Forge version 1.2.0

[Chris H.]

I thought that I might go ahead and start a topic for our upcoming Forge version 1.2.0 release. I am not going to associate a date of release at this point.

The new UI seems to be coming along nicely. Most of the functionality has been added. Some of the bugs have been fixed.

I guess that the biggest concern at this point is the slow downs that some people have noticed. I have not been able to replicate this problem but I have 6 GB of memory in my machine and I tend to play only a few games/matches at a time.

[Milod]

yes there are slowdowns in this version also in the newest snapshot i got win7 intel dual core 2ghz 2gb memory and decent video card

[jmartus]

I got them from 12/16 release on it athlon x4 635 6.0 gb memory

[Doublestrike]

I've got a few things to finalize and then I'll start playtesting. Once I actually start playing this game again I can probably fix this. I'll try to put in some good hours of play time this week and see if I can't nail it down (if others haven't nailed first).

[ArsenalNut]

I have noticed the slow down too. I find the slow down starts to get noticeable after about 45 minutes of playing. Eventually, the UI gets unresponsive after about two hours. I only the play the constructed mode. I am typically play testing a deck I found on the net against the AI so I might keep playing the same decks for a long time.

I did a little investigation. Something is going on with the memory for the new UI. When I start up Forge, the java process shows about 700MB being used. When I start to notice serious issues with UI response, the java process has over 3GB being used. My machine is a Win7 64bit with 16GB RAM. I am fairly new to Java and Eclipse so I haven't tried to do any profiling to find where all the memory is getting allocated. If there is something you want me to try, let me know.

[Doublestrike]

I had a (very) quick attempt at fixing the slowdown - the observers weren't being deleted before fresh ones were added at the start of each match, which might very well be the problem.

Will playtest when I have time; would be interested if anyone notices a difference.

[friarsol]

I had a (very) quick attempt at fixing the slowdown - the observers weren't being deleted before fresh ones were added at the start of each match, which might very well be the problem.

Will playtest when I have time; would be interested if anyone notices a difference.

Is there any reason we need to delete the observers one at a time? If we're going to re-init the observers each initMatch() shouldn't we just call deleteObservers() on any MyObservable before adding any new ones? Is there an equivilant of "deinitMatch"? That's really where we should be clearing things out at the end of a match.

[Doublestrike]

Is there any reason we need to delete the observers one at a time? If we're going to re-init the observers each initMatch() shouldn't we just call deleteObservers() on any MyObservable before adding any new ones? Is there an equivilant of "deinitMatch"? That's really where we should be clearing things out at the end of a match.

All great things, no reason for the one-at-a-time. A lot more streamlining can take place here with the careful amputation of the old UI stuff.

[silly freak]

A library I once used handles this with "Disposable"s. I guess it's a pattern of some sort; it boils down to this:

Code: Select all

//do some GUI initialization action in regard to the game.
Disposable d = gui.init(game);

//play a game
...

//remove the GUI's initialization from the game
d.dispose();


This keeps the code that does init and dispose close together, which, I think, is a good thing. Futhermore, it makes resetting something universal, not a special named method: The Disposable could be given to the game for automatic disposal when it's over.

Just my 2 cents

[Chris H.]

I thought that I might go ahead and start a topic for our upcoming Forge version 1.2.0 release. I am not going to associate a date of release at this point.

`
I would like to release the next beta in a few days, say this upcoming Friday.

[jmartus]

I tried the 12/26 beta version and it seams to slow down a lot after a few matches just trying to help out.

[slapshot5]

I tried the 12/26 beta version and it seams to slow down a lot after a few matches just trying to help out.

Try the next nightly build to see if this helps (should be a 12/27 build or later). I'm not sure it's completely fixed, but it should be somewhat better.

-slapshot5

[slapshot5]

Do we have any things that needs to be done before we do a beta release of 1.2.0?

I think we should turn off the View Any Hand and View Any Library Enabled/Disabled buttons in Dev Mode because I don't think they do anything.

-slapshot5

[Chris H.]

Several people have noticed a crash report stating that "Buffers have not been created". This appears to be an uncommon event.

Snacko posted a brief message about this bug in message Re: Current Known Bugs list

I am unsure on how to address this one problem. Apparently it is caused by:

This error shows up because you call Swing methods outside of EDT.

`
This one may take awhile to track down, at least it does not pop up very often and I have not seen it myself.

I think that we should go ahead a release the beta on Friday. We could use some feedback.

[Chris H.]

I think we should turn off the View Any Hand and View Any Library Enabled/Disabled buttons in Dev Mode because I don't think they do anything.

`
I clicked on both of the hand and both of the library icons and got a window displaying a list of cards. It looks like there is currently no check to see if this is enabled or disabled.

I turned off dev mode and both of these options became disabled.