Hi Chris,
I noticed today that several MWorks errors don’t stop the experiment. In particular, if I try to index past the end of a sequence, or index an integer, it shows an ERROR in the console but the expt continues with bad values.
We want to be sure to notice and fix any errors, so stopping the experiment is what we want. (An equal alternative would be sounding a siren or vibrating the floor. Stopping is easier).
I vaguely remember there might be a setting for this. If there’s not, can you add one? That would also remove the need to stick stop_on_failure=‘YES’ to every run_python call we make.
Thanks,
Mark
Hi Mark,
There’s not a setting for this, but I can add one. It’ll probably be a new system variable ( #stopOnError
).
Chris
Sounds great. That means we can set it in XML? Do we use a variable with name=“#stopOnError” default value = “yes”, or just use an assignment on every trial?
Mark
That means we can set it in XML? Do we use a variable with name=“#stopOnError” default value = “yes”, or just use an assignment on every trial?
My thought was to make it like any other configuration variable, meaning you would set it either via MWServer’s preferences window or by editing setup_variables.xml
directly. You could also set it per-experiment by including a variable_assignments
section in your XML.
Chris
Great - having a preference setting for it on MWServer would be good.
thanks,
Mark
Hi Mark,
#stopOnError
and the associated MWServer preference are now in the nightly build.
Cheers,
Chris