[3.3.2] Assertion failed in mon-time.c

[Gorbad]

In the current competion (119) I encountered an assertion in

monster/mon-timed.c line 247: timer > 0

I couldn't find a similar issue on rephial, but I seem to remember reading it before on oook.

I have/will keep a savefile if needed.

[CunningGabe]

In the current competion (119) I encountered an assertion in

monster/mon-timed.c line 247: timer > 0

I couldn't find a similar issue on rephial, but I seem to remember reading it before on oook.

I have/will keep a savefile if needed.

Thanks -- this was just fixed recently in 3.4-dev.

Magnate: maybe we should port this change to 3.3.2 and update it for the competition?

In the meanwhile, avoid impact hounds >_>

[Gorbad]

In the meanwhile, avoid impact hounds >_>

Good advice at all times

[Magnate]

Magnate: maybe we should port this change to 3.3.2 and update it for the competition?

Hmmm. I hadn't realised there was another 3.3.2 comp running. I hope that we'll be fixing a 3.4 release date on Saturday, so I'd rather not release a 3.3.3 as well. But if it's going to be a long time then that's a different story. Let's revisit the issue at the weekend.

Meanwhile, was it only the one line that needed fixing? ISTR from PowerWyrm that there were other values in project_m that might end up at 0 by the same sort of accident, but I'm not certain.

[CunningGabe]

Hmmm. I hadn't realised there was another 3.3.2 comp running. I hope that we'll be fixing a 3.4 release date on Saturday, so I'd rather not release a 3.3.3 as well. But if it's going to be a long time then that's a different story. Let's revisit the issue at the weekend.

Meanwhile, was it only the one line that needed fixing? ISTR from PowerWyrm that there were other values in project_m that might end up at 0 by the same sort of accident, but I'm not certain.

I think everything else should be okay other than that one line. In each case, before we call mon_inc_timed, we already check whether the timer increment is nonzero. The only problem with stunning is that we potentially modified that number in between checking and sending the call to mon_inc_timed.