the message test to be. Perhaps we want to test there is some message
> or that there is a string such as "SQL" or name of the column in it. Any
> opinions?
>
> Yes, I can remove the test on the error message, probably I was just too
> pendant...
> The important thing is that an exception is raised!
>
In few cases I was testing that the error message contains e.g. the name of
the raster map which does not exist and caused the error.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/grass-dev/attachments/20151016/2bad9e16/attachment.html>