Re: [lmi] BOOST_TEST needed for automated GUI testing?

[Greg Chicares]

On 2014-03-18 16:58Z, Václav Slavík wrote:
> 
> On 18 Mar 2014, at 17:45, Greg Chicares <address@hidden> wrote:
[...difference between this command-line option...]
>>  /opt/lmi/bin[0]$./wx_test.exe --ash_nazg --data_path=/opt/lmi/data
[...and this one, which triggers an assertion...]
>>  /opt/lmi/bin[0]$./wx_test.exe --mellon --data_path=/opt/lmi/data
> 
> So what should the GUI-test binary do here? I always use —ash_nazg
> (it’s been a few years, but IIRC it wasn’t possible for me and Vadim
> to run it otherwise?), should the test always run in that mode? Or
> would that be undesirable?

I've followed the GPL advice to display a disclaimer and copyright
notice when a program is started "in the simplest and most usual way".
Specifying '--ash_nazg' is not the "usual way" for end users; it's a
"trapdoor" that (among other things) bypasses that notice. Of course,
*we* habitually use the trapdoor, and it's okay that the GUI-test
binary fails otherwise; so does the 'system_test' target, but end
users should never run any of these tests.