[AtomicStryker]

Originally Posted by olj EDIT: You use TIMER_FLAG_NO_MAPCHANGE, why you trying to kill them on map end??

An attempt to make Sourcemod recognize the Timers as INVALID_HANDLE after a mapchange.


@D1maxa
Try it in a game. Youll get the same errors.


The one thing i could do is remove the OnMapEnd() function entirely, which should work since the Timers have the nomapchange flag..

[D1maxa]

i have no errors anymore with my edited version ;)
Only correct work in programming with handles - and u don't have any problems. The problem wasn't in sourcemod

[AtomicStryker]

What you posted is how version 1.0.0 looked. You'll get handle errors, trust me

[D1maxa]

Originally Posted by AtomicStryker What you posted is how version 1.0.0 looked. You'll get handle errors, trust me

give me your version 1.0.0
with your version 1.0.2 server got errors every 10 minutes, 500 kb error log for one day

[AtomicStryker]

I dont have version 1.0.0 anymore but be assured it used KillTimer just as your recode does. It would be a lot easier if ... does setting a Handle invalid kill its timer already?

[D1maxa]

Originally Posted by AtomicStryker I dont have version 1.0.0 anymore but be assured it used KillTimer just as your recode does. It would be a lot easier if ... does setting a Handle invalid kill its timer already?

no, you have nothing in sourcemod. There is task_exists(id) in amx mod x, but not in sourcemod. So u must set INVALID_HANDLE to your timer's handle when u kill timer or timer function executed

[AtomicStryker]

There, i scrapped using KillTimer or CloseHandle altogether. I just ran an entire Campaign with lots of Ragequitting (about 40) and not a single Sourcemod error. The Plugin worked fine.

Have 1.0.4

[D1maxa]

there are no errors of course because u removed all of CloseHandle and KillTimer. But u don't control when your timer's function executes.
Look on next example:
l4d_kickloadstuckers_duration is 45, i connected in 40 sec before map ends, map is ending, server is changing map and CheckClientIngame is executing: i'm loading new map so i'm not in game and server is kicking me. Is this normal?
If u want save your code, u must use next check:

public Action:CheckClientIngame(Handle:timer, any:client)
{
    if (!IsClientConnected(client)) return; //onclientdisconnect should handle this, but you never know
    if(LoadingTimer[client] == INVALID_HANDLE) return;
    if (!IsClientInGame(client))
        {
            decl String:name[256];
            GetClientName(client, name, sizeof(name));
            PrintToChatAll("%s was kicked for being stuck in connecting state for %i seconds", name, RoundToNearest(GetConVarFloat(CvarDuration)));
            
            KickClient(client, "You were stuck Connecting for too long"); //he is connected but not ingame after an entire minute? SMITE HIM!!!
        }
}

my previous code with KillTimers workes without any problem on real public game server during whole day (there are hundreds of connects/disconnects on it), so it's not like your version 1.0.0, you were wrong.
And final: u don't need OnMapEnd function because

Are clients disconnected on mapchange? All clients are fully disconnected before the map changes. They are all reconnected after the next map starts.

so OnClientDisconnect(client) does all job when map ends

[AtomicStryker]

That is why i make use of OnMapEnd(). Why don't you playtest it before saying it doesnt work

It invalidates all running Timers, including the one you claim would kick you.

My approach works and doesn't generate errors anymore, im happy with it.

[D1maxa]

Omg! don't u see that OnMapEnd and OnClientDisconnect does same work when map ends? If there are no errors, this doesn't mean what code is correct. Your code has logical errors. I have said about possible situation already.

LoadingTimer[client] = INVALID_HANDLE <- u think that this kills timer? do u really think that?

I can not understand how your plugin approved. Perhaps version 1.0.0 was more clear