[grs4]

Maybe read_file which i am using? From yesterday, no errors

[Arkshine]

I can't do much unless you're giving me a reproducible way of the issue or some hints where I can look into.

[grs4]

Originally Posted by Arkshine I can't do much unless you're giving me a reproducible way of the issue or some hints where I can look into.

Yesterday I read this: https://forums.alliedmods.net/showth...=69386?t=69386

I had changing player models. Also i changed all files from .ini to .cfg. and as i said, from yesterday no errors.

If it will appear again, i will PM to you, because I want to repair this.

[Arkshine]

I've no idea what you mean why changing .ini to .cfg, but if related to plugins spamming players, can't do much about it.
If you are using amxx dev version, you should not have issues with setting player's model with cs_set_user_model. Your issue is probably something else.

Anyway, let's try to be logic if your issue is related to player's connection, this has nothing to do with player's model.
You should take a look at plugins which do stuff at this time (plugin_precache, client_connect, etc.).
Ideally, you should try to reproduce the issue, then disable the plugins one by one which does stuff at player's connection, and if related to AMXX, we should find the issue easily.

[grs4]

Originally Posted by Arkshine I've no idea what you mean why changing .ini to .cfg, but if related to plugins spamming players, can't do much about it. If you are using amxx dev version, you should not have issues with setting player's model with cs_set_user_model. Your issue is probably something else. Anyway, let's try to be logic if your issue is related to player's connection, this has nothing to do with player's model. You should take a look at plugins which do stuff at this time (plugin_precache, client_connect, etc.). Ideally, you should try to reproduce the issue, then disable the plugins one by one which does stuff at player's connection, and if related to AMXX, we should find the issue easily.

If there is a problem with .ini, then chaning to .cfg should fix it.
About cs_set_user_model, there is some bug. If player has a new model set by cs_set_user_model and he is going to spec or discoonecting from server, plugins are registering a touch from this player like he is ingame.

We had talk about Arrays, this function is crashing server. Sometimes this function is working great, sometimes bad.

What is saying?

PHP Code:

if(DEBUG)    
        log_to_file(LOG_FILE, "DEBUG: Getting map index with most votes. Size: %d / %d", 


DEBUG: Getting map index with most votes. Size: 2 / 3
Segmentation fault data..
uploading dump..

What does this function do? It's calculating votes for next map. If there is a draw, &ret1 is first map id and &ret2 is second map id. If map won, &ret2 = 0.
This function is 100% causing server to crash. (Segmentation fault data)

PHP Code:

public GetMapIndexWithMostVotes(&ret1, &ret2)
{
    if(DEBUG)    
        log_to_file(LOG_FILE, "DEBUG: Getting map index with most votes. Size: %d / %d", g_VotedForNextMapNum, VOTE_MAP_COUNT)
        
    new Array:aVotes = ArrayCreate(g_VotedForNextMapNum, VOTE_MAP_COUNT)

    for(new i = 0 ; i < g_VotedForNextMapNum; i++)
         ArrayPushCell(aVotes, GetPlayersNumInVoteZone(i+1));

    new tempvotes, mostVotes, secondMostVotes = -1, firstMapId, secondMapId;

    /* Get map with most votes */
    for (new i = 0; i < g_VotedForNextMapNum; i++)
    {
        tempvotes = ArrayGetCell(aVotes, i);
        if(tempvotes > mostVotes)
        {
            mostVotes = tempvotes;
            firstMapId = i;
        }
    }

    /* Get second map with most votes */
    for(new i = 0 ; i < g_VotedForNextMapNum; i++)
    {
        if(i == firstMapId)
            continue;
        
        tempvotes = ArrayGetCell(aVotes, i);
        
        if(tempvotes > secondMostVotes)
        {
            secondMostVotes = tempvotes;
            secondMapId = i;
        }
    }

    // if we have a draw
    if(mostVotes == secondMostVotes)
    {
        ret1 = firstMapId;
        ret2 = secondMapId
    }
    else
    {
        ret1 = firstMapId;
        ret2 = -1;
    }
    
    ArrayDestroy(aVotes);
} 


[Arkshine]

If there is a problem with .ini, then chaning to .cfg should fix it.

This makes no sense. What are you even talking about? .ini/.cfg of what? Please be precise.

About cs_set_user_model, there is some bug. If player has a new model set by cs_set_user_model and he is going to spec or discoonecting from server, plugins are registering a touch from this player like he is ingame.

This makes no sense too. You will have to elaborate a bit more. Really difficult to understand you.

We had talk about Arrays, this function is crashing server. Sometimes this function is working great, sometimes bad.

How can you be sure the crash comes from this function while a segmentation is not going to tell exactly where it crashed? The code looks fine at first glance. If it's reproducible, can you make a simple test plugin with this code?

I think you should try to ask someone who can speak in english, because it's going to be hard to understand you fully.

[grs4]

Originally Posted by Arkshine This makes no sense. What are you even talking about? .ini/.cfg of what? Please be precise. This makes no sense too. You will have to elaborate a bit more. Really difficult to understand you. How can you be sure the crash comes from this function while a segmentation is not going to tell exactly where it crashed? The code looks fine at first glance. If it's reproducible, can you make a simple test plugin with this code? I think you should try to ask someone who can speak in english, because it's going to be hard to understand you fully.

This makes no sense. What are you even talking about? .ini/.cfg of what? Please be precise.

As it was told before, i'm talking about loading/saving from .ini files.
You told that You don't have idea, why I changed .ini to .cfg.

Take a look at this. Depresie said he had same problem and he thought the problem is in loading/saving to .ini files. So i'm trying all paths to fix it. Not only you are trying to help me.

There is sure something wrong related either to reading from the .ini file or to the dynamic arrays

I dont want to argue with you and also be patient plase, no one can know english perfectly. (:

About cs_set_user_model, there is some bug. If player has a new model set by cs_set_user_model and he is going to spec or discoonecting from server, plugins are registering a touch from this player like he is ingame.

I thought when player disconnected from server and before plugin set him model by cs_set_user_model, plugins thought that player still is on server. Why i thought that? I had entity which had register_touch on a player. When player left, plugin thought he is still on server. but i made an mistake, i forgot about !is_user_connected.

How can you be sure the crash comes from this function while a segmentation is not going to tell exactly where it crashed? The code looks fine at first glance. If it's reproducible, can you make a simple test plugin with this code?

PHP Code:

    if(g_VoteTime == 0)
    {
        if(DEBUG)    
            log_to_file(LOG_FILE, "DEBUG: Voting ended.")
        
        new result[3]
        GetMapIndexWithMostVotes(result[0], result[1]);
                
        set_hudmessage(0, 255, 255, 0.06, 0.70, 1, 1.0, 4.5, 0.2, 0.2, -1)
        ShowSyncHudMsg(0, g_SyncHudInfo, "Calculating results...");
            
        RemoveVoteMapEntitiesAndResetUsers();
        set_task(5.0, "CheckVoteResults", _, result, 2);
    } 


Because I saw "DEBUG: Voting ended." In console, and only as i said before, GetMapIndexWithMostVotes had been called. I did not saw a hud with "Calculating results.." and no ents were removed from RemoveVoteMapEntitiesAndResetUsers();. Only this `if` was executed.

If it's reproducible, can you make a simple test plugin with this code?

Yesterday I removed Dynamic Arrays to check if this was a reason.
[check attachment]
Yesterday 18.55 UTC +1 server was crashed, so then i changed it.
Is hard to create at now reproducible code, because as I said, sometimes It was working, sometimes no and i still don't know why. If crash will appear again, I will be know that dynamic arrays were not a reason, if no, then i will roll back to dynamic arrays to try find a solution.

Sorry for my english, i'm still learning on my bads, i don't want anyone to help me.

[Arkshine]

"ArrayCreate(g_VotedForNextMapNum, VOTE_MAP_COUNT)"

The first argument is wrong (it expects the size in cell). It should be 1 because you save only integers (you usually set > 1 when you're working with array or string to provide the max length).
It might be the cause of the potential crash.

Despite all the checks, it's very easy to crash AMXX when the API is misused.

[Depresie]

One question, you are running the server on linux or windows ?

[grs4]

I changed to size 1 and no crash at this time.

Server is linux, plugins are compiled on windows.