Patch 1.23b, JASS, GUI, and Gaming

[Eleandor]

I assume Griffen meant that dead code should not be allowed. conditional code (if/then/else) should ofcourse still be allowed. Anyway, a) won't solve the problem completely, since you would still be able to return wrong types through something such as "if true then return h else return 0; return 0". Besides, most compilers I've used still allow dead code, they only give warnings but still compile it. In the end, dead code shouldn't cause any trouble at all, it's no typechecking that causes the trouble here.

 

[Captain Griffen]

I assume Griffen meant that dead code should not be allowed. conditional code (if/then/else) should ofcourse still be allowed. Anyway, a) won't solve the problem completely, since you would still be able to return wrong types through something such as "if true then return h else return 0; return 0". Besides, most compilers I've used still allow dead code, they only give warnings but still compile it. In the end, dead code shouldn't cause any trouble at all, it's no typechecking that causes the trouble here.

Dead code's a mistake. b would be necessary and sufficient really to fix the return bug. a would just be extra.

 

[Dark_Dragon]

wow i just had hard time with my librarys making coz some natives dont work...

local hashtable ht = InitHashtable()
call SaveInteger(ht,0,0,5)
call SaveReal(ht,0,1,bj_PI)
call RemoveSavedInteger(ht,0,0)
call RemoveSavedReal(ht,0,1)
call BMsg(HaveSavedInteger(ht,0,0)) // result: true
call BMsg(HaveSavedInteger(ht,0,1)) // result: true

for handle it works did not test for boolean and string...
lols

~DD

 

[Captain Griffen]

Are you on 1.24? If so, report it.

 

[Dark_Dragon]

i am on 1.24 but i had problems with my account ... if u want u can do it (i mean only if u want) and btw i am not sure where to post it (ofc to blizzard but maybe here Report)
...

nvm i was an idiot... found it but as i said have some problems it tells me that it cant detect my account name or password!

so again if u want feel free to do so!

~DD

EDIT: lol i was not on battle.net for so long it seems my account got deleted or wtf?
well well its always like that... i dont even remember when i was last time on battle.net lol played dota and owned some freaks then they kick me and i was so pissed so i did not play from that time :S

 

[maskedpoptart]

I have dowloaded 1.24. I did tests with all the primitive types and it seems none of them get removed (you can still load the values after they are removed). This results in HaveSaved<type> to screw up and always return true after a value has been stored in that slot for the first time. I will report this to blizzard.

 

[Dr Super Good]

maskedpoptart, atleast that is an improvement over before, they used to always return false even if there was a saved value.

 

[maskedpoptart]

Ah, that's good. (I didn't get a chance to test hashtables till now.)

Can anyone tell me why the only hashtable functions that show up for me in the JNGP function list are SaveX, LoadX, and InitHashtable? RemoveSavedX and HaveSavedX aren't showing up for some reason... they all compile without error...

 

[Dark_Dragon]

Ah, that's good. (I didn't get a chance to test hashtables till now.)

Can anyone tell me why the only hashtable functions that show up for me in the JNGP function list are SaveX, LoadX, and InitHashtable? RemoveSavedX and HaveSavedX aren't showing up for some reason... they all compile without error...

i dont know what u mean but they show for me...

and btw i cant explain the bug but sometimes SaveInteger does not overwrite the value i mean if i save 0 over 6 for example it will not save the 0 but 6 will remain... crazy bugs!

still cant find why that happens (its in special cases) and i get an access of memory violation error randomly in my map which means that HaveSavedX is kind bugged... in some cases... still cant find them its used in mass systems and algorithms are complicated so yeah...

~DD

 

[DoOs_101]

Ok so basically they turned every custom hash table systems made into natives?

And if 1.24 screws up the return bug then is there a way to resolve this because in all 4 of my maps including 1 I'm making, they all use CSCache, XE, and other systems that use that return bug.

 

[Eleandor]

Replace those systems with the 1.24 version, which will undoubtedly be out before the patch will be available...

 

[DoOs_101]

What damned employees would intentionally or unintentionally screw the wc3 community. I mean seriously they should know things like this if they're updating.

Controversially, it could be that they don't give a damn about this game cuz they already got a new extravaganza at work, sc2...

 

[Dark_Dragon]

What damned employees would intentionally or unintentionally screw the wc3 community. I mean seriously they should know things like this if they're updating.

Controversially, it could be that they don't give a damn about this game cuz they already got a new extravaganza at work, sc2...

this update patch (1.24) is out to revive wc! you just dont know it but this is really awesomeness what blizzard is currently doing...

 

[Yixx]

I dont know if you have paid any attention to this, DoOs_101, but Blizzard has had close contact with Vexorian and others (i have heard),
And all you have to do is make sure the H2I function is private or public, and replace

return h
return 0

with return GetHandleId(h)

Only when you use any I2H (like I2Trigger or I2Timer) you got a problem, but there are other solutions to this problem that are as well usable as I2H,,

 

[Dr Super Good]

Or for even more efficency just use a search and replace of all references of H2I( with GetHandleId(. This is not only easier, neater but also makes the resulting code faster as it takes advantage of the native's new speed instead of wrapping it for some reason.

Why did vexorian not think of this is beyond me but it should be far more effective than simply just changing it to a wrapper function.

 

[Captain Griffen]

Or for even more efficency just use a search and replace of all references of H2I( with GetHandleId(. This is not only easier, neater but also makes the resulting code faster as it takes advantage of the native's new speed instead of wrapping it for some reason.

Why did vexorian not think of this is beyond me but it should be far more effective than simply just changing it to a wrapper function.

...except that Vexorian's JassHelper will inline it? Meaning that your efficiency argument goes out of the window.

 

[Dr Super Good]

Are you sure it copies the contense of the function to where its called? Or do you mean something else with your crazy inline.

Anyway, searching and replacing them can also work outside of JNGP so theres your argument out of the window as well.

 

[Captain Griffen]

...you don't know what inline is? I really don't get your question. 'Contense' is not a word...

Working outside of JNGP is working outside of sanity, so screw those people?

 

[DoOs_101]

I dont know if you have paid any attention to this, DoOs_101, but Blizzard has had close contact with Vexorian and others (i have heard),
And all you have to do is make sure the H2I function is private or public, and replace

return h
return 0

with return GetHandleId(h)

Only when you use any I2H (like I2Trigger or I2Timer) you got a problem, but there are other solutions to this problem that are as well usable as I2H,,

Alright thats good!

I don't know if I use I2H and other Iinteger 2 something, but does CSCache use all of them? If so then I'm damned and need of an update on that system @@.

 

[Yixx]

DSG: Yes it is better to search for the call H2I( and replace it,
but lazy people can just change the content of the H2I function as long as it is public or private, since Blizzard is gonna make their own H2I native (as you probably know,,).

 

[Captain Griffen]

The blizzard making a H2I native is a decent reason to change the calls to H2I to GetHandleId, though, since we don't know what behaviour the native will have.

 

[DoOs_101]

The blizzard making a H2I native is a decent reason to change the calls to H2I to GetHandleId, though, since we don't know what behaviour the native will have.

R they going to fix the I2H as well?

 

[PurplePoot]

Naturally.

 

[scorpion182]

when the 1.24 officially come out?

 

[Captain Griffen]

R they going to fix the I2H as well?

If by fix you mean remove, yes. If by fix you mean make a native equivilent, then damned well no.

 

[krisserz]

Ehh Blizzard is so patching wc3 and we last times...

 

[xp3r009]

when the 1.24 officially come out?

Not so sure, the only way to get it currently is through Westfall server on Battle.net as a BETA patch, but if you would like to play DotA and some popular scenario's I wouldn't recommend patching it yet.

For now, we would have to wait until its real release.

 

[Dr Super Good]

The guy on the blizzard forum for diablo II said they neary had completed 1.24.

They are probably holding it off release until DotA is updated to work in which case the difference of release should be a few hours atmost. Although this sounds crule, if they did not do this WC3 would die over night.

 

[Blue_Jeans]

offtopic: I think Dr Super Good explained the broken primitives at least 10 times...

I just thought I'd mention that amusing tidbit.

 

[Captain Griffen]

They are probably holding off until the return bug is fixed.

 

[Vosty]

Ummm. How come I cant find these in Blizzard.j/common.j?
I got v1.23b....

 

[Dr Super Good]

1.24 is the lattest beta, use that and ask.

 

[maskedpoptart]

Okay I feel like a noob asking this, but I've been wondering it for a while now:

Where do I find the Blizzard.j/common.j files? How do I open them? I have JassCraft and JNGP.

 

[Dr Super Good]

In the patch mpqs in your WC3 install directory. I think they are in a script or code subfolder.

 

[PurplePoot]

War3Patch.mpq\scripts\common.j

(And the same for Blizzard.j)

 

[maskedpoptart]

Yep, I found it. Out of curiosity, what do you guys use to open mpqs? I ended up using War3ModelEditor's MPQ Browser, and then exporting the .j files.

 

[Anachron]

They are probably holding off until the return bug is fixed.

You mean, finally fixed, after it has been again detected with a small code change?

 

[PurplePoot]

I use Ladix's MPQ Editor or MPQMaster, and W3ME's MPQ reader when I want to browse models. The problem with W3ME's is that it blows up when the MPQ lacks a listfile.

 

[Vosty]

Ok, I just have to go to westfall beta, right?
EDIT: Why do all the save functions return booleans? Is it checking if it already there, or putting it there and telling you if it did or something?
And should I just replace this in local handle vars, or just drop local handle vars completely?

 

[M_Vegeta]

its a good news for who know how fix it ............ most of my trigger are imported i dont know alot of jass, i read lot post in topics and since im french i failed at understanding

.... Hope some1 could help me fix it i use newjassGen i did change j.file and since i saw gui could use returns bug im more screwed............ lol

help would be nice and imba credits would be gave :O lot chance for poeple to get credits ^^, and im sure im the the only 1 that probly have this bug and dont know how to fix it or what it is

edit : i dont know if its normal i cannot index an special effect variable

is there anyplace i can learn how to use new hashtable ?

 

[Vosty]

There is a new tut for GUI hashtables.
Learn it.