Spring RTS Engine

Yeah...

I tried to leave 2 comments about the more than 30 second stalling, but they never showed up.

So I will continue posting complaints and suggestions here.

Mwha?!?!

Moth3r must have edited Wordpress specifically for sourceforge so I'll ahve to update it to the official version sorry.

The antistall algorithm will have to take into account planned constructions whereas atm it only takes into account ongoing cosntruction......

As for RC19 source, sorry about that as usual my efforts are thwarted by time. I'll try and get it to you before midnight

I just thought I should let you know, an impending patch of mine adds "Fight" (id=16) to the available commands of mobile, air, and factory units, which breaks NTAI (the com fails to move, but will use his d-gun, unlike a truly dumb com).

It is not the fact that "Fight" exists, but the fact that it is an available command that breaks it.

The RC19 release is REALLY great. Totally smooth as silk, regardless of how long the game goes. Great stuff, guys :)

AF wrote:The antistall algorithm will have to take into account planned constructions whereas atm it only takes into account ongoing cosntruction......

I really don't thinks so; the problem is new for RC19 and it just has to do with NTai ignoring either:

MaxStallTime

metalIncome

metalStored

Of possibly just a problem with the blocking of commands.

I set metalIncome=1, metalStored=1, and MaxStallTime=6. But the 1st thing NTai built was an advanced vehicle plant, which caused stall for about a minute.

Something is screwed up.

reduce the stall tag, and chane the multipleirs to a non '1' value.

Are you usign B_FACTORY or B_FACTORY_CHEAP or B_RULE?

coravp.

Which used to be caught properly, but now isn't.

I really hope you didn't get rid of the specified units not being caught...

I guess I'll add it to always_antistall, just in case.

But I want everything to always_antistall, unless they are defined in never_antistall.

I didnt change the antistall.

the onyl economy rule I changed was the estorage and mstorage which was meant to fix a crashbug...

It'll be your rule value modifiers, check antistall=3;

Hmm.. Well, this is a new problem.

I'll poke around, and make sure I didn't make a syntax error.

fiddle your values....

I agree with vassago; lag gone. dead. entirely. Much smaller logs too. Sw33t.

huh? wouldnt ti be better to prefix CMD_ to the define? CMD_FIGHT...

It shouldnt affect NTai, can i see the logs?

http://ntai.from-hell.net

Comments and page postign work again now, I also tied in a login/logout/register int eh top menu so people cn finalyl post in the forums.

there's also a pastebin if you wanna put logs in there....

PS:: The menu I'm talking about is the one at the very top, press the link and the menu will pull down

btw dgunning was disabled in RC17 as it had a bug and needs rewritting, are you sure you're using RC18 code?.

Sorry about not posting the source code last night, I had to fix darkstars somewhat

I fixed lindirs problem with antistall algorithms giving the go ahead for things it shouldnt have, and ti works better now. Although I ahve oen rot wo issues witht he buildtree currently in the installer.

AK's dotn do anythign they're nto beign picked up by any fo the classes in NTai, where they accidentally remvoed from the attackers/scouters tags?

NTai is now energy stallign more, this needs changing.

I'm gonan try and release RC20 soon enough, whcih will include proper ranged atatcks back in, and lindir, I'm making it look for plans too, so fi con a is due to build a adv kbot lab nearby and con b has a normal lvl 1 lined up narby to its planned position, it'll abandon and start the planned cosntruction of the adv kbotlab earlier while the original builder is still moving to the position.

I'll attempt to integrate plans into the antistall system too.

AF wrote:I fixed lindirs problem with antistall algorithms giving the go ahead for things it shouldnt have, and ti works better now. Although I ahve oen rot wo issues witht he buildtree currently in the installer.

Excellent.

AK's dotn do anythign they're nto beign picked up by any fo the classes in NTai, where they accidentally remvoed from the attackers/scouters tags?

Yes, in the version in the installer I made a typo. Actually I think you made a typo. Someone made a typo.

It's fixed in v0.8.0.1, and in v0.8.1, which I'll release tonight if I have access to a computer

NTai is now energy stallign more, this needs changing.

Now that antistalling is working again, I'll twiddle with the values.

I'm gonan try and release RC20 soon enough, whcih will include proper ranged atatcks back in, and lindir, I'm making it look for plans too, so fi con a is due to build a adv kbot lab nearby and con b has a normal lvl 1 lined up narby to its planned position, it'll abandon and start the planned cosntruction of the adv kbotlab earlier while the original builder is still moving to the position.

I'll attempt to integrate plans into the antistall system too.

Cool! I don't think it is absolutely neccessary, but it will still improve NTai's performance.

This is probably the wrong place to ask but how does the current metatag system work? I've tried using a [TAGS] section and adding some key=unitlist entries but I still get errors that the tag doesn't exist.

use lowercase tagnames

sent email earlier rc20 not compiling under linux
also tried removin Potential.cpp/.h same result

AF wrote:use lowercase tagnames

That didn't change anything, I'm still getting those errors and it even tells me repair is an invalid task keyword.

List of changes to source sofar is
Tobi's fixes for RC18
Removed SunParser.cpp/.h
Removed Potential.cpp/.h

error on compiling under gentoo/linux is http://pastebin.ca/76457

scons: *** ParentOfRoot instance has no attribute 'must_be_a_Dir' ????

did quick google might need to edit SConstruct will check it out later. Was using svn revision 1523 cause of ai problems with it. Might test it later but gtg atm

opps