better support set_cycle in drgroup #2016
Conversation
… group is initialized
…oup's cycle boundary. Updated test to test this case.
|
@excaliburtb I lost track of this one over the holiday, I can't recall were you waiting on us to test this branch out in ramtares? |
| long long cycle_tics = (long long)round(rate_in * Trick::JobData::time_tic_value); | ||
| rate_in_tics = cycle_tics; | ||
| next_cycle_in_tics= exec_get_time_tics(); | ||
| if((curr_tic % cycle_tics) != 0) |
There was a problem hiding this comment.
I found an instance of Ramtares users calling dr_group.set_cycle(0.0) in the input file today. It ends up throwing a floating point exception. Will fix on our end but it'd be a nice update to have a check for this to prevent the divide by zero. Not sure what the Trick way of handling this is but an ALL CAPS FLASHING RED TEXT ALERT IS NOT NEARLY ENOUGH.
There was a problem hiding this comment.
lol. sorry that FPE's aren't severe enough of a signal to not do that. I'll add some code to dox the user and have hitman-looking person stop by user's houses.
There was a problem hiding this comment.
it's a good ask. i'll try to include it in this PR
…me_set_cycle_isnt_robust
excaliburtb
force-pushed
the
2015-datalog_run_time_set_cycle_isnt_robust
branch
from
28e4670 to
7d15a78
Compare
… rates and reset the next_cycle tics to the proper value
3 participants
Fixes #2015