Bounty: Find the exact reasons of the "timer set to null" bug

08-05-2006, 12:45 AM	#1
Vexorian	I am going to give 50 rep to anyone that is able to find out the exact reasons of that bug. No I don't want you to reproduce it, anyone can, I have like 5 spells that suffer this bugs, but I want to know the exact reasons.

08-05-2006, 04:13 AM	#2
Rising_Dusk	I think you could put a bounty of 500 rep on it and noone would figure it out. Really, I think it's just caused by a quirk... Or something.

08-05-2006, 07:40 AM	#3
EveningStar	care to explain it first? I've faced a timer bug where pausing and un-pausing a periodic timer will cause it to stop running after 1~3 subsequent iterations. I end up having to just destroy the timer and re-make each time I need to "pause" it.

08-05-2006, 07:43 AM	#4
vile	It's simple, there's a bug that if you start a timer and then you null its variable, the timer will eventually after many casts, will freeze.

08-05-2006, 09:19 AM	#5
Captain Griffen	Maybe keeping (and leaking) a variable assigned to the timer could protect it against being dumped upon by the DestroyTimer bug? EDIT: No, as if you don't make it a variable at all, then it doesn't bug...

08-05-2006, 09:23 AM

Jacek

Quote:

Originally Posted by vile

It's simple, there's a bug that if you start a timer and then you null its variable, the timer will eventually after many casts, will freeze.

What would happen if you don't pause the timer? I never do, just destroy them and never met any bugs ;/

08-05-2006, 09:26 AM	#7
Wyvernoid	Also, if you try to pause a non-started timer, it would directly crash the game (at least on my machine)

08-05-2006, 09:50 AM	#8
vile	It never crashes my pc if i try to pause or destroy a non started or a non existing timer. The only thing that caused a crash for me was using something like DestroyTimer(GetTriggeringTrigger()) or the other way around.

08-05-2006, 09:55 AM	#9
Daelin	Vile, what version of the game you have? I personally never tried this bug on different versions, but I know that I used to have 1.18 and moving an unit outside the map (SetUnitX, SetUnitY) would not cause a crash! ~Daelin

08-05-2006, 10:04 AM	#10
vile	1.20E, the latest, i never tried out that setunitx/y bug, but i do have a function made by someone, i think vex, that gets a safe x/y to set the x/y without crashes.

08-05-2006, 10:09 AM	#11
Daelin	That's not a problem. You can check when the unit gets outside the map and not allow it, that's not difficult. But I could actually move the unit even outside the border and it worked (with 1.18) without crashing at all. In 1.19 I think, it crashed. Sorry for the off-topic. ~Daelin

08-05-2006, 11:01 AM

#12

The)TideHunter(

Quote:

Originally Posted by vile

It never crashes my pc if i try to pause or destroy a non started or a non existing timer. The only thing that caused a crash for me was using something like DestroyTimer(GetTriggeringTrigger()) or the other way around.

Writing:

JASS:

call call DestroyTimer(GetTriggeringTrigger())

Just causes a compile error, it dosent even run.

08-05-2006, 11:18 AM	#13
vile	Ah sorry, not that one. It caused a crash when I was trying to use a function relating to that. My bad.

08-05-2006, 11:24 AM	#14
Daelin	Umm... Could anyone point me to a sample code where the bug takes place? Cuz I do not really know what this is all about... Is it supposed to be something like this? JASS: local timer t = CreateTimer() call TimerStart(t, 0.01, true, function Whatever) set t = null Does function Whatever stop working after a large number of repeats? ~Daelin

08-05-2006, 11:48 AM	#15
Blade.dk	No, it's not that simple. Take a search for the latests posts made by me if you want, there should be a map attached to the other gamecache thread where the problem occurs.