Builders helpers for UT Editor

[Nelsona]

First post has been updated with latest updates for this year:
- MapGarbage December 2021
- XC_PathsWorker December 2021

See ya next year... Because... some people really don't understand combos and where a JumpSpot needs to be placed.
Here is how plain default DevPath is causing a bugger ReachSpec... due to mapper's "intelligence".

Stupid_WayBack.PNG (1.01 MiB) Viewed 6748 times

Pawn will really fail to get back on platform if it wants to move there...

And then... at least builder might go more smarter than a walls-drawer... in January 2022.

No_Stupid_WayBack.PNG (1.01 MiB) Viewed 6748 times

[Nelsona]

In January MapGarbage will have a bit of polishing regarding to Nodes without incoming paths. Several dudes were informed about missing paths to InventorySpots. After some time they replied that "I don't have those". They are not really understanding why those things were vanished because for sure they WERE in map before.
Such a report will include WHICH INVENTORY was part of a previous borked InventorySpot, because the newer one created will have the same NO-ROUTE after each build without anything positive, a bad location will stay as it is, doesn't matter how many times are you pushing build button and junk data are just overloaded without to fix anything. Such a report will be like here showing which Inventory is not a goal for A.I.

Code: Select all

NoIncomingPath: PathNode78 is not having incoming connections.
NoIncomingPath: PathNode71 is not having incoming connections.
NoIncomingPath: InventorySpot580 -> HealthVial30 is not having any incoming connection...
NoIncomingPath: InventorySpot150 -> HealthVial62 is not having any incoming connection...
NoIncomingPath: InventorySpot921 -> FlakAmmo8 is not having any incoming connection...
NoIncomingPath: InventorySpot922 -> FlakAmmo10 is not having any incoming connection...
NoIncomingPath: InventorySpot923 -> RocketPack2 is not having any incoming connection...
NoIncomingPath: InventorySpot90 -> RocketPack9 is not having any incoming connection...

[Nelsona]

During examinations to these builders, at a moment I was thinking to develop a test feature concerning whatever CTF map (or MonsterHunt, Assault) in order to gain data if Bot Pawns can navigate map and probably showing broken routes. I studied what's suitable as method (because I don't know what is used in UE1) from these:
https://medium.com/omarelgabrys-blog/pa ... 5a8902eb40
Initially a sort of pre-alpha code was doing something in simple maps but not exactly in complex ones due to missing things (I need more arrays...). I tried to emulate a "FindPathToward" (with native 1000 Paths boundary) but... it needs much more variables and a lot of code polishing - I discarded idea so far but I might solve it in future - maybe. I have to admit that for me is difficult using UScript resources because I'm not that skilled in programing things and facing limitations. But... such a check would answer if map is suitable for Bot games under 2 seconds.

In other hand I didn't see any solution so far for examining vertexes if are well aligned at grid and options for solving them...

[Nelsona]

First Post has been updated with two things: MapGarbage January 2022, XC_PathsWorker January 2022.
Tiny notes:
#1 MapGarbage:
- adding which Inventory has paths missing not only reporting which InventorySpot has no incoming paths because the boys are not understanding what they do;
- adding an option for testing classes concerning if class is Abstract - unusable directly and if this is compiled or it's a lousy wreck inside map.
#2 XC_PathsWorker
- a small checker has been added trying to prevent creating a TWO-WAY jumper combo but heading to an impossible way back - people are not understanding how to setup a jumpy route and preventing lousy return moves, just like that. It's not an offense, it's a reality demonstrated by themselves.

[Nelsona]

First post has been updated with MapGarbage version for May 2022.
Update Info:
- adding bClientNav option for fixing as static stuff navigation actors that are ruined during On-Line gaming in clients - I think this stuff should stay in map without having paths referenced to nowhere, even if we talk about a client game - said points will be set bNoDelete - exactly, in Editor technically editing these "constants" are not only innocent but improving future product released from there;
- rewriting main codes for preventing links at BotPack - checks in certain SP COOP map doesn't need any BotPack and links loading in memory useless assets. Features concerning BotPack seem operational due to solution applied for tracking them.

Another stuff added is XC_PathsWorker version for May 2022.
Update Info:
- rewriting codes for decreasing some iterations;
- rewriting codes for preventing links with BotPack;
- reducing log data during fixing ReachSpecs - it will have only those fixed. Reporting Paths is already a feature available.

[Nelsona]

Update bump:
PathsLinker from first post has been removed. It doesn't worth wasting time with extra explanations.

[Nelsona]

Bumping versions
First post has been updated with XC_PathsWorker for July 2022.
Quoting "what's new" - aiming various test-works during mapping stage:
- added sub-option for fixing ReachSpecs - if used, lost specs are NOT loaded into chain but nullified - cannot be completely deleted with current XC assets;
- added a re-indexing option for remapping ReachSpecs and leaving nullified ones at last positions in their array. Here we can have the image about what paths are done and what is truly calculated in routing process during map's run-time.
- autoconnector based on probing will be able to connect two related combo specific points (LiftCenter-LiftExit) in original style by just linking them without tests. It will test only normal paths.
Details are explained in PDF document included in archive.

[Nelsona]

Next update info.

The future version of August mainly aims at some impediments that come with the new updates of the UT editing application. We have some old problems that are rather moved than solved. In the 469C version that I can't test on my system, I see results that are not too fascinating - there are maps edited there and publicly distributed. The hidden navigation pathways (shortcuts) do not seem arranged at all, they are still overloading without any benefit. We will be able to see them by changing them temporarily into normal paths and making the editor application to show them - putting them in original state it's another option which is activated automatically. Also we can do them Null and use only what is useful to us. The arrangement and improvement operations will operate without impediments up to the 469b version by loading the corresponding XC_Engine version.
Even if we talk about older maps from stock, we can also have an image about how big is the load with paths, because we can see every single path.

Sample - a spot from CTF-Gauntlet. This is what is shown normally as Paths using XC_Engine.

Gauntlet_Normal.PNG (293.59 KiB) Viewed 5066 times

And this is the Real Load...

Gauntlet_RealLoad.PNG (293.69 KiB) Viewed 5066 times

I'm not surprised to see Engine skating in certain cases...

[Nelsona]

For next update session MapGarbage will have something added - related with other junks that are hosted in maps such as Empty groups, more specific groups that are created or copied from elsewhere and having nothing. Yes, I found such works, one of them after a clean-up task was reduced with 3 MB.
What builder does it is reporting/counting how many actors are in certain group specified by checker guy. Groups without any sort of actors are just useless...

[Nelsona]

Next days after updating docs concerning information for builders and update notes, I will post them and... probably something new...

PathsChecker.PNG (215.26 KiB) Viewed 4942 times

Yup... I called it PathsChecker but I can use another name if one more suitable will be recommended...
Purpose of builder is preventing mapper/editor (me and/or others) from spectating Bots during more minutes just for figuring if they want or not to move at items/actors based on Navigation Paths which map is having. All testing is needed in a game session with regard to jumping with JumpBoots or ImpactHammer (these need to be tested "LIVE").
Builder is launching a navigation test. We are testing logic engine routes from every single PlayerStart to all sort of Actors (even if map doesn't use direct paths from PlayerStarts - lol):
- Inventories;
- FlagBases;
- AlternatePaths;
- DefensePoints;
- ControlPoints;
- MonsterWaypoints (exactly... you read well);
- MonsterEnd.
Here we will know in advance if what we have are valid paths or just dummy stories. Builder works enough accurate, Editor environment is not like game environment but is still conclusive in this regard.
As a technical detail, Builder DOESN'T NEED any BotPack or MonsterHunt packages installed in Editor - not everyone wants games installed after all.