Re: [certi-dev] asserting in RTIA::execute

[Timi Tuohenmaa]

Gosh... That changing break to continue was so stupid that I just woke
up four in the morning out of disappointment and came to tell that
here.
It should have been completely obvious for such a little while loop
that it does exactly same thing to break or continue.

I need to find why that assert hits and is that an actual problem
(release version seems to crash randomly and debug version stops
there, so I think there might be a relation). However this "fix" made
me feel so stupid I had to come here before someone else tells me
that. Now back to bed...


-Timi


2016-05-24 15:57 GMT+03:00 Timi Tuohenmaa <address@hidden>:
> Hi,
>
> I probably found a bug as my fairly large federation sometimes asserts
> in RTIA::execute()'s switch-cases default section (line 142 probably
> in git master).
>
> It seems that tm->_tick_state is TICK_NEXT and after some reading I am
> thinking that
>
> RTIA_federate.cc should have in processOngoingTick function in
> TICK_CALLBACK section on line 1349 a "continue" instead of "break".
> That would go to the TICK_NEXT handling immediately instead of letting
> it out to eventually main loop. It would also make us to do what
> comment after break suggests (goto TICK_NEXT).
>
>
> Kind regards,
> Timi Tuohenmaa