War Machine WIP

[dosraider]

*Downloaded*
Will try it out this weekend or so .....

[Larry Laffer]

Was gonna post this on your blog, but this ought to be just as good.


My findings:

[Bug] Units and menus are drawn on different(and wrong) layers: Menus should be on top. linky

[Bug] Menus are click-through in certain conditions(possibly because of the above layers issue). If you order the blue unit in that screenshot to move, you'll probably click over one of the factories. As such it'll bring up the build menu instead of moving the unit.

[Bug] Enemy units don't count as obstacles/uncrossable terrain. In that image, the red unit can move to and from the square between the mountains and the blue tanks.

[Bug] Very possible memory leak. While idle, the game was using between 50 and 75% of my 2.4GHz Celeron and had an average memory allocation increase of 600 kb/s. I've left it running for around an hour, came back to find it unresponsive and with well over 450 megs of memory in use.

[Feature] Infantry does not move across mountains(they do in the original games).



Still, far better than the last version

[CPT Worm]

Thanks, Larry. I think I have corrected the a few of the bugs you noticed:

[Bug] Units and menus are drawn on different(and wrong) layers: Menus should be on top. linky

[Bug] Menus are click-through in certain conditions(possibly because of the above layers issue). If you order the blue unit in that screenshot to move, you'll probably click over one of the factories. As such it'll bring up the build menu instead of moving the unit.

I just needed to move the image generation for the sprites before the menus.

I also added if clauses that prevent build menus from opening when trying to move a unit.

[Bug] Very possible memory leak. While idle, the game was using between 50 and 75% of my 2.4GHz Celeron and had an average memory allocation increase of 600 kb/s. I've left it running for around an hour, came back to find it unresponsive and with well over 450 megs of memory in use.

I think I found the problem to that as well. Apparently, a few items escaped my shut down loop, but they should all be now freed from memory before the game exits.

As for the other two....those involve the movement algorithm. It's pretty juryrigged at the moment, and I don't want to screw with it while it's stable.
I think I have a good idea on how to correct it, but I will wait until before the release of the next version to fix it.

Thanks again, you've been a great help thus far!

[Larry Laffer]

No problem, glad to help!

Linky!

Circled numbers: CPU in use / Working Memory.

Seems the memory issue is still there. Between the time I've taken that screenshot and finished editing it, it had doubled its used memory.


As for the movement algorithm, it's still quirky, it doesn't fully calculate the paths. In that screenshot, all units have issues with their moveable areas(it *seems* it might not be able to go around the corners properly but I'm not willing to bet money on it).

Also, while in the testing period, think you could disable the end-of-round music or at least end the round before it finished playing? Ought to make testing a tad faster.

[CPT Worm]

Circled numbers: CPU in use / Working Memory.

Seems the memory issue is still there. Between the time I've taken that screenshot and finished editing it, it had doubled its used memory.

Does the problem persist after you close the program?
I think this API is very clunky, and it might be memory-intensive over efficient...

As for the movement algorithm, it's still quirky, it doesn't fully calculate the paths. In that screenshot, all units have issues with their moveable areas(it *seems* it might not be able to go around the corners properly but I'm not willing to bet money on it).

I think I know the problem for that one. That unit is in the square that helps units turn corners. It's probably not operating correctly since it's in that square. I'll fix that later.

Also, while in the testing period, think you could disable the end-of-round music or at least end the round before it finished playing? Ought to make testing a tad faster.

Well, I'm not happy with that transition, so I'm going to completely recode it.

However, my plan for the next week and change is to get the battle calculator functioning. Someone already has the data for that written out, so that should be easy enough. Only thing I have to worry about is animation, really.

[Larry Laffer]

The leak itself seems to be fixed but there's still the other error(guess I could've described it better) where the program keeps allocating memory while running. Yes it releases it upon closure, but it just keeps hoarding memory.

Really no problem with being memory-intensive over efficient, but at least thus far I think there's an infinite loop in memory allocation somewhere.



Alright, didn't really get it to 1 GB memory(it died when I tried to attach a debugger) but extrapolating, the current version should get there within 30-40 minutes.


Yea, I've seen games use less

[Larry Laffer]

Since we know that Mr. Worm is not dead... or at the very least he's a self sufficient zombie, is there any update on this?

[CPT Worm]

Since we know that Mr. Worm is not dead... or at the very least he's a self sufficient zombie, is there any update on this?

Well, to be quite honest, I haven't been working much on this as of late. Work kinda got in the way.

However, I am planning to get back to work on it soon; I need to redesign the engine, though.

[Larry Laffer]

Faster! Work faster, I say!