E8S Mirror Mirror 3 sometimes does wrong directions

[Stealth3si]

I was asked to submit this here instead by quisquous

[valarnin]

I'm going to quote wexxlee here (maybe we should have an issue template that asks for this info?)

1. With ACT and FFXIV running, in ACT: go to Plugins -> OverlayPlugin.dll -> General and hit "Copy Tech Support Info to Clipboard", then reply and paste it here. It will likely include your windows username in the various file paths (assuming you're on Windows) - e.g. "C:\Users\AppData..." - so feel free to replace that with something like "xx".

2. Attach a log of the fight where this occurred. Use the Log Splitter, drop your log file where indicated, check "Anonymize Log", check the box for a fight where this occurred, hit "Export", and attach it to your reply.

(I vaguely recall something like this happening while running this for mount farm in December but it didn't really matter because we were doing it unsynced and just steamrolled it even with the damage downs, so I didn't bother reporting it)

[wexxlee]

I'm going to quote wexxlee here (maybe we should have an issue template that asks for this info?)

(Yeah, I was thinking something similar. There's also some nice stuff with GitHub actions - auto-tagging, closing stale issues, etc. -that I think would be worth looking at too. Adding to my list. 😄)

(I vaguely recall something like this happening while running this for mount farm in December but it didn't really matter because we were doing it unsynced and just steamrolled it even with the damage downs, so I didn't bother reporting it)

I vaguely remember seeing something similar under the same conditions (and not reporting it either). There's some comments in e8s.ts about some MapEffect locs that may or may not be right, so that may be one potential place to start.

A log that corresponds to the screenshots, as well as either the actual mirror placement during those screenshots (or, at a minimum, the "safe quadrant" that cactbot should have called) would be extremely helpful.

[JLGarber]

Is it possible there's a timing issue with the getCombatants call that populates her facing here? There shouldn't be, but if I'm reading the screenshots correctly, a mistake with her facing would cause the issues we're seeing here.

(I sort of wonder whether that getCombatants call is even necessary now, since we have better log line support for facing and positioning directly.)

[Stealth3si]

I'm going to quote wexxlee here (maybe we should have an issue template that asks for this info?)

1. With ACT and FFXIV running, in ACT: go to Plugins -> OverlayPlugin.dll -> General and hit "Copy Tech Support Info to Clipboard", then reply and paste it here. It will likely include your windows username in the various file paths (assuming you're on Windows) - e.g. "C:\Users\AppData..." - so feel free to replace that with something like "xx".
2. Attach a log of the fight where this occurred. Use the Log Splitter, drop your log file where indicated, check "Anonymize Log", check the box for a fight where this occurred, hit "Export", and attach it to your reply.

(I vaguely recall something like this happening while running this for mount farm in December but it didn't really matter because we were doing it unsynced and just steamrolled it even with the damage downs, so I didn't bother reporting it)

1. Here is the "Tech Support Info" you requested.

   
   Plugin Name             Enabled Version  Path
   ------------------------------------------------------------------------------------------------------------------------------------------------
   FFXIV_ACT_Plugin.dll    True    2.6.9.9  C:\Users\xxx\AppData\Roaming\Advanced Combat Tracker\Plugins\FFXIV_ACT_Plugin.dll
   ACT_DiscordTriggers.dll True    1.0.2.1  C:\Users\xxx\AppData\Roaming\Advanced Combat Tracker\Plugins\ACT_DiscordTriggers.dll
   OverlayPlugin.dll       True    0.19.26  C:\Users\xxx\AppData\Roaming\Advanced Combat Tracker\Plugins\OverlayPlugin\OverlayPlugin.dll
   CactbotOverlay.dll      True    0.31.2.0 C:\Users\xxx\AppData\Roaming\Advanced Combat Tracker\Plugins\cactbot-0.24.18\cactbot\CactbotOverlay.dll
   TriggerTree.cs          True             C:\Users\xxx\AppData\Roaming\Advanced Combat Tracker\Plugins\TriggerTree.cs
   SlimDX_XAudio2.dll      True    1.1.0.2  C:\Users\xxx\AppData\Roaming\Advanced Combat Tracker\Plugins\SlimDX_XAudio2.dll

Overlay Name URL

Triggers file:///C:/Users/xxx/AppData/Roaming/Advanced%20Combat%20Tracker/Plugins/cactbot-0.24.18/cactbot/ui/raidboss/raidboss.html?timeline=0&alerts=1 Ember https://goldenchrysus.github.io/ffxiv/ember-overlay/#/ Kagerou https://hibiyasleep.github.io/kagerou/overlay/

Various Settings Value

Plugin Language English Machina Region Global Game Version 2024.02.05.0000.0000 Inject and use Deucalion for network data False Hide Chat Log (for privacy) False Use WinPCap-compatible library for network data False Disable high-performance network filter False Disable Combine Pets with Owner False Disable Damage Shield estimates False (DEBUG) Enable Debug Options False (DEBUG) Log all Network Packets False (DEBUG) Also Show 'Real' DoT Ticks False (DEBUG) Simulate Individual DoT Crits False (DEBUG) Graph Potency, not Damage False (DEBUG) Enable Benchmark Tab False Cactbot User Dir C:\Users\xxx\AppData\Roaming\Advanced Combat Tracker\Plugins\cactbot-0.24.18\cactbot\user\

2. I attached the log file of where this fight occurred. 
[Eden's Verse_ Refulgence (Savage)_20240222_085742.322_7m_wipe.log](https://github.com/OverlayPlugin/cactbot/files/14391260/Eden.s.Verse_.Refulgence.Savage._20240222_085742.322_7m_wipe.log)
3. Below is the full edited screenshot that corresponds to this log file. The actual mirror placement during this screenshot is Blue Mirror is West, Green Mirror is North, Red Mirror is East. The "safe quadrant" that cactbot should have called is "NW => NE => SE (fast)"
![image](https://github.com/OverlayPlugin/cactbot/assets/76997301/9c521e90-1f19-4372-8bdb-df61268984fd)

[xiashtra]

Inject and use Deucalion for network data       False

Intermittent problem could easily be dropped network packets. You also have old install paths that could interfere with cleanly updating.

[Stealth3si]

Inject and use Deucalion for network data       False

Intermittent problem could easily be dropped network packets. You also have old install paths that could interfere with cleanly updating.

Roger that, I just enabled "Inject and use Deucalion for network data" and changed the install path to "C:\Users\xxx\AppData\Roaming\Advanced Combat Tracker\Plugins\cactbot\user"

I will report back the results.

[Stealth3si]

It seemed to have fixed the issue fn. Thanks!

If it happens again next week I will repost here.

[wexxlee]

I took a look at the log and triggers, and I think @JLGarber is right - it looks like a facing issue. The alerts in the screenshot would have been correct if Shiva was facing south, but she wasn't. The trigger treats any abs(heading) between 3.0 and 3.14 as north, and anything else as south, so if the combatant data polling picked up Shiva's (non-north) heading before it was updated, that would explain it.

(Incidentally, running the log through raidemulator shows combatantData at the time of trigger execution with a heading of -2.22, but I don't recall how we simulate that from a log read -- I assume it's based on the most recent prior log line with combatant position data?)

It seems plausible that fixing the install paths and using deucalion could impact how quickly combatant data is updated, but since valarnin and I also remember something like this happening, it may also just be an intermittent issue (which, iirc, is not unheard of for combatantData). Either way, this is probably still something that could be fixed.

[xiashtra]

I took a look at the log and triggers, and I think @JLGarber is right - it looks like a facing issue. The alerts in the screenshot would have been correct if Shiva was facing south, but she wasn't. The trigger treats any abs(heading) between 3.0 and 3.14 as north, and anything else as south, so if the combatant data polling picked up Shiva's (non-north) heading before it was updated, that would explain it.

(Incidentally, running the log through raidemulator shows combatantData at the time of trigger execution with a heading of -2.22, but I don't recall how we simulate that from a log read -- I assume it's based on the most recent prior log line with combatant position data?)

It seems plausible that fixing the install paths and using deucalion could impact how quickly combatant data is updated, but since valarnin and I also remember something like this happening, it may also just be an intermittent issue (which, iirc, is not unheard of for combatantData). Either way, this is probably still something that could be fixed.

She teleports mid, facing north, and [optionally] turns to face south while casting here; it's possible that the combatantData polling is happening while she's turning.

[valarnin]

Adding ActorMove and ActorSetPos lines is on my TODO list for OverlayPlugin which may be a better solution for things like this (and UWU titan jumps) going forward. I'm not sure if it's better to adjust the getCombatants call here to be more delayed now and then re-adjust later, or if it's better to just wait for the new lines to be available.

Assuming that timing is the root cause, of course.

[wexxlee]

It seemed to have fixed the issue fn. Thanks!

If it happens again next week I will repost here.

@Stealth3si , just checking in. Any further issues since changing your install path and enabling deucalion?

[xiashtra]

Since the original poster has not responded with any further problems, I am closing this issue. @Stealth3si if you are still having this problem, you can file another issue report.