Question DG_Script Crash Issue

I am using an older copy of the CircleMud base, and was curious if the issue I am about to describe has been solved in the latest TBA, or if someone is able to point me to where in the code I might address a specific crash related issue I have discovered.

Issue: When running a script with a wait period, if the %actor% being targeted by the script is not present, but is at the main menu of the game, it crashes the mud.

To be more clear about this in examples: Here is a short trigger snippet.

So, in this trigger for example, If by chance the player dies in that wait 1 s period, when the game goes to teleport that %actor%, it crashes. I've attempted identifiers such as if %actor%, and !%actor%, which solved it partially, but not totally. In instances where the arguments are placed, if a character is in game it works fine. If a character dies and disconnects immediately, the trigger ceases and does not crash. If a character dies and is at the main screen of the mud, it crashes. This got me thinking it has something to do with creating a way for the code to detect if a chars descriptor state is CON_PLAYING or not.

Has anyone looked at this, heard of this, have any ideas, etc? It's driving me nuts because I have a lot of triggers I want to use a wait period on to act upon a player, but I don't want to have an unstable mud.

This sounds like a reproducible crash to me. I haven't got a pc for testing it right now, but I'd suggest trying to trigger the bug while running through gdb. There's a post here somewhere about how that's done.

So, I see the error coming from Handler.c's char_from_room function as teleport is directly linked to it.

I made a workaround for the time being which I will simply apply to all dg_scripts acting upon a character at the main screen of the game. As of now, the fix works, but this is not a good fix. It just solves my stability for the time being. With the fix I will include below, it does not kick a player from the trigger trying to fire with a 'wait' argument sitting there, preparing to call on a player. This is an issue because if a player dies, comes back into game, they will still have that function called upon them. But you know... I weighed the cost. Stability or the rare instance where this will happen and a mortal might be upset...? Stability.

From dg_mobcmd.c

Original:
New:

This is a bit weird - when you're at the main menu, you should fail the check in valid_dg_target() (which is called from get_char): I mean, on the main menu, you're not CON_PLAYING!?

And as such, should be enough to filter you out...

I've tested your trigger with the latest tbamud code. This will no longer crash the system, so your version is apparently outdated.

Simple test trigger: I attached it to a room, created a test character and transferred him to just outside the room. Then he entered:
Next, I summoned a death knight for killing the testchar: As you see, all we get are a couple of log warnings.

I just realized, I never replied and said thank you. I applied your fix as soon as you sent it and it worked. Sorry for the 4 month delay, not intentional! Thanks!