ClearTargets sometimes stops prematurely

[Lyricalpanda]

From report: "pressing cleartargets again , they continue"

Console prints out "TrueTarget has no spawn count"

[RekkasGit]

Note there are limits on what is allowed to target. Anything that isn't LoS will not be picked up. I did several hours of testings and it seemed to work fine as long as I didn't lose LoS.

foreach (var s in _spawns.Get().OrderBy(x => x.Distance))
{
    //find all mobs that are close
    if (!s.Targetable) continue;
    if (!s.Aggressive) continue;
    if (!MQ.Query<bool>($"${{Spawn[npc id {s.ID}].LineOfSight}}")) continue;
    if (s.Distance > 60) break;//mob is too far away, and since it is ordered, kick out.
    if (s.TypeDesc == "Corpse") continue;
    if (s.Name.Contains("'s pet'")) continue;
    if (s.Name.IndexOf("Chest", StringComparison.OrdinalIgnoreCase) > -1) continue;
    if (s.Name.IndexOf("a box", StringComparison.OrdinalIgnoreCase) > -1) continue;
    if (s.Name.IndexOf("crate", StringComparison.OrdinalIgnoreCase) > -1) continue;
    if (s.Name.IndexOf("hollow_tree", StringComparison.OrdinalIgnoreCase) > -1) continue;
    if (s.Name.IndexOf("wooden box", StringComparison.OrdinalIgnoreCase) > -1) continue;
    //its valid to attack!
    MobToAttack = s.ID;
    break;
}

[RekkasGit]

Please reopen if the issue is continuing or you have more information.

[Lyricalpanda]

This happened to me twice, added some debugging to try and narrow down a bit what's happening. Maybe MobToAttack isn't clearing for certain mobs? https://github.com/RekkasGit/E3Next/blob/master/E3Next/Processors/ClearXTargets.cs#L37

Added in some print lines, saw this interesting tidbit:

I had called assist on ID 3. Immediately after, I get two bots saying that's a corpse and my driver actually changed targets to ID 108 to cast a spell on. Then I see a bunch of spam that we're still waiting on ID 3 to die, turn into a corpse, and proceed to next XTarget.

It's looking like it's not stopping prematurely as much as a potential issue with MobToAttack and ClearTargets thinking the target is still alive.

So you can see where the extra writes are coming from:

                if (MobToAttack == 0)
                {
                    int counter = 0;
                    foreach (var s in _spawns.Get().OrderBy(x => x.Distance))
                    {
                        counter += 1;
                        //find all mobs that are close
                        if (!s.Targetable) continue;
                        if (!s.Aggressive) continue;
                        if (!MQ.Query<bool>($"${{Spawn[npc id {s.ID}].LineOfSight}}")) continue;
                        if (s.Distance > 60)  {
                            MQ.Write($"ClearTargets kicking out because distance {s.Distance}");
                            break;//mob is too far away, and since it is ordered, kick out.
                        }
                        if (s.TypeDesc == "Corpse") continue;
                        if (s.Name.Contains("'s pet'")) continue;
                        if (s.Name.IndexOf("Chest", StringComparison.OrdinalIgnoreCase) > -1) continue;
                        if (s.Name.IndexOf("a box", StringComparison.OrdinalIgnoreCase) > -1) continue;
                        if (s.Name.IndexOf("crate", StringComparison.OrdinalIgnoreCase) > -1) continue;
                        if (s.Name.IndexOf("hollow_tree", StringComparison.OrdinalIgnoreCase) > -1) continue;
                        if (s.Name.IndexOf("wooden box", StringComparison.OrdinalIgnoreCase) > -1) continue;
                        //its valid to attack!
                        MobToAttack = s.ID;
                        MQ.Write($"ClearTargets found Mob ID {s.ID} name {s.CleanName}");
                        break;
                    }
                    MQ.Write($"ClearTargets looked through {counter} mobs");
                    if (MobToAttack == 0)
                    {
                        //we are done, stop killing
                        Enabled = false;
                        MQ.Write("\agClear Targets complete.");
                        return;
                    }

                    //mobs to attack will be sorted by distance.
                    if (MobToAttack > 0)
                    {
                        //pop it off and start assisting.
                        Int32 mobId = MobToAttack;
                        Spawn s;
                        if (_spawns.TryByID(mobId, out s))
                        {
                            MQ.Write("\agClear Targets: \aoIssuing Assist.");
                            Assist.AllowControl = true;
                            Assist.AssistOn(mobId);
                            MQ.Delay(500);
                            E3.Bots.BroadcastCommandToGroup($"/assistme {mobId}");
                        }
                    }
                } else
                {
                    MQ.Write($"ClearTargets Mob to attack {MobToAttack}");
                }

[Lyricalpanda]

Also, I actually can't reopen issues 😆. But will ping once I have some more information.

[Lyricalpanda]

Not sure why, but just happened again.

We're trying to attack ID 31, which is reported to be a corpse right after assist call. Driver is now casting on a new mob. After combat, I checked for target id 31, which resulted in nothing.

I had no looters on, so the corpse should still be there.

Looks like a missing else for when !(_spawns.TryByID(MobToAttack, out ts) { MobToAttack = 0} ? I'll try the following locally

                if (MobToAttack > 0)
                {
                    Spawn ts;
                    if (_spawns.TryByID(MobToAttack, out ts))
                    {
                        //is it still alive?
                        if (ts.TypeDesc == "Corpse")
                        {
                            //its dead jim
                            MobToAttack = 0;

                        }
                    } else {
                        //Can't find mob, look for new one
                        MQ.Write($"ClearTargets Can't find spawn ID {MobToAttack}, clearing.");
                        MobToAttack = 0;
                    }

                }

[RekkasGit]

I will be adding a delay before we refresh the spawns, this should make sure that we have the most up to date spawn data. Reopening this so we can test once i push it into master.

[RekkasGit]

[RekkasGit]

no update, closing for now.

[Lyricalpanda]

Made sure I pulled in the latest copy, just ran into it three times on the same fight in guk ldon (particularly unlucky). Should we have a safe guard of MobToAttack=0 if _spawns.TryByID fails? To cover if a corpse doesn't spawn or if the corpse is looted before the next system tick on the driver.

Funnily enough, FUKU is also a good test for this (someone tried this on our attempt 😆) as the adds don't leave behind corpses