Map assessment and pathing enhancements

[Nelsona]

Until some recommended stuff shows up, allow me to introduce a few parts from working stage.
Because map is large and mainly a sort of plane field, builder chosen for executing paths will be PathsBuilder from XC_EditorAdds embedded in XC_Engine 24. Yes, it does lousy jumps but right now I don't have issues with those "options" as long as I can delete buggers and cleaning up trash jumps.
This is an image from a work-in-progress stage...

CheckMate_0.PNG (293.35 KiB) Viewed 2603 times

If a Bot is trying to jump down there from the tower he can make a tattoo on his back on which it says: I'm so dead... For me is hard to think what was all about these ideas and how ReachSpecs calculations were executed... But... it's not a problem... XC does the dirt, XC will remove the dirt, just like that.

Why using this option ?
Builder does a better optimization in overcrowded spots (towers with ammo and LiftExit-s) and it's suitable for long paths in order to minimize processing with less points used for connections. It also works clean without PrunedPaths (garbage assets if you want my "private-public" opinion). Nothing is suffering from anything if these are missing. If DevPath can calculate an "Entry" into network, the rest of navigation process can go smoother like a good Engine oil. At end of task map will get flushed by any byte which is not part of map but a result from routing process which is dynamic and this data is created and changed multiple times per second during map's run-time.

[evilgrins]

Given rifles are at the corners, probably would be placing weapons along the chess board where the squares cross.

If you want me to add those first, before you path, I can get on that.

Probably gonna put UTDMT around the edges of the board, likely only 2. Anybody falls down there won't be there for long!

Added: Got a thread about this over in ut99.org, to identify the map in the first place. New replies indicate the one I sent you before isn't the original... so hold off in editing while I do stuff on the original.

I got a post on the mapper's facebook group waiting on permission if the mapper is still around.

[Nelsona]

Look at this, we can have other way if this is not suitable...

DM-Checkmate_rS1023.7z

(2.26 MiB) Downloaded 113 times

Weapons with too slow projectiles were not added - maybe not suitable in this stage but... we can have them later if are needed.
Paths have been built with XC_PathsBuilder and adjusted with XC_PathsWorker. It doesn't include null paths or such useless data due to latest updated XC_PathsWorker...

WIP journey facts.
It took me longer because first stage went flawed. I used some checker for fixing certain brushes said with duplicated models - geometry is probably bad here - as result of tweaking I had to rebuild map. That was the stage when things were turned into crap. One surface was gone, either way surface was drawn but other one was gone, and so on. I had to resume the work from original map and no longer "fixing" anything, so to speak geometry is original. In this stage with miss-aligned and rotated brushes, errors are not a surprise for me...

[evilgrins]

Hypothetical: Maps that are edited with the new patch cannot be edited with the previous version.

Is there a way to make a map edited under 469 edited by previous editors?

Reason I'm asking · https://ut99.org/viewtopic.php?f=5&t=15451 · he edited that with 469 and while I can open it with my editor, any changes I make don't stay as saving crashes the editor.

Yes, I still don't have the new patch as it does bad things to UTDM-monsters on the Red Team.

[Nelsona]

I opened map - but let me figure how many resources are taking the map because otherwise it won't be saved (- and yes, I do not use 469b unless I have to try solving whatever issue with something). I had some of those ugly stock issues solved years ago and I won't drop them away just because certain 469 function is no longer crashing game any more but it's the same miserable crap as in 436. They are working harder to break backward compatibility... but doing this nice and slow...

Resuming topic. Let me look at it and I will tell you if this is a "469-ONLY" map or a common UT map.

2 Minutes Later
I'm hating myself when I do good predictions...

Torlan_Exclusive469.png (327.38 KiB) Viewed 2490 times

Original DevPath will clamp seeking at 1000 Nodes - we already have 1700++. It won't work.
And I was asking Mr. Trash about useless stuff embedded... He responded me (not using e-mail but the log):

Code: Select all

NoIncomingPath: PathNode7317 is not having incoming connections.
NoIncomingPath: PathNode7399 is not having incoming connections.
NoIncomingPath: PathNode7512 is not having incoming connections.
NoIncomingPath: PathNode7520 is not having incoming connections.
NoIncomingPath: PathNode7559 is not having incoming connections.
NoIncomingPath: PathNode7624 is not having incoming connections.
NoIncomingPath: PathNode7629 is not having incoming connections.
NoIncomingPath: PathNode7644 is not having incoming connections.
NoIncomingPath: PathNode7646 is not having incoming connections.
NoIncomingPath: PathNode7653 is not having incoming connections.
NoIncomingPath: PathNode7697 is not having incoming connections.
NoIncomingPath: PathNode7723 is not having incoming connections.
BreakPointsFound: These 12 navigation points aren't that used as valid paths. They might be break points if are a piece of a road.
....
PotentialAITrap: PathNode8082 has at least one path for reaching at it, but zero paths for going out.
TrapsReporting: 1 NavigationPoint actors are considered trap nodes in this map.
...
UnReachableNodes: PathNode7784 looks much over ground for normal Human Size, check it.
UnReachableNodes: PathNode8647 looks much over ground for normal Human Size, check it.
UnReachableNodes: PathNode8658 looks much over ground for normal Human Size, check it.
UnReachableNodes: PathNode8659 looks much over ground for normal Human Size, check it.
UnReachableNodes: PathNode8673 looks much over ground for normal Human Size, check it.
UnReachableNodes: PathNode8685 looks much over ground for normal Human Size, check it.
UnReachableNodes: PathNode8693 looks much over ground for normal Human Size, check it.
UnReachableNodes: PathNode8787 looks much over ground for normal Human Size, check it.
UnReachableNodes: PathNode8789 looks much over ground for normal Human Size, check it.
UnReachableNodes: Here we have 1 points skipped, 9 Navigation Points which are too high and 1730 good ones.

I don't say that I won't try re-converting map back to UT - because I can clearly see useless nodes added in the same place, but I did a checks concerning paths.

Code: Select all

LostData: ________________________________________________________________
LostData: ReachSpecs missing from PrunedPaths[0-15] 	= 31.
FixedWithSuccess: ReachSpecs successfully attached 	= 0.
FixFailures: ReachSpecs attaching failures - probably no place = 31.
LostData: ________________________________________________________________
SpecTest: Testing and fixing specs took: 
	0.178303 seconds.
SpecCheckLocator: Checking ReachSpecs Localization has been finished.

And here...

Code: Select all

XC_Counter: Total ReachSpecs found = 20578

Interesting solutions... Pretty useless...
Optional pointing another "469 logic"... Flight paths over ground

Code: Select all

ProcessStarted: Start Clock...
WalkLinkFound: Connected via ReachSpec 19802 ...
NoFlightNeed: Already connected in reverse PathNode8799 to PathNode6930 with Walking flag, which means that flight is not Needed.
WalkLinkFound: Connected via ReachSpec 5260 ...
NoFlightNeed: Already connected in reverse PathNode7348 to PathNode7746 with Walking flag, which means that flight is not Needed.
WalkLinkFound: Connected via ReachSpec 8423 ...
NoFlightNeed: Already connected in reverse PathNode7796 to PathNode7746 with Walking flag, which means that flight is not Needed.
WalkLinkFound: Connected via ReachSpec 19336 ...
NoFlightNeed: Already connected in reverse PathNode8757 to PathNode7789 with Walking flag, which means that flight is not Needed.
WalkLinkFound: Connected via ReachSpec 8405 ...
NoFlightNeed: Already connected in reverse PathNode7795 to PathNode7828 with Walking flag, which means that flight is not Needed.
WalkLinkFound: Connected via ReachSpec 8426 ...
NoFlightNeed: Already connected in reverse PathNode7796 to PathNode7829 with Walking flag, which means that flight is not Needed.
WalkLinkFound: Connected via ReachSpec 2009 ...
NoFlightNeed: Already connected in reverse PathNode7006 to PathNode8588 with Walking flag, which means that flight is not Needed.
ProcessEnded: Time Taken 1.399661

If we can walk from node B to Node A (no jump needed) route from A to B it's definitely a ground path. If things are not reporting valid routes it might be about some BSP issues which this type of maps are doing from time to time. XC Tester did not say anything about a walk or flight path over there because of something coming from geometry. I'll study options in my testers for preventing false reports but I don't think they are miss-reporting things.

[evilgrins]

OjitroC edited me a version without the vehicles so I could assess bot-pathing in my own way... of watching bots handle the map.

Buggie said the map is intended for players, not bots... and he claims he fixed the visual issues I noted before.

Mostly, though, I was just curious if it was possible to de-469 it.

[Nelsona]

What is GOOD TO KNOW is I could not save map as a copy without anything touched. After DELETING paths map was saved normally and this thing is demonstrating once again that getting over 436 constants is a bad practice - this is what it does those alleged "memory corruptions" - "MAPPER" more exactly. And then ? And then the BullSh!t story about "Players-Only" it's something like fake news - I don't even listen(/read) any lousy and pathetic explanations from those who keep demonstrating that they don't fully understand these devs.

Edit:Allow me a few seconds to show you something...

CurrentState0_Torlan.png (370.77 KiB) Viewed 2481 times

DevPath works here - 440/436. What did I do more exactly ? I used the PathsMaker coded by the same boy who always is recommending an "update" to 469 for no purpose and releasing X versions of the same map which he cannot do well in first editing stage... Hypocrisy now days has a Level which you cannot even imagine...
All network is good using an automated script and returning during a few seconds a normal UT map, limit being set at 985 Nodes and Not 2000 which is twice over 436-451 Engine capacity.

[evilgrins]

Cool.

This forum doesn't have a "like button" but if it did I would've liked that!

[Nelsona]

Perhaps not everything has to be "liked". I removed those top items which common player cannot see and Bot won't get there. For winning a CTF match you need to capture enemy Flag and not escalating to the sky while enemy is attacking your base. But... we edit the edit which is the result of an edited edit of an edited map....
I suppose v105 will get wrapped in big parts or we can wait v208 or whatever...
And then what makes the difference between a "469-only" and an "UT map" at pathing subject ? Let's see...
- In 469 we have an extra-power needed to process more USELESS data - PrunedPaths lists are mainly full filled. Are these even needed ?
- Actually no, if you drop out the extra-charge, Navigation is getting turned back into UT compatibility and... game works normally.
What I found disturbing ?
While some operational code from MapGarbage was "bugged" - the mostly not very optimized, these maps are perhaps veeeeery optimized. I did not know that optimizing things means more useless garbage to be processed, more nodes for moving from this spot 1 feet away and so on. I'm so "impressed". For dealing with such useless debris we need an update which is NOT HAVING XC assets available and it's not a bug-free thing. Fascinating !
The thing is that map MH-GardenOfDeath (the version with normal paths) IT'S NOT working in said 469b - supposed powered up. But the map works normally in 436 + XC assets. And then ? And then to me this looks mainly as a down-date not an up-date - coming with new Editor "reactions".

[Nelsona]

What's the "connected" problem ?
Map has under 1000 Nodes but it goes over 4000 ReachSpecs which maybe I don't like. And then ? And then I touched a bit PathsChecker builder which I'm using for testing paths. This is an environment where I'm really curious to check links from anywhere to everywhere. What's the deal ? The deal means testing if DevPath responds at calls from each Navigation Point for finding routes to every single Navigation Point (item, goal). In certain adjusted maps, this is not possible due to some One-Way routes done on purpose, but in a common map in theory Bot triggered in a sudden moment should be able to find an enemy if this one is somewhere closer to Navigation Network... using optional sub-option _bOmniSeeking which will be there in next builder update.
Q: But... hey... do you have any map where everything is connected everywhere ?
A: Yes, just look at DM-Agony. Map has debris data but still having valid links. I have to admit that DM-Malevolence has flaws.

Q: What flaws ?
A: We cannot find any route to PathNode4... because PathNode4 is not in map and then... I don't think you'll have any enemy over there...

Q: And here ?
A: And here I'm looking for lost nodes that were added on hills and not connected for various reasons - and removing them.

As opinion about how Bots are playing, in 436 we don't have Paths problems, but I see funky mod issues making all this completely useless to me. Bot getting weapon and trying to jump in vehicle is suddenly dying, not seldom, but often. Rendering goes very slow here, there are some holes into building from the middle and... probably more after using it more times for capturing other details... It might not be completely an UT436 map as long as I see that XV it looks the mostly for some 469 version whatever which probably in future will have special hard-coded junks in order to fix the flaws which such a mod is having by default and this way altering other mods based on sudden new "Engine Rules".