I tested your game out again, and actually encountered a fatal error; which usually is caused by the memory (RAM) limit being reached; which would be caused by leaks. The lag wasnt bad up until that point; but it was a BOOM, crash, report error. Maybe 20 minutes in (about level 12) Many of us were clearing 22 bosses. There weren't many warlords, nor spammed units; which means that the lag may not be coming from armies that people make.
TriggerDestroy only decouples events from the function (and sets a var to null) and TriggerCreate is able to create a new trigger and can then be used to register new events. if you redo the lagging part with a more efficient impl. and just register the new event to the function (you might need to check pure map script in order to know the function name for the trigger (gt_sleepTrigger_Func)
trigger gt_sleepTrigger = TriggerCreate("gt_sleepTrigger_Func"); creates a new trigger
TriggerAddEventPlayerEffectUsed(gt_sleepTrigger, -1, "DummyEffect"); adds an event to the trigger
hopefully no map obfuscating is actived since this would make everything a whole lot more complicated.
This method should work to replace any event of any trigger.
Hi, I have also encountered the same lag when creating my map using war3 mod (GA). Before seeing your post, I did some optimization by removing footprints (Renee told me it would work), which reduced the lag to some extent especially for human tank. However, when the game went on to about 1 h and 30 min, it makes the game un-playable. I am trying your solution (Actor set timer), hope it works.
War3_SwapableMissileAttackImpactModel is indeed a bug, I thought I've fixed long ago.(it doesn't need timers to be fixed, actually)
But it seems the one in the battle.net still have this issue. Perhaps due to some data rollback
I will upload a fix, but since the AI mod is also closed to be done, I think it may better to be uploaded with the AI one.(Depended on if the AI can be done in this week.)
On a seconed thought, I just uploaded the new fix right now.
You need to both update the data mod and the extra mod to make it take effect.
This new verion contain lots of fixes which current I don't have time to provide a full list.(for example, all war3 buildings can be built on ramps now)
Would provide detail information after the orc AI lives.
The errors people are posting sound like they are not running in test mode. If you test the map it should log all errors with full path names to the specific data causing it. It is only in multiplayer that simplified logs are used with totally meaningless internal numbering.
I have a strong feeling that the logs themselves can be considered a leak performance wise. In almost all cases where you get a long error log you have very bad map performance. I fixed such a problem for SFE5 a month or so ago with double attack actor creation. However this is still a chicken and egg discussion as it is unclear if the log or what is being logged is the cause of the bad performance.
Hi, have you figured out how to remove Scope[Effect, Effect] B[P XX XX 8.00]? I think it might be this problem. However, I do not know how to do it ~
It's NOT a bug nor anything leaks. Scope[Effect, Effect] are generated by behaviors, Each behavior owns one Scope[Effect, Effect]. Just like each unit needs a unit scope.
Puls, it's only a scope with nothing else in it, it have 0 affect to the perfermance
At least the buffs them selves comsume much more resourse than a thousand empty scopes. Remove unessary buff if you can.
The errors people are posting sound like they are not running in test mode. If you test the map it should log all errors with full path names to the specific data causing it. It is only in multiplayer that simplified logs are used with totally meaningless internal numbering.
I have a strong feeling that the logs themselves can be considered a leak performance wise. In almost all cases where you get a long error log you have very bad map performance. I fixed such a problem for SFE5 a month or so ago with double attack actor creation. However this is still a chicken and egg discussion as it is unclear if the log or what is being logged is the cause of the bad performance.
Basicly the war3 mod itself generates 0 error messages in all sample maps. Unless you develop other extra bad things on it:)
Rollback Post to RevisionRollBack
To post a comment, please login or register a new account.
I tested your game out again, and actually encountered a fatal error; which usually is caused by the memory (RAM) limit being reached; which would be caused by leaks. The lag wasnt bad up until that point; but it was a BOOM, crash, report error. Maybe 20 minutes in (about level 12) Many of us were clearing 22 bosses. There weren't many warlords, nor spammed units; which means that the lag may not be coming from armies that people make.
Skype: [email protected] Current Project: Custom Hero Arena! US: battlenet:://starcraft/map/1/263274 EU: battlenet:://starcraft/map/2/186418
TriggerDestroy only decouples events from the function (and sets a var to null) and TriggerCreate is able to create a new trigger and can then be used to register new events. if you redo the lagging part with a more efficient impl. and just register the new event to the function (you might need to check pure map script in order to know the function name for the trigger (gt_sleepTrigger_Func)
trigger gt_sleepTrigger = TriggerCreate("gt_sleepTrigger_Func"); creates a new trigger
TriggerAddEventPlayerEffectUsed(gt_sleepTrigger, -1, "DummyEffect"); adds an event to the trigger
hopefully no map obfuscating is actived since this would make everything a whole lot more complicated.
This method should work to replace any event of any trigger.
Hi, I have also encountered the same lag when creating my map using war3 mod (GA). Before seeing your post, I did some optimization by removing footprints (Renee told me it would work), which reduced the lag to some extent especially for human tank. However, when the game went on to about 1 h and 30 min, it makes the game un-playable. I am trying your solution (Actor set timer), hope it works.
War3_SwapableMissileAttackImpactModel is indeed a bug, I thought I've fixed long ago.(it doesn't need timers to be fixed, actually)
But it seems the one in the battle.net still have this issue. Perhaps due to some data rollback
I will upload a fix, but since the AI mod is also closed to be done, I think it may better to be uploaded with the AI one.(Depended on if the AI can be done in this week.)
On a seconed thought, I just uploaded the new fix right now.
You need to both update the data mod and the extra mod to make it take effect.
This new verion contain lots of fixes which current I don't have time to provide a full list.(for example, all war3 buildings can be built on ramps now)
Would provide detail information after the orc AI lives.
The fixes are in all regions now.
Hi, have you figured out how to remove Scope[Effect, Effect] B[P XX XX 8.00]? I think it might be this problem. However, I do not know how to do it ~
The errors people are posting sound like they are not running in test mode. If you test the map it should log all errors with full path names to the specific data causing it. It is only in multiplayer that simplified logs are used with totally meaningless internal numbering.
I have a strong feeling that the logs themselves can be considered a leak performance wise. In almost all cases where you get a long error log you have very bad map performance. I fixed such a problem for SFE5 a month or so ago with double attack actor creation. However this is still a chicken and egg discussion as it is unclear if the log or what is being logged is the cause of the bad performance.
It's NOT a bug nor anything leaks. Scope[Effect, Effect] are generated by behaviors, Each behavior owns one Scope[Effect, Effect]. Just like each unit needs a unit scope.
Puls, it's only a scope with nothing else in it, it have 0 affect to the perfermance
At least the buffs them selves comsume much more resourse than a thousand empty scopes. Remove unessary buff if you can.
Basicly the war3 mod itself generates 0 error messages in all sample maps. Unless you develop other extra bad things on it:)