| 09-19-2006, 02:28 PM | #1 |
If you have a very simple spell that suffered from the "timer set to null" bug then I would like to take a look on it , the least code and things done in the spell the better. Attach it in a map - not just the code so I could actually test it. All my spells with this problem are really complicated so I can't really see the exact thing that causes the problem. Edit: If you also have one that makes the bug appear with pretty few casts I'd also like to see it. |
| 09-20-2006, 01:46 PM | #2 |
no luck? I guess I will have to try to reproduce it alone... BTW I am days away from finding its cause and solving this problem without any sort of alchemy like pools and things like that |
| 09-20-2006, 01:50 PM | #3 |
I would help out, if I believed this bug exists. Personally I am of the opinion that this is just erroneous use of timers by people(as in the timer is destroyed before it associated objects are done away with). Did a test with several spells on autofire. I went out for a couple hours, and left it running. It managed to create upwards of 300,000 timers(used a count each time I created one), with no visible issues left over from the spells. Edit: Do we have any code we know to cause these issues laying around that we could peruse? |
| 09-20-2006, 01:54 PM | #4 |
this issue exist, modiffy hydra's source to not use a wrapper but use the timer directly and set the timer to null, it will eventually have issues: http://www.wc3campaigns.net/showthread.php?t=80577 |
| 09-20-2006, 02:25 PM | #5 |
The map I attached (swirl spell template thing) in one of the recent threads about this problem is pretty simply coded, and it causes the bug without too many casts. |
| 09-20-2006, 04:36 PM | #6 |
days away you say... neat |
| 09-20-2006, 04:53 PM | #7 | |
Quote:
|
| 09-20-2006, 05:09 PM | #8 |
blade, I can't make it bug, but have you checked your code? There is only one case where the constant function to determine if you set it to null would work. In the other one the if then else is commented and in the main one (which seems to always be which causes the bugs) you are setting t and d to null later even if the boolean is false.. |
| 09-20-2006, 08:15 PM | #9 |
I don't remember the code, it is about 1.5 years old.. But make sure it sets it to null, then it bugs out for me in a few minutes. |