[MapScript] Erebus Continent

[Milaga]

You know... I was playing around with some changes to the normilization code so I may have broken something. It should be clearing out roughly half of the forest and/or jungles in a 7x7 area centered on your starting plot. I will also need to make some adjustments to forest placement when the smart climate kicks in, I keep forgetting to do that.

I look into that as soon as I have a chance, today and tomorrow are busy days for me at work so I probably won't get anything done before this weekend.

I'm not sure that's it. Or at least, that's not all of what is going on. I finally met the Ljosolfar and their start was in Floodplains. All the other civs seemed to start in areas with lots of forests EXCEPT the Ljosolfar, the ones you'd expect to be in the forest. And it's not as though all the forest starts are taken. It looks like most of the tiles that aren't desert have forests on them. I increased the number of peaks but left the values for all the other tiles alone.

 

[Seven05]

I'm not sure that's it. Or at least, that's not all of what is going on. I finally met the Ljosolfar and their start was in Floodplains. All the other civs seemed to start in areas with lots of forests EXCEPT the Ljosolfar, the ones you'd expect to be in the forest. And it's not as though all the forest starts are taken. It looks like most of the tiles that aren't desert have forests on them. I increased the number of peaks but left the values for all the other tiles alone.

Were you using FlavorMod or Flavor Starts in RiFE or Wild Mana? I can't duplicate the effect at all here, forests/jungles are being cleared fine and nobody is ending up horribly out of place.

If you can enable logging and see what it says while placing the civs that may help.

I've also uploaded 2.56 which has a number of changes including reduced forest density.

 

[[to_xp]Gekko]

care to explain what was changed in 2.56 aside from forests?

 

[[to_xp]Gekko]

nice work on the new version, although the handling of unique features may still be buggy ( all unique features on, the ring of carcer was nowhere to be seen) . did you increase deserts btw? compared with 2.55 at identical settings I'm getting too much deserts.

 

[Seven05]

care to explain what was changed in 2.56 aside from forests?

It's classified!

Ok, not really. I changed the way forest chance is calculated so they're not as dense overall unless you go with the increased forest setting. I added support for RiFE's upcomming 'Flavor Patch' which includes 'deep oceans' and a new 'reef' features as well as alternate ancient ruins to address the sanctify issue. I added a noise component to the tectonic plates to help randomize the mountain ranges and coastlines a little more and I fixed an unwanted side effect of the detail map overlay caused by it not being shrunk properly when the rest of the map was. I also moved the kelp, reef, ancient forest and haunted forest code so they are all covered by the 'use advanced features' option.

For player starts I tuned up the MP version of the Doviello and Illians flavors to see if they're more playable in MP games.

I didn't do anything with deserts or any other terrain settings directly or any of the climate code (aside from the forest chance changes) however with the different coastlines and mountain ranges it's entirely possible that you'll get more defined desert regions. When you have a more defined desert region fewer of the desert tiles are removed during the clean-up process that turns desert tiles next to grassland into plains. On my large sized test maps the overall desert percentage is mostly unchanged but I can see that having an impact on smaller maps and low cohesion maps with the smart climate enabled.

I'll check the city ruin placement again, I'm pretty sure there is nowhere left where it can remove existing improvements EXCEPT city ruins and I ran several tests with all unique features on without any missing.

 

[[to_xp]Gekko]

thanx for the info and for the constant improvements

 

[Valkrionn]

care to explain what was changed in 2.56 aside from forests?

Support for several new features in RifE: Reefs, Deep Ocean, Ancient Ruins (city ruin replacement for the map option, no sanctifying it).

 

[Valkrionn]

Here's a screenshot of the new ocean depths, Coast, Ocean, and Deep Ocean:

 

[Seven05]

And the asteroids... erm, I mean REEF feature

 

[Valkrionn]

Yeah, that's a placeholder.

 

[Milaga]

Were you using FlavorMod or Flavor Starts in RiFE or Wild Mana? I can't duplicate the effect at all here, forests/jungles are being cleared fine and nobody is ending up horribly out of place.

I was using the Magister Mod in FfH vanilla, so no Flavor at all.

I really don't know what caused the bad start for the Ljosolfar, but I can't replicate it either. It's possible that, since they started close the the Balseraphs, they moved into their spot before founding a city. The Ljosolfar then were bumped north, so they picked a plot with a bunch of floodplains.

Spoiler Snippets of the log :

Random seed (Using Python rands) for this map is 807906779824036
peak count = 1
64,112 = 1.0

Evaluating continent percent: -0.252005
Massive Continent, reducing desert and plains
Distributing rainfall
Finished generating plot types.
Adding Terrain
Finished generating terrain types.
Adding Rivers
Adding Lakes
Adding Features

Trying to place 1 of BONUS_DEER
Successfully placed 1 of BONUS_DEER
Trying to place 1 of BONUS_BANANA
Could not place 1 of BONUS_BANANA
Trying to place 14 of BONUS_COW
Could not place 9 of BONUS_COW
Trying to place 4 of BONUS_GEMS
Could not place 4 of BONUS_GEMS
Trying to place 8 of BONUS_SUGAR
Could not place 8 of BONUS_SUGAR
Trying to place 5 of BONUS_DYE
Could not place 5 of BONUS_DYE
No room at all found for BONUS_GEMS!!!
No room at all found for BONUS_SUGAR!!!
No room at all found for BONUS_BANANA!!!
No room at all found for BONUS_DYE!!!
No room at all found for BONUS_SHEUT_STONE!!!
No room at all found for BONUS_TOAD!!!
number of starting areas is 1
1153 plots available for 9 starting plots
8 tile seperation grain
Player 0 starting plot 22 value is 86
1023 plots remaining
Player 6 starting plot 11 value is 36
823 plots remaining
Player 1 starting plot 17 value is 36
736 plots remaining
Player 5 starting plot 76 value is 36
579 plots remaining
Player 8 starting plot 95 value is 36
420 plots remaining
Player 3 starting plot 194 value is 56
340 plots remaining
Player 4 starting plot 267 value is 36
196 plots remaining
Player 2 starting plot 498 value is 36
143 plots remaining
Player 7 starting plot 93 value is 32
130 plots remaining
9 plots remaining after clean-up
Removing peaks and hills ...
Player0 preference found- Jungle Val: -2 Forest Val: 2 Hill Val: 0
Player1 no preference found, using defaults
Player2 no preference found, using defaults
Player3 preference found- Jungle Val: -2 Forest Val: 0 Hill Val: 0
Player4 no preference found, using defaults
Player5 no preference found, using defaults
Player6 no preference found, using defaults
Player7 no preference found, using defaults
Player8 no preference found, using defaults
Removing bad features ...

In these maps there are still a lot of forests. I had a few starts that had all of my plots covered in forests. The log mentions removing peaks and hills but no mention of the removal of forests.

There is no jungle on the map (hence no jungle resources.) Again, I was using Smart climate: modify. If all eight civs have no preference for Jungle, will it be completely eliminated?

Maybe don't need to worry too much about this. If I see anything funny with the new version I'll bring more logs.

 

[Seven05]

Thanks, later tonight I can test with the exact random seed to

The line that reads: Removing Bad Features... is referring to forests and jungles. I'll have to check the civs in Magistar's mod since apparently most of them are using the default values (or, not supported by me). Did he add new civs? Or did you just pick all of the civs that don't need anything specific? Although I don't think I've left 7 of the original 18 out as most have at least one or two values defined. Thank god for Smart Climate then, with that many civs fighting over grassland they would have had some horrible starts and I see that none of them even needed a second attempt at finding a start plot.

It looks like you're using the 'Modify Selection' smart climate option and with no jungle civs it's dropped it by 2, this means it's the same as selecting 'Low' from the map options. If you were using the 'Override Selection' option it turned them off completely. Depending on the map it's entirely possible that there is no room for jungles at that setting, obviously you didn't get any as it was unable to place any of the jungle resources like bananas. If you want to ensure that you get at least some jungles even if you have no jungle civs (same applies for deserts and tundra) you can try setting the map option for them to 'Increased' and use the 'Modify Selection' option, although you need to be prepared for massive jungles/deserts/tundra if you end up with a few of those civs in the game. I have some code to add to the smart climate to account for maps without enough room for jungles if they aren't turned off entirely but I'm still testing it as it's a bit more complicated than it was for deserts and tundra.

I'm not familiar with Magistar's mod at all so I don't know if he has anything in place that could prevent the normilization from working. It's possible that he did something similar to Flavor Mod and disabled calls to the normilization callbacks, if so it's ignoring that step in the mapscript entirely.

And yes, please keep me posted if you see anything funny with the new version. Also, if you can, let me know what Civs are in MagisterMod that I don't have flavor values for (and feel free to suggest values for them) so I can add propper support for them.

 

[Milaga]

Did he add new civs? Or did you just pick all of the civs that don't need anything specific?

You know, I didn't realize until a while after I posted that I don't really know everything that Magister's mod has changed. I picked the same 9 civs for the original report, Bannor, Elohim, Ljosolfar (+2 forest, -2 jungle), Sidar (-2 jungle, I think), Grigori, Amurites, Sheiam, Calabim, Balseraphs.

No major changes in the MC Mod. Just a lot of really awesome tweaks.

Thank god for Smart Climate then, with that many civs fighting over grassland they would have had some horrible starts and I see that none of them even needed a second attempt at finding a start plot.

Yeah, these settings produce a lot of grassland and very little plains and tundra. There still is a fairly large desert in the middle of the maps, though. Smart Climate looks like it makes for some balanced, if not boring, maps in vanilla FfH.

I'm not familiar with Magistar's mod at all so I don't know if he has anything in place that could prevent the normilization from working. It's possible that he did something similar to Flavor Mod and disabled calls to the normilization callbacks, if so it's ignoring that step in the mapscript entirely.

Only he can say for sure. Come to think of it, there are a whole bunch of new Unique Features. Their placement definitely could be interfering.

 

[Seven05]

You know, I didn't realize until a while after I posted that I don't really know everything that Magister's mod has changed. I picked the same 9 civs for the original report, Bannor, Elohim, Ljosolfar (+2 forest, -2 jungle), Sidar (-2 jungle, I think), Grigori, Amurites, Sheiam, Calabim, Balseraphs.

Yep, no flavor settings for most of them, the only difference with the Sidar is a negative neighbor value

Yeah, these settings produce a lot of grassland and very little plains and tundra. There still is a fairly large desert in the middle of the maps, though. Smart Climate looks like it makes for some balanced, if not boring, maps in vanilla FfH.

Yeah, it's trying to compensate for extreme cases, like 9 civs that all want grassland starts

When you have a more balanced civ line-up it's not so extreme, so if you have the Malakim and either the Doviello or Illians in there you don't see as much of the world tuning into a grassland paradise. With FF, RiFE and Wild Mana it's easier to get civs that like the jungle, desert or tundra so it's not so severe. For example if you end up with 3 desert civs, 3 tundra civs and 3 'default' civs you'd get a fairly even mix of grassland, desert and tundra.

Only he can say for sure. Come to think of it, there are a whole bunch of new Unique Features. Their placement definitely could be interfering.

The unique features shouldn't be having any significant effect. My normilization code skips tiles with those on them but there would have to be one on almost every tile to make that big of a difference. However, if he was working with FF, RiFE, Wild Mana or Flavour Mod for his base DLL he'll have the Flavour Mod code in there which disables the normilization. It could also happen if he made any changes in the DLL to the map normilization callbacks in an attempt to speed up the game by removing them.

Unfortunately, while the mod can detect the mapscript and make adjustments as needed I'm not sure that I can do the opposite. And even if I could, I don't think I'd be able to do it in a way that didn't break feature in other mods. It's really wierd though as I can't see why it should be happening in the first place.

Since you have logging enabled, next time you generate a map see if you get any errors in PythonErr.log (ignore the PythonErr2.log), that records all of the python exceptions (PythonErr2.log is full of mostly junk, not the exceptions we need).

 

[Seven05]

Ok, I have some changes lined up for the next version, I just want to make sure I haven't forgotten soemthing since last week was a little busy here.

I have added jungles to the basic smart climate code where it will compensate for maps with little or not southern land by decreasing the minimum temperature needed for jungles. This is similar to the existing effects where it will reduce deserts on massive continents and reduce tundra when you have a lot of land in the northern half of the map.

I've also made some adjustments to the way mp games are handled. When it detects a multiplayer game the souther edge of the snow/tundra region has some random noise added to the temperature map, this will cause some extra plains & grasslands to be placed mixed in with the tundra. The effect is stronger on riverside tiles so the starts for the Doviello and Illians are more likely to have some riverside grasslands available for early farms.

The smart climate options have been tuned up a bit. In particular, using the 'modify selection' option is less likely to turn jungles or tundra off if left at default and no civs are in the game that prefer those terrains.

If it looks pretty good with the latest changes I have a second map script to upload that will offer some of the map types that people have asked for but I wasn't able to add to this script. I hate to have two different map scripts up, but it will keep the number of map options to pick through down to something reasonable.

 

[[to_xp]Gekko]

sounds good, although I think that the tweaks you are reserving for MP would be good to have in SP as well. Flavour can go to hell, balanced starts are way more important for an enjoyable game

 

[Seven05]

Maybe I can make it a map option... I hate to add more but I should be able to merge it with the single tile unit placement option.

 

[[to_xp]Gekko]

that would be great

 

[Milaga]

EDIT: Uh, nevermind. I reread the first page and it doesn't look like this works with Orbis.

 

[Seven05]

Well, it should work with Orbis but I haven't added anything specifically for Orbis. If somebody wants to let me know what the new civs are and their 'flavor' I can get them added very easily. If you use it now they'll still be placed, just using the default 'I love grasslands and hate deserts, jungles and trees' rule.

And, 2.57 is being uploaded, it'll take a few minutes as I have to re-write a bunch of stuff on the first page. For those of you following along, here's the changes so you don't have to read the python file for the change log

- Added correctional adjustment for jungles if there is not enough land in the south. Since jungle use rain AND temperature I only adjust temperature here so it will only reduce the minimum temperature required if there is too few land tiles in the south.

- Added y-wrap awareness to the smart climate. Sounds fancy, right? Basically it means that the smart climate will not adjust tundra, snow and jungle temperatures if y wrap is enabled.

- Removed all flavor related code to detect multiplayer games (see below)

- Removed unit placement map option (see below)

- Added new option to control starting plot rules. The three basic choices are full flavor, reduced rules and minimal flavor, you can all choose either of those with the single tile unit placement. So you get six choices but it's really only three. Reduced flavor changes the flavor settings for the tundra based civs, minimal flavor further changes the tundra based civs and adds in the ability for normilization to convert marshes, tundra and snow to grassland near player starts. This means you can essentially turn off flavor starts, even in SP games, however some preferences will persist so the elves won't clear trees, the lanun will start on a coast and the malakim will start in deserts.

I'll post a note when I'm done updating the first post.