Bug Reports (snapshot builds)

[timmermac]

With the most recent snapshot build as of this morning, I am noticing some fairly severe performance degradation. Forge is almost sticking at various points, particularly as I'm starting a match.

Edit: I'm also noticing that Forge is using somewhere over a gig of memory right now. Is there a memory leak happening somewhere in the code right now?

Edit 2: It eventually crashed with a heap space error.

[Diogenes]

I also came to report that I experienced slowdown in yesterday's daily (edit: looks like there's a new one up.) It began at the start of the seventh match in a series in sealed deck, all prior games ran smoothly. The UI noticeably lagged, and Forge was using 50% of my cpu (on a dual-core.) RAM usage was at 550 megabytes, which I think is just a little above normal and not anything unexpected.

I have seen this in the past when playing many consecutive matches without a restart, but I haven't had a marathon session in a while. On a related note, everything Theros looks really good so far.

(Edit: I'm going to move all other issues into the next post for readability.)

[Diogenes]

Anthousa's heroic ability goes on the stack, but the user is never prompted to choose target lands for the ability to affect.

Dragon Mantle gives the enchanted creature a constant +1/+1 in addition to the pump ability.

A few costs are wrong:

Dragon Mantle should cost R, not 1R
Reaper of the Wilds should cost 2BG, not 2BR
Hammer of Purphoros should cost 1RR, not 2R

UI: There's a minor graphical issue with the round buttons in the dock panel. A white square appears around them after they've been highlighted, and does not go away. This doesn't affect the targeting-arc button.

[friarsol]

Anthousa's heroic ability goes on the stack, but the user is never prompted to choose target lands for the ability to affect.

Dragon Mantle gives the enchanted creature a constant +1/+1 in addition to the pump ability.

A few costs are wrong:

Dragon Mantle should cost R, not 1R
Reaper of the Wilds should cost 2BG, not 2BR
Hammer of Purphoros should cost 1RR, not 2R

Thanks these should be fixed now.

[drdev]

I'm guessing the memory issues are the result of my FSkin changes. In order to support changing skins on the fly, I had to track any components that use a skin-based property. I'm guessing some places must be creating lots of temporary components that are getting stuck in memory because the skin code is maintaining a pointer.

I had tried to build some garbage collection into what I did, but I couldn't figure out how. I wasn't worried at the time since i didn't see any noticable problems. Does anyone with more java experience have suggestions. At the very least we could add a function to remove a component from the skin tracking collection.

As for performance problems, when and where are people noticing a slowdown? Could it be tied to the memory issue after leaving Forge open long enough?

[Diogenes]

As for performance problems, when and where are people noticing a slowdown? Could it be tied to the memory issue after leaving Forge open long enough?

I don't think that's the case for me, I think it has more to do with Forge eating up one of my cores. Normally Forge uses about 20-40% of my cpu-time during gameplay, and falls to near 0% when unfocused or idle, but sometimes it just maxes out (today this was at the start of a match before hands were even dealt.) This has only ever happened during extended play sessions. I'm still not sure if it spikes suddenly or ramps up, but I'll look into it.

The only memory issue I've noticed is that in the last week Forge fails to load ("failed to create virtual machine") if I have my web-browser open. Granted, firefox takes between 250-600 megabytes of physical memory (I have a lot of open tabs ^_^), but there should still be gigs left over.

Edit: I got it to happen again. This time it didn't go off until the last game of match seven. RAM usage was still under 600. The cpu wasn't capped at 50% this time (got into the 60s), but the slowdown wasn't as bad.

[Diogenes]

Heliod's description text is missing his token creation ability. Shipwreck Singer's description text is missing the "an opponent controls" clause to its first ability. This is a text-only issue, both abilities behave correctly.

[swordshine]

Heliod's description text is missing his token creation ability. Shipwreck Singer's description text is missing the "an opponent controls" clause to its first ability. This is a text-only issue, both abilities behave correctly.

Fixed. Thanks, Diogenes.

[Hanmac]

cards like Sway of Illusion that has "Any number of target creatures" does not trigger the Heroic abillity

[timmermac]

As for performance problems, when and where are people noticing a slowdown? Could it be tied to the memory issue after leaving Forge open long enough?

I don't think that's the case for me, I think it has more to do with Forge eating up one of my cores. Normally Forge uses about 20-40% of my cpu-time during gameplay, and falls to near 0% when unfocused or idle, but sometimes it just maxes out (today this was at the start of a match before hands were even dealt.) This has only ever happened during extended play sessions. I'm still not sure if it spikes suddenly or ramps up, but I'll look into it.

The only memory issue I've noticed is that in the last week Forge fails to load ("failed to create virtual machine") if I have my web-browser open. Granted, firefox takes between 250-600 megabytes of physical memory (I have a lot of open tabs ^_^), but there should still be gigs left over.

Edit: I got it to happen again. This time it didn't go off until the last game of match seven. RAM usage was still under 600. The cpu wasn't capped at 50% this time (got into the 60s), but the slowdown wasn't as bad.

With snapshot version 15, I could play 2 or 3 matches in quest mode before it slowed to an absolute crawl. At that point, Forge memory usage is usually up in the 900+ megabyte territory.

[spr]

As for performance problems, when and where are people noticing a slowdown? Could it be tied to the memory issue after leaving Forge open long enough?

I don't think that's the case for me, I think it has more to do with Forge eating up one of my cores...

If you are using Eclipse you might want to check out JVM Monitor. This lets you monitor memory/thread usage of any java app regardless of how it was started - see screen shot (running version 1.4.7 exe in the middle of a match).

Cheers,
Steve

[timmermac]

As for performance problems, when and where are people noticing a slowdown? Could it be tied to the memory issue after leaving Forge open long enough?

I don't think that's the case for me, I think it has more to do with Forge eating up one of my cores...

If you are using Eclipse you might want to check out JVM Monitor. This lets you monitor memory/thread usage of any java app regardless of how it was started - see screen shot (running version 1.4.7 exe in the middle of a match).

Cheers,
Steve

I think Java itself comes with something similar - but you might have to hunt around a bit in your Java directory to find the file.

[RumbleBBU]

I think the memory leak/whatever problem is a bit more serious than just slowdown. This just happened to me, played ~5-7 Quest duels, visited the spell shop in between matches. I noticed increasing sluggishness in the game. Then, visiting the spell shop again, the game just freezes for a moment, and then:

| Open

EDT > java.lang.OutOfMemoryError: Java heap space
at com.mortennobel.imagescaling.ResampleOp.doFilter(ResampleOp.java:155)
at com.mortennobel.imagescaling.AdvancedResizeOp.filter(AdvancedResizeOp.java:81)
at forge.gui.toolbox.imaging.FImagePanel.getResampledImage(FImagePanel.java:228)
at forge.gui.toolbox.imaging.FImagePanel.doQualityPaint(FImagePanel.java:210)
at forge.gui.toolbox.imaging.FImagePanel.paintComponent(FImagePanel.java:204)
at javax.swing.JComponent.paint(Unknown Source)
at javax.swing.JComponent.paintChildren(Unknown Source)
at javax.swing.JComponent.paint(Unknown Source)
at javax.swing.JComponent.paintChildren(Unknown Source)
at javax.swing.JComponent.paint(Unknown Source)
at javax.swing.JComponent.paintChildren(Unknown Source)
at javax.swing.JComponent.paint(Unknown Source)
at javax.swing.JComponent.paintChildren(Unknown Source)
at javax.swing.JComponent.paint(Unknown Source)
at javax.swing.JComponent.paintChildren(Unknown Source)
at javax.swing.JComponent.paint(Unknown Source)
at javax.swing.JComponent.paintChildren(Unknown Source)
at javax.swing.JComponent.paint(Unknown Source)
at javax.swing.JLayeredPane.paint(Unknown Source)
at javax.swing.JComponent.paintChildren(Unknown Source)
at javax.swing.JComponent.paint(Unknown Source)
at javax.swing.JLayeredPane.paint(Unknown Source)
at javax.swing.JComponent.paintChildren(Unknown Source)
at javax.swing.JComponent.paint(Unknown Source)
at javax.swing.JComponent.paintToOffscreen(Unknown Source)
at javax.swing.BufferStrategyPaintManager.paint(Unknown Source)
at javax.swing.RepaintManager.paint(Unknown Source)
at javax.swing.JComponent._paintImmediately(Unknown Source)
at javax.swing.JComponent.paintImmediately(Unknown Source)
at javax.swing.RepaintManager$3.run(Unknown Source)
at javax.swing.RepaintManager$3.run(Unknown Source)
at java.security.AccessController.doPrivileged(Native Method)

I think this problem is relatively serious and needs to be addressed somehow before the next beta. As an ultimate (absolute last) resort, disable the troublesome code. Given the choice between a stable game and changes of skins without a restart, I believe most players would prefer a stable game...

[spr]

I think this problem is relatively serious and needs to be addressed somehow before the next beta. As an ultimate (absolute last) resort, disable the troublesome code. Given the choice between a stable game and changes of skins without a restart, I believe most players would prefer a stable game...

It is not necessarily related to the recent skins update since a suspiciously similar error was generated by someone using 1.4.7-r23046M. I will look into it.

Cheers,
Steve.

[MIC132]

Sometimes in commander the AI taps his mana for his commander (judging from the amount) but instead of the commander going tot he stack it just disappears. It's not in the command zone, not in grave, not in exile, nowhere. Never happened to me, just AI.