[Ydiss]

By the way, I'm having serious issues compiling this map. Despite the fact that it's smaller in size and VIS data than dmxfire (which takes around an hour to compile) it is taking 12+ hours to compile which I simply cannot understand.

I've posted this on the .map forums but have had no reply helping me yet so until this issue is resolved, this map will not ever be available oOo:

[Bazooka_Joe]

12 hours to compile!!? Holy shiz nit! Hrm. . .are you doing anything to piss off MOHRadiant, like custom textures, or cutting down on any of the geometry? That's a long time, man. Well, so far it looks GREAT, and I wish you the best of luck. I'll move this and leave a copy here.

[Ydiss]

[quote="Bazooka_Joe":f9c7c]12 hours to compile!!? Holy shiz nit! Hrm. . .are you doing anything to piss off MOHRadiant, like custom textures, or cutting down on any of the geometry? That's a long time, man. Well, so far it looks GREAT, and I wish you the best of luck. I'll move this and leave a copy here.[/quote:f9c7c]

No custom textures, geometry is pretty smart and ordered (I never use CSG subtract, nor hollow and only overlap brushes when I absolutely have to, which is rare). Nothing to cause the VIS tree to clog up like it is.

It doesn't make any sense at all... All evidence points to this new map to compile faster than dmxfire. Less numportals, fewer faces and brushes way fewer static models (like 59 compared to 250+).

I've used detail brushes extensively.

The map only has 240k vis data (dmxfire has 330k+).

dmxfire takes 1 hour no more to compile on my new machine.

I abandoned my new map's compile process after it had been stuck on 99% VIS for 5 hours.

I'm going to spend this week (well, if all goes well hopefully just one night) on debugging the map.

That is I'm going to make copies and start deleting chunks of it to see if there's a particular area causing this issue.

Other than that, I'm not sure what to do... I can't add any more detail brushes or I may as well just not bother trying to VIS at all and the map would be slow as shit to play.

I'll need a lot of that luck, I'm afraid cry:

[Zoner]

Originally Posted by Ydiss

Soon as it's ready for testing I'll send you a copy, Zoner.

Actually, it would prolly be better if you sent Boo the map...I don't have Rcon or anything.

Drop by allaboutnothing.net and post 'er up when you get it done.

[Ydiss]

OK, will do. Is the AAN server SH or AA? Realism or stock?

[Bazooka_Joe]

Originally Posted by Ydiss

OK, will do. Is the AAN server SH or AA? Realism or stock?

Spearhead, with AAN Realism. I'm sure that Boo could adjust toyour wants for the map for proper testing.

[Ydiss]

Oh it's ok... aM plays stock AA or so I recall, so it'd be good to have two different game types testing it.

I would have to say the map would suit stock better. It's quite open and spacious. Plenty of cover to be used but it'd be a bit of a frag fest with realism settings. It'll be worth testing on each to see if there are any flaws in the design for either mode.

I thought a little about an objective version for this map, as requested. I would only make it a simple blow-up-the-target objective, but there's nothing preventing me from making the target different to a flak cannon.

I was thinking of having the Allies spawn in the Hotel Car park (South West corner - check the blue print out) and the Axis spawn in the opposite corner and Town Hall balcony area.

The objective would have to be on the balcony or park (North East of map)... Any suggestions what this could be? If I had to I would extend the map a little bit (add an interior to an empty building, or a new area).

Any suggestions will be considered and credited. Otherwise I'll come up with something myself.

[wintersforge]

Originally Posted by Ydiss

Oh it's ok... aM plays stock AA or so I recall, so it'd be good to have two different game types testing it.

I would have to say the map would suit stock better. It's quite open and spacious. Plenty of cover to be used but it'd be a bit of a frag fest with realism settings. It'll be worth testing on each to see if there are any flaws in the design for either mode.

I thought a little about an objective version for this map, as requested. I would only make it a simple blow-up-the-target objective, but there's nothing preventing me from making the target different to a flak cannon.

I was thinking of having the Allies spawn in the Hotel Car park (South West corner - check the blue print out) and the Axis spawn in the opposite corner and Town Hall balcony area.

The objective would have to be on the balcony or park (North East of map)... Any suggestions what this could be? If I had to I would extend the map a little bit (add an interior to an empty building, or a new area).

Any suggestions will be considered and credited. Otherwise I'll come up with something myself.

Just let me know when its ready and we will put it throught the paces...

aM>AAN

[Ydiss]

I'm currently debugging the map through deduction. I deleted an entire half of it and am compiling it now. The VIS completed in reasonable time so, by deduction, I should be able to presume the half I deleted might have the problem in it.

I'll let this finish then I'll delete the other half instead. If it takes ages then I should know which half the error is in.

Then it's a case of deleting half of that half, and then I'll start homing in on particular structures.

Hopefully I'll have the problem sorted by this weekend.

However... Worst case scenario is that I can't find the problem (ie, it's a problem with the whole rather than a part) and the map will have to be binned.

If it comes to that stage I'll have to ask someone here who I can trust to compile it for me, see if that works.

[Bazooka_Joe]

I don't know much about MOHRadiant itself, but could it simply be some sort of unfinished Geometery in the map? Meaning, that a corner is not connected, or something in that sense?

[Ydiss]

No, I think I know what the problem is. It's likely inefficient VIS. There is VIS data but it's not efficient at blocking off portals from eachother so too many areas can be seen from too many other areas causing a lot of VIS trees (if you've ever seen a fractal diagram and how it splits into infinite trees as the calculations become more dense that's roughly how VIS works.. the more complex the data the more branches and the longer it takes).

I may have to completely redesign the map into 4 quarters that are all totally separated from eachother by VIS (that is each area has no line of sight to the other).

It might make the map uglier but at least I'll be able to compile it and it will run smoothly.

[nomar]

Hey Ydiss I'm an admin on the AAN server and we'll definitely want to try it out. I'll just check back to this thread to see when it's ready.

And yeah, we'd love an OBJ version too. It would be cool if the Axis were the attacking team because there are so many OBJ maps where Axis defends.

Good work. We still have xfire courtyard in our standard DM rotation.

Jeff

[Ydiss]

Ok, I'll consider that. Won't be ready for a while though... It looks like I'll have to completely rework the entire map to fix the VIS issue.

Could be a few weeks now. oOo:

no matter how good the map is, no one will play it. No one plays custom maps hake:

[PeskyAtheist]

Originally Posted by Sclass12

no matter how good the map is, no one will play it. No one plays custom maps hake:

bullshit they dont. ive played plenty of custom maps on different servers. good map bro, nice work.