[Lo6idZe]

Thanks... Other opinions ?

[Arkshine]

To do a countdown, set_task is perfectly fine. Though, the way how you do the thing is badly coded.

[Lo6idZe]

Mmm... Can you edit it , please ?

[Arkshine]

That's not the right forum to ask requests.

[Bad_Bud]

I just got around to using an entity instead of a "b" flag set_task for the first time...

Considering the reason behind switching to an entity is for optimization, I find the following code segment somewhat troubling in concept:

PHP Code:

#include <amxmodx>
#include <amxmisc>
#include <fakemeta>

new g_Classname[] = "roflmao"

public plugin_init()
{
    register_plugin("ROFL","1.0","Hawk552")
    
    new Ent = engfunc(EngFunc_CreateNamedEntity,engfunc(EngFunc_AllocString,"info_target"))
    set_pev(Ent,pev_classname,g_Classname)
    set_pev(Ent,pev_nextthink,20.0)
    
    register_forward(FM_Think,"ForwardThink")
}

public ForwardThink(Ent)
{
    static Classname[33]
    pev(Ent,pev_classname,Classname,32)
    
    if(!equal(Classname,g_Classname))
        return FMRES_IGNORED
    
    log_amx("ROFLMAO thinks.")
    
    set_pev(Ent,pev_nextthink,20.0+halflife_time())
    
  return FMRES_IGNORED
} 


We check a string on every call.

I chose to do it a different way, and I haven't had problems with it yet, but I'd like to know if there's something wrong with the way I'm doing it, because for replacing a "b" flag set_task, this seems a lot more logical.

PHP Code:

#include <amxmodx>
#include <amxmisc>
#include <fakemeta>

new g_thinker

public plugin_init()
{
    register_plugin("255kwaH","0.1","LFOR")
    
    g_thinker=engfunc(EngFunc_CreateNamedEntity,engfunc(EngFunc_AllocString,"info_target"))
    set_pev(g_thinker,pev_classname,"doesntmatter")
    set_pev(g_thinker,pev_nextthink,20.0)
    
    register_forward(FM_Think,"ForwardThink")
}

public ForwardThink(Ent)
{
    if(Ent!=g_thinker)
        return FMRES_IGNORED
    
    log_amx("ROFLMAO thinks.")
    
    set_pev(Ent,pev_nextthink,20.0+halflife_time())
    
  return FMRES_IGNORED
} 


But still, I don't think I've ever understood why using an entity would be more efficient than a set_task. Does anyone care to explain?

[Alka]

Because setting a task with flag "b" aka LOOP, have some random delays, like if you set a task with 1.0 delay will have delays like -> 0.8, 0.9, 1.1,...a fix for this will be to not set flag "b" but to call task everytime in that public. Also set_task is a native, making a ent and make it think is faster that native.

[Bad_Bud]

Oh, that's cool. Thanks.

[ConnorMcLeod]

Use register_think and register your custom entity classname, so only that ent will call the handler.

[Arkshine]

@Hawk552 :

Sorry to bump an old thread but connor was saying me he has a big lag when a vote appears from the mapchooser ( or galileo or timeleft plugin, etc) plugin and set_task might be the cause. You say :

What few people know is that it is a very expensive operation, the CPU has to literally check every cycle if it is time for the set_task to be executed.

Checking the Amxx source code, you have a check done in StartFrame() with g_task_time to make sure g_tasksMngr.startFrame(); is called each 0.1 second, not every cycle. Next, CTaskMngr::startFrame() will still loop anyway all the time through all the registered tasks, then executing executeIfRequired() to check if a task can be executed.

My point is : I don't think few more checks (such simple checks and of course multiplied by the number of registered tasks) will add a significant change in the cpu consumption when you have already much forwards in Amxx or others modules called very often with many checks. I would assume that using set_task is perfectly fine even when using the "b" flag and you would use an entity only when you need to control a precise and <0.1 think time.

About the connor's problem, having a lag (it appears not always) when a set_task is executed, I don't know enough the C++ to say if the problem is really from there (though I don't see anywhere what it can be the cause), and that's not like he has searched throughly, so I can't give much informations.

Anyway, if you don't mind, I would like to hear your thoughts.

[Bugsy]

I saw a big decrease in transfer time (file sent much faster) with my FTP include when I changed from set_task w\ flag "b" to thinking entity (using 0.1 interval for both) . I will try to dig up the older copy and re-test to make sure something else was not causing this.