It is currently 11 Sep 2025, 22:25
   
Text Size

Bug Reports xmage 1.4.13

by BetaSteward

Moderators: North, BetaSteward, noxx, jeffwadsworth, JayDi, TheElk801, LevelX, CCGHQ Admins

Re: Bug Reports xmage 1.4.13

Postby jeffwadsworth » 13 Jul 2016, 16:59

Kfchucky wrote:I had trouble yesterday trying to log onto server. Kept saying 'Unable to connect. Expecting previous version'. Today I uninstalled and reinstalled and it worked. However, now I cant load the mana symbols. It keeps going through the process, gets to 657/657 and then never finishes. I go to my plugin/images file and see the symbols folder (not a zip file and pref. has it clicked to save as zip file), I click on it and see small, medium and large folders. When I click on the folder, I see a gif with a picture of a cone, square and cirlcle. Im assuming the pictures never loaded and are stuck. How can I fix this problem? I have moved the folders, deleted just the gifs, and renamed folder. No luck

So the guy after my post gets a response, yet I havent gotten a response on my issue?
Did you check to see if any images are missing from your client once it is started up? Note that the symbols downloader does not exit on its own. You have to hit the X to close the window.
jeffwadsworth
Super Tester Elite
 
Posts: 1172
Joined: 20 Oct 2010, 04:47
Location: USA
Has thanked: 287 times
Been thanked: 70 times

Re: Bug Reports xmage 1.4.13

Postby rigeld2 » 13 Jul 2016, 18:05

20:01: rigeld2 casts Dark Salvation [72b] targeting rigeld2 Thalia, Heretic Cathar [840]
20:01: rigeld2 puts a Zombie [abd] token onto the battlefield
20:01: rigeld2 puts a Zombie [60e] token onto the battlefield
20:01: rigeld2 puts Dark Salvation [72b] from stack into his or her graveyard
20:01: rigeld2 attacks with 1 creature
20:02: Attacker: Kalitas, Traitor of Ghet [a9f] (3/4) blocked by Thalia, Heretic Cathar [840] (3/2)
20:02: Thalia, Heretic Cathar [840] deals 3 damage to Kalitas, Traitor of Ghet [a9f]
20:02: rigeld2 gains 3 life
20:02: Kalitas, Traitor of Ghet [a9f] deals 3 damage to Thalia, Heretic Cathar [840]
20:02: rigeld2 moves Thalia, Heretic Cathar [840] from battlefield to the exile zone

Dark Salvation didn't kill Thalia.

escplan9: fixed beyond 1.4.13v1
rigeld2
 
Posts: 30
Joined: 27 Feb 2015, 02:12
Has thanked: 0 time
Been thanked: 4 times

Re: Bug Reports xmage 1.4.13

Postby ShivaFang » 13 Jul 2016, 19:01

There is a memory leak in the mad AI - specifically the SimulatedNode class. The game field references a simulated copy of the Game and the customData field in the Game class is set to be a SimulatedNode.

This creates a looping reference that the GC does not clean up. The Mad AI creates a few simulated games this way every priority (to test the new min/max values for each possible action)

I discovered it when working on my MonteCarlo search AI and I was the mad AI as a base class as a default (mostly because the Computer Player AI throws errors during attacking and blocking) Since MonteCarlo simulates hundreds of games every priority these seriously added up. Once I stopped extending ComputerPlayer7 and the SimulatedPlayer the memory leak went away. I was sure it was my code for the longest time before I did reasearch on how to do a Heap Dump and discovered the real culprit!

I've tried clearing the customData field before creating each new game, but that didn't seem to do the trick (I'm not sure where the base class sets it)
ShivaFang
 
Posts: 101
Joined: 25 Jun 2016, 01:15
Has thanked: 26 times
Been thanked: 3 times

Re: Bug Reports xmage 1.4.13

Postby jeffwadsworth » 13 Jul 2016, 19:17

ShivaFang wrote:There is a memory leak in the mad AI - specifically the SimulatedNode class. The game field references a simulated copy of the Game and the customData field in the Game class is set to be a SimulatedNode.

This creates a looping reference that the GC does not clean up. The Mad AI creates a few simulated games this way every priority (to test the new min/max values for each possible action)

I discovered it when working on my MonteCarlo search AI and I was the mad AI as a base class as a default (mostly because the Computer Player AI throws errors during attacking and blocking) Since MonteCarlo simulates hundreds of games every priority these seriously added up. Once I stopped extending ComputerPlayer7 and the SimulatedPlayer the memory leak went away. I was sure it was my code for the longest time before I did reasearch on how to do a Heap Dump and discovered the real culprit!

I've tried clearing the customData field before creating each new game, but that didn't seem to do the trick (I'm not sure where the base class sets it)
You will notice Mage_Ai project on the magefree github site. This was BetaSteward's project that has apparently been abandoned for now. You may find some interesting things in there.

https://github.com/magefree/mage/tree/M ... provements
Last edited by jeffwadsworth on 13 Jul 2016, 19:19, edited 1 time in total.
jeffwadsworth
Super Tester Elite
 
Posts: 1172
Joined: 20 Oct 2010, 04:47
Location: USA
Has thanked: 287 times
Been thanked: 70 times

Re: Bug Reports xmage 1.4.13

Postby ShivaFang » 13 Jul 2016, 19:18

Define 'disabled'? It's still one of the options when you download the client and server. - Human, mad, and draftbot.
EDIT: Nevermind, you edited your post. But,yes, the mad AI is still there and it has a memory leak.

I'm not familiar enough with gitHub to be understand what you mean by 'project'. There is a MCTS (that someone else pointed me to) but it threw all kinds of errors and I found it completely unusable.
ShivaFang
 
Posts: 101
Joined: 25 Jun 2016, 01:15
Has thanked: 26 times
Been thanked: 3 times

Re: Bug Reports xmage 1.4.13

Postby leshrac » 13 Jul 2016, 19:41

Song of the Dryads and Blood Moon interact in a weird way: I enchanted my opponents Blood Moon with Song of the Dryads. The result was: All non-basic lands were still mountains, and couldn't be tapped for mana any more.

I'm pretty sure the Song should just overwrite the Blood Moon and all the lands should work normally.


Edit: And another bug with Emrakul, the Promised End
First time i cast it. I got control over my opponents turn, and wanted to kill him by tapping his 2 Ancient Tombs and then playing a fetchland from his graveyard, he had 2x Crucible of Worlds in play. I couldn't do either, tried multiple times. Tapping the tombs didnt work, when i tried to play the land i got the usual prompt with 2 Crucibles in play (so i didn't mix up the priority), but the land never entered the battlefield.
Edit 2: As it worked in a different game, i might still have mixed up the priority.

And while i'm at it, this bug is older (last version, encountered it a week ago), but i doubt it's fixed: I had a Gitrog Monster in play. During my upkeep with the sacrifice trigger on the stack my opponent bounced it with Karakas. I let that resolve and then of course didn't sacrifice a land to the Gitrog. Then (i realized that a few turns later, but that must have triggered it) i had the Gitrog Monster in hand and in the graveyard (i only play one, couldn't be a different card). Trying to cast the one in hand didn't work.
Last edited by leshrac on 14 Jul 2016, 11:59, edited 1 time in total.
leshrac
 
Posts: 33
Joined: 31 Dec 2014, 00:05
Has thanked: 2 times
Been thanked: 0 time

Re: Bug Reports xmage 1.4.13

Postby Moosejuice » 14 Jul 2016, 01:47

Not sure why the UI is bugging out. I did a forum search and can't seem to find anyone else with this issue. The picture is of the ui while it is bugging. It keeps changing randomly.

Screenshot is attached.
Attachments
screenshot.png
Moosejuice
 
Posts: 1
Joined: 14 Jul 2016, 01:42
Has thanked: 0 time
Been thanked: 1 time

Re: Bug Reports xmage 1.4.13

Postby zefferal » 14 Jul 2016, 03:32

Crop Sigil won't let me activate at opponent's end step with delirium. Not sure whether it's timing, targeting or some other bug, but I can't activate ability.

Edit: Tried activating at various other times as well to no avail. I have creatures in the GY, no lands though.

Edit 2: It worked finally once I had a land in the graveyard. Programming seems to require both legal targets to activate, rather than none ("up to" language, often used on Planeswalker's + abilities).

Here's something else though - When I did finally get it to work, it triggered two instances of bloodbriar. Now, the Crop Sigil was sacrificed, but there were no bloodbriars on the battlefield when this went down, so I'm not sure why two of them (checked the log, the triggers referred to two unique IDs) triggered when it was Sacrificed.

escplan9: Bloodbriar issue is fixed beyond 1.4.13v1. Crop Sigil also fixed beyond 1.4.13v1 now.
Attachments
bloodbriar.PNG
crop.PNG
Last edited by zefferal on 14 Jul 2016, 03:54, edited 1 time in total.
zefferal
 
Posts: 7
Joined: 24 Jun 2016, 07:40
Has thanked: 0 time
Been thanked: 0 time

Re: Bug Reports xmage 1.4.13

Postby zefferal » 14 Jul 2016, 03:49

Not sure what card caused this particular weirdness, but I blocked all (four) attacking creatures with my watcher in the web, in this order:
Ulrich, Kessig Dire Swine (w/ trample), Howlpack wolf, Incorrigible youths. I misplayed, and deserved a full 6 trample damage from the swine instead of 1. But, for some reason, I took 7. Transcript below, screenshot as well.

05:38: Turn 20 zorkmead (9 - 8)
05:38: zorkmead draws a card
05:38: zorkmead casts Incorrigible Youths [910]
05:38: zorkmead puts Incorrigible Youths [910] from stack onto the Battlefield
05:38: zorkmead attacks with 4 creatures
05:38: Attacker: Ulrich, Uncontested Alpha [f18] (6/6) blocked by Watcher in the Web [48f] (2/5)
05:38: Attacker: Kessig Dire Swine [e6e] (6/6) blocked by Watcher in the Web [48f] (2/5)
05:38: Attacker: Howlpack Wolf [f86] (3/3) blocked by Watcher in the Web [48f] (2/5)
05:38: Attacker: Incorrigible Youths [910] (4/3) blocked by Watcher in the Web [48f] (2/5)
05:38: Ulrich, Uncontested Alpha [f18] deals 6 damage to Watcher in the Web [48f]
05:38: zefferal loses 7 life
05:38: Howlpack Wolf [f86] deals 3 damage to Watcher in the Web [48f]
05:38: Incorrigible Youths [910] deals 4 damage to Watcher in the Web [48f]
05:38: Watcher in the Web [48f] deals 2 damage to Kessig Dire Swine [e6e]
05:38: Watcher in the Web [48f] died
05:39: zorkmead casts Kessig Prowler [87b]
05:39: zorkmead puts Kessig Prowler [87b] from stack onto the Battlefield
05:39: Turn 21 zefferal (9 - 1)
05:39: Ability triggers: Crop Sigil [001] - At the beginning of your upkeep, you may put the top card of your library into your graveyard.
05:39: Ability triggers: Ulrich, Uncontested Alpha [f18] - At the beginning of each upkeep, if a player cast two or more spells last turn, transform Ulrich, Uncontested Alpha [f18].
Attachments
watcher.PNG
zefferal
 
Posts: 7
Joined: 24 Jun 2016, 07:40
Has thanked: 0 time
Been thanked: 0 time

Re: Bug Reports xmage 1.4.13

Postby Mikail » 14 Jul 2016, 17:23

Gruul Charm has some debug string in it's hover-over tooltip as seen on attached screenshot (2nd mode, the "Gain control of..." part).
gruul charm ttip.png


LevelX: Fixed beyond 1.4.13v1
Mikail
 
Posts: 17
Joined: 07 May 2016, 06:43
Has thanked: 0 time
Been thanked: 0 time

Re: Bug Reports xmage 1.4.13

Postby MGA » 14 Jul 2016, 17:28

after Brisela, Voice of nightmares died i wasn't able to cast spell with less than 3 CMC the following turns,
BUG CHECK PLEASE

LevelX: Fixed beyond 1.4.13v3
Last edited by MGA on 14 Jul 2016, 17:37, edited 1 time in total.
MGA
 
Posts: 19
Joined: 26 Jun 2015, 11:56
Has thanked: 0 time
Been thanked: 0 time

Re: Bug Reports xmage 1.4.13

Postby Poit » 14 Jul 2016, 17:35

Hey, I noticed today that the “becomes unattached” ability on Stitcher's Graft as well as Grafted Wargear does not result in a forced sacrifice of the equipped creature if the "unattachement" is triggered by the destruction of the equipment.

I noticed this in a sealed EMN game on the xmage.de server, but replicated it here with the AI to confirm that the sac-on-unequip ability actually triggers and that the issue also exists with the Grafted Wargear.

In this replication, I attached Stitcher's Graft to a Stoneforge Mystic and then destroyed the equipment via Disenchant. The ability triggers, but keeps the Mystic [1c4] in play without any further message, the log says:

6:52 PM: Poit activates: Equip. from Stitcher's Graft [458] targeting Stoneforge Mystic [1c4]
6:52 PM: Poit casts Disenchant [9ff] targeting Stitcher's Graft [458]
6:52 PM: Stitcher's Graft [458] was destroyed
6:52 PM: Poit puts Disenchant [9ff] from stack into his or her graveyard
6:52 PM: Ability triggers: Stitcher's Graft [458] - Whenever Stitcher's Graft [458] becomes unattached from a permanent, sacrifice equipped permanent.

I dug a bit into the code, and noticed that the SacrificeEquippedEffect ability says
Code: Select all
Permanent equipment = game.getPermanent(source.getSourceId());
if (equipment != null && equipment.getAttachedTo() != null) {
    do stuff to trigger sacrifice...
}
return false;
I have no knowledge of the mage framework and actually never really used Java before (some C++), but could the problem be caused by the ability applying the checks when Disenchant already resolved and the equipment is in the graveyard, returning null on the getAttachedTo method?

Disenchant Graft.png


LevelX: Fixed beyond 1.4.13v3
Poit
 
Posts: 12
Joined: 18 Jun 2016, 14:35
Has thanked: 0 time
Been thanked: 0 time

Re: Bug Reports xmage 1.4.13

Postby Agetian » 14 Jul 2016, 17:54

v1.4.13r1: My opponent cast Dark Salvation (X=2), but my creature (which was supposedly targeted) didn't get -2/-2 until end of turn. Could have been a miscast by my opponent, or could be a bug, but my opponent said that the targeting/debuff portion did not work as it was meant to.

- Agetian

escplan9: Fixed beyond 1.4.13v1
Agetian
Programmer
 
Posts: 3489
Joined: 14 Mar 2011, 05:58
Has thanked: 684 times
Been thanked: 572 times

Re: Bug Reports xmage 1.4.13

Postby Mikail » 14 Jul 2016, 19:01

Falkenrath Gorger + Asylum Visitor (& probably any Vampire with printed Madness) + Olivia, Mobilized for War - Haste part of the triggered effect may not affect entering Vampire properly, please read further for more details.

When I cast Falkenrath Gorger and then discarded Asylum Visitor with Olivia, Mobilized for War 's triggered ability, two Madness pop-ups appeared, I have used the first one, Asylum Visitor entered the battlefield, Olivia triggered again and I used the ability to give Asylum Visitor Haste and +1/+1. Then the second Madness trigger resolved for the Visitor and I have cancelled it (by choosing "No" at first pop-up, but from what I have tested, the choice at this point does not matter). Asylum Visitor lost Haste and was both visually and functionally affected by Summoning Sickness.

I was able to avoid this issue by cancelling the first Madness pop-up and then using only the second one.

Please watch attached video for more details.

This issue exists probably since SOI addition and was noticed on local server playing against AI.

Double madness trigger.gif


LevelX: Fixed beyond 1.4.13v3
Mikail
 
Posts: 17
Joined: 07 May 2016, 06:43
Has thanked: 0 time
Been thanked: 0 time

Re: Bug Reports xmage 1.4.13

Postby MGA » 14 Jul 2016, 21:29

the +1 ability of TAMIYO seems not working on GIDEON as creature

attacked with 1 human and gideon only 1 card was drew.
CHECK

escplan9: Fixed beyond 1.4.13v1
MGA
 
Posts: 19
Joined: 26 Jun 2015, 11:56
Has thanked: 0 time
Been thanked: 0 time

PreviousNext

Return to XMage

Who is online

Users browsing this forum: No registered users and 23 guests

Main Menu

User Menu

Our Partners


Who is online

In total there are 23 users online :: 0 registered, 0 hidden and 23 guests (based on users active over the past 10 minutes)
Most users ever online was 7967 on 09 Sep 2025, 23:08

Users browsing this forum: No registered users and 23 guests

Login Form