passing arguments to code/boolexpr

05-13-2008, 02:20 PM

I think I can do it in a way that isn't absurdly lame, anyways these are the ideas right now:

JASS:

function aragat takes unit u, integer v returns boolean
    call SetUnitUserData(u, GetUnitUserData(GetFilterUnit() ) + v)
    return false
endfunction

//...
    call GroupEnumUnitsInRange(g, x,y, 500.0 , condition_function aragat(GetTriggerUnit(), 5) )
//.

into:

JASS:

function c__ aragat takes nothing returns boolean
 local unit u= f__arg_unit1
 local integer v = f__arg_integer1
    call SetUnitUserData(u, GetUnitUserData(GetFilterUnit() ) + v)
 set f__arg_unit1=u
 set f__arg_integer1=v
 return false
endfunction

function c__GroupEnumUnitsInRange_1 takes group g, real x, real y, real radius, unit u, integer v returns nothing

   set f__arg_unit1=u
   set f__arg_integer1=v
   call GroupEnumUnitsInRange(g,x,y,radius,Condition(function c__aragat) )

   
endfunction


//...
    call c__GroupEnumUnitsInRange_1(g, x,y, 500.0 , GetTriggerUnit(), 5 )
//.

Anyways, the bad news:

JASS:

local boolexpr b= condition_function aragat(GetTriggerUnit(), 5) //Syntax error: illegal usage of condition_function

05-13-2008, 02:41 PM	#2
emjlr3	not bad, in the end not many extra steps, and takes the hassle out of needing to do the variable transfer yourself good going

05-13-2008, 02:57 PM	#3
Vexorian	It will use the same global variables all the time, I think a lot of systems use its own variables to do this stuff, I think the next challenge is to reduce the amount of global variables as much as possible.

05-13-2008, 05:51 PM	#4
ADOLF	all my long life transferred value through global variables... also i feel happily) to write a pair of instructions of assignment it is completely not difficult, and code from first post obviously is not optimum=/ // define waits when it will added)))

05-13-2008, 06:55 PM	#5
Vexorian	nobody will force you to use it. Speed is but a single metric of many to consider, the code I posted is the only one with correctness.

05-13-2008, 07:22 PM	#6
grim001	Guess it is useful for non-periodic groupenums that find units to affect when a spell is cast. It would allow you to have a library of useful filters.

05-13-2008, 09:58 PM	#7
Toadcop	it's called vex have idea for an pizdato useless feature. or short. 100500 wrappers will not make it better. (vJass) ofc you can do what you want but it's a publick thread :P // to hear everytime omg awesome is also boring ^^

05-14-2008, 01:19 AM	#8
emjlr3	who the hell cares if there is an extra wrapper? if it makes my life easier, any the minimal slowdown is never apparent in game, its well worth it

05-14-2008, 01:42 AM

Vexorian

Quote:

it's called vex have idea for an pizdato useless feature.

or short. 100500 wrappers will not make it better. (vJass)
ofc you can do what you want but it's a publick thread :P
// to hear everytime omg awesome is also boring ^^

The wrapper is necessary, it is not my fault you people are used to pass global variables in an irresponsible way that will break things up if there's the smallest nesting, that's another of the things I am intending to fix with this one.

Also, the wrapper is most likely to get inlined in case it is not necessary when it turns out I upgrade the inliner. And for crying out loud, this little function call is nothing in comparison to the amount of calls a ForGroup or a GroupEnum do already. And perhaps, the reduction in the globals count cancels the issue already.... This is as most of vJass is general purpose solution, thus half-assed solutions that are not robust enough are not what I want to accomplish with it, that's the reason behind all those wrappers.

05-14-2008, 01:48 AM	#10
TheSecretArts	I like it. It makes scripting easier for me.

05-14-2008, 01:58 AM

#11

Vexorian

Anyways, I made the whole post to ask about the syntax I am using, could somebody please give any feedback about it?

JASS:

call GroupEnumUnitsInRange(g, x,y, 500.0 , condition_function aragat(GetTriggerUnit(), 5) )

//Forgot about the version for ForGroup:
call ForGroup(g,  function aragat(GetTriggerUnit(), 5) )

should work in ForForce and Item/Destructable enums (Who the hell uses ForForce though?)

05-14-2008, 09:34 AM

#12

ADOLF

Quote:

100500 wrappers will not make it better

1000 useless commands - powerful argument (c) DioD

// still it is possible to ask what for to invent a bicycle...

imho the suggested method of transfer of argument will not be useful...)

Quote:

nobody will force you to use it

yes, but made 100500 wrappers too there is no sense...

05-14-2008, 11:19 AM	#13
Toadcop	the main problem in this is intializing local vars as for me. yes i know you need to keep value queue... btw you could use a global stack... but it's not much better. well yeah to make it "user friendly" i guess you need to add such feature. but i wouldn't use it :P i dislike code generation without my knowledge.

05-14-2008, 12:13 PM	#14
Captain Griffen	I'm not a great fan of the condition_function syntax, partly due to the length, and partly due to the '_' which looks ugly and requires that I use two keys. I'd prefer call GroupEnumUnitsInRange(g, x,y, 500.0 , condition aragat(GetTriggerUnit(), 5) ), but I can see that might cause confusion...

05-14-2008, 12:16 PM	#15
emjlr3	if it were up to me JASS: CF aragat() would be a fine, a little harder to understand, but I think easier to remember and easiest to write or perhaps JASS: CF.aragat() I am not good with syntax things