| 08-20-2006, 02:15 PM | #1 |
For whoever wanted to know, the cause of the desync on my map was a dummy unit with -10000 sight range. When I set that to 0, netsplits stopped. |
| 08-20-2006, 04:57 PM | #2 |
hey the caster system's dummy uses -10000 sight range, I always thought there were at least a couple of guys that used the caster system but I am now starting to doubt. Either that or it is the result of combining the -10000 sight range with an specific situation or trigger. But I would want to know how to make a caster owned by the player but not give information. |
| 08-20-2006, 05:01 PM | #3 |
I'm using caster system, and no server splits. Sounds like you may want to dig a little deeper. |
| 08-21-2006, 02:53 AM | #4 |
Why -1000? -1 or 0 have same effect |
| 08-21-2006, 01:42 PM | #5 |
I used the caster system like, years ago, so the dummy still stayed on my map. Update: there are still rare netsplits, but their frequency has dramatically dropped since. |
| 08-21-2006, 03:57 PM | #6 |
Yeah but did you have the desyncs back then? I think that there might be another thing affecting this. I am gonna still try to fix it, maybe setting casters to do not share vision with the player. |
| 08-21-2006, 07:27 PM | #7 |
yes, i had them back then. btw, a lil off topic, but can you add your optimizer an option NOT to touch the war3map.j? No matter which options i put, it always HAS to touch it. I mean, I got my own obfuscation methods now, and i want it to lay off the j, but there are still useful options there that i'd like to use. |
| 08-21-2006, 07:30 PM | #8 |
It will always clean white space and comments, and I see no reason to let people disable that. That method is 100% safe, and white space and comments are worthless |
| 08-22-2006, 09:37 AM | #9 |
I dont see any reason to use this tool, i can easy read any trigger inside "protected" map (if i can others can do this too) only strings moving from wts file have effect. |
| 08-22-2006, 11:44 AM | #10 | |
Quote:
You're wrong. It decreases the map's size. Period. |
| 08-22-2006, 12:53 PM | #11 | |
Quote:
The 25% of map space it can cut can decide life or death of a bnet map... Then you have using short variables and function names that does make JASS much faster. Or what about merge that decreases the overhead of loading time? |
| 08-22-2006, 01:13 PM | #12 |
I can cut not needed files self. Init merge may cause big problem if using triggers init on fly. No need to init spell triggers for hero before is picked. Mass events realy slow down map. |
| 08-22-2006, 01:21 PM | #13 | |||
Quote:
Quote:
Quote:
|
| 08-22-2006, 05:52 PM | #14 |
Of course you can do everything "yourself". But 50,000 years ago our race noticed this funny thing: you use tools and suddenly you've got a spaceship on the goddamn moon. But, I agree that there ought to be an option to turn off chopping white space/comments or not touch script at all. For example there could be a bug in even something that well tested. |
| 08-22-2006, 06:06 PM | #15 |
The fact is that if I add that option, that would happen in 5 months once I feel like coding in Delphi again. And I don't really have to cause vile can add the new script manually, cause I am guessing he would run his obfuscation methods after optimized otherwise he wouldn't have an issue with it. But if it could fail, the odds that just removing white space would cause issues are so extremelly low right now. There used to be problems but that was before preparing against every possible scenario of string contents. Right now I would be really surprised if that causes problems. Now, if somebody finds a huge problem caused by the optimizer, I will have to update it and I would add that option back then. Notice that the basic obfuscating would be required to use the inline strings option though. |
