Evil Genius at work

[RulerofRails]

Sorry for the mis-info about the cgo weights. Didn't review them, just went off memory. Gumboots, thanks for the correction. The correct info was in the post I linked.

Jim, what you said about events never being totally deleted is true. Never heard of someone hexing around with them. Unless there's a way for that my understanding (may be WRONG): Better to re-use than delete. Moving events should be avoided as well. Just leave the "retired" ones in a deactivated state (for example no conditions).


With the regular March and September economic events, your method will work. But, FWIW here's a way I might use
Test at the beginning of the month for
(Whatever condition to limit the event to be active during only part of the game by a certain value of an activation variable or simply the game year test.)
Current Game Month = 3
OR Current Game Month = 9
Economic State != 0

Then same effect:
Set Economic Status to 0.


If you have an event that tests weekly you might tell the week from the count of times processed on the event debugging page. Some maps have used events that save the week to a variable for a finer control over when events will fire. I.e. you can choose to check the event on one specific week by specifying a particular value of that variable, instead of just a constant check each week.

The timer on issuing stock is tied to a finer measurement of time than weekly. The status event updates quite frequently. I never bothered to work it out. Milo (guy who made the 1.06 patch) talks about a "scenario clock".

PS.
In typical RT3 style, the editor has plenty of quirks. Check out this trick. Those that have been discovered are scattered at various points around the forum. Most of the quirks can be worked-around, at least for a partial achievement of the desired effect. Though there are still plenty of things that haven't been tried, at least if they were, no report was given on their bugginess or lack thereof.

[Gumboots]

Not that I'm in a rush to edit my way through 7,800,000+ hex values to trim out useless bytes, but if a man were bored enough (and careful enough) is there a guide or note that documents what is safe to strip out? In this case, I would strip Go West!.gmp down to its base and start over.

I think you are boldly going where no lunatic has gone before.

On the bright side, there does appear to be a pretty high limit before events cause problems. I've never hit it, and have done quite a bit of playing around with events when testing.

There is a guide to where the events section of the .gmp begins and ends (have you got the folder named Milo's Stuff*?) so it should be possible to strip the entire events section in one fell swoop. That wouldn't be too arduous. I expect that manually picking out individual events would be trickier. Having said that, I also expect there would be some consistent syntax for starting and finishing the bytes for any event, and once that was known it should be usable.

If you want to look into this, I'd start with the most bog basic blank map available in the editor and save that sucker, then add one or two events to it and do a comparison on the hex.

*Here's what Milo's Stuff says about events:

Code: Select all

------------------------------------------------------------------------
Events:
  Event block header
  Events
  Conditions
  Effects
  Event dependencies:  cargos, trains, etc

000468a0:00066940?  EVENT BLOCK HEADER: several 4-byte counts or flags
000468a0:000468a3     number of events
000468c8:000468cb     number of events?

EVENTS:  2199 bytes each
0004695a:0004695d   00 00 00 00 for no cond
                    40 0d f8 05 for ALWAYS TRUE
                    10 87 78 03 for ALWAYS TRUE
                    d8 99 70 03 for ALWAYS FALSE and constructed
0004696e:00046971   event index, starting with 1
00046972:00046990   event name, null-terminated
00046a3a:00046d55   dialog 1, null-terminated
00046d5b:00046e23   dialog 2, null-terminated
00047148:00047148   0 = dialog, 1 = choice, 2 = newspaper, 3 = game msg
00047149:00047149   event frequency:
                       0  yearly (beginning of year)
                       1  yearly (end of year)
                       2  monthly (beginning of month)
                       3  monthly (end of month)
                       4  weekly (beginning of week)
                       5  weekly (end of week)
                       6  when track or station is placed
                       7  when a company is started
                       8  at start of scenario
                       9  status (in annual report/ledger)
                       a  after an event choice
0004714a:0004714a   test against territory:  0=all, #=territory, 63=don't test
0004714e:0004714e   0 = single-player, 1 = multiplayer, 2 = both
0004714f:0004714f   0 = repeating event, 1 = one-time event
00047153:00047153   test against:
                       0  don't test against companies  
                       1  all companies
                       2  on-screen player's company only
                       3  AI companies only
00047154:00047154   test against:
                       0  don't test against players
                       1  all players
                       2  on-screen player only
                       3  AI players only
00047155:00047155   apply effects:
                       0  to the company\player\terr for which cond is true
                       1  to the whole game
                       2  to all companies
		       3  to human companies only
		       4  to AI companies only
		       5  to territories
		       6  to all players
		       7  to human players only
		       8  to AI players only
00047159:00047159   0 = apply effects repeatedly; 1 = apply once per event
00047169:0004716c   index of territory to apply effect to (0 = all terrs)

00047209            next event name

So that doesn't look too bad. As long as you have a record of all past and present event names, it should be possible to remove any of them.

Edit: So I took a look in one of my .gmp's and Milo's offsets don't correspond to my file. Just for added fun. I suspect they correspond to whatever test map he was using at the time, rather than being intended as a general purpose roadmap.

However I have managed to track down where the same bits are in my file. I'll look into mapping it a bit more later, but it doesn't seem like it's going to be too much drama.

For example, his reference 0004695a:0004695d corresponds to 0024c88a:0024c88d in my file.

[Gumboots]

I was just looking through what MIlo wrote, and found this in his NOTES.txt file.

5c8de0 is the first of a long built-in list of buildings. I smell duct tape.
Referenced at 41029F and 41EC17, both times passed to 5A57CF, but swapping
places with the other operand. More evidence that 5A57CF is strcmp.
5c8df0 is the second in the list, referenced from 41027F and 421118... tell
me they aren't strcmping everything. Oh, god, this is awful, they're doing it
twice.

EDIT: This is interesting too. Apparently Milo extended 1.06's upper limit of locomotives by carving chunks out of the cargo limits and re-purposing them. This implies that if somebody wanted a greater range of cargoes, and was happy with being limited to 100 locomotives, it would be possible to create a hybrid 1.05/1.06 .exe that had up to 61 cargoes (as opposed to 1.05's 41 cargoes and 1.06's 51 cargoes).

Cargo types are limited to 61 on the conditions page and 71 on the effects
page. That almost certainly means I can carve the last 10 effects off and turn
them into loco-avail effects. I also have the empty space from 457-502,
which I can use if I can add an if statement to eval code. Should add
5+1+23 = space for 29 new locos, with 1 if statement required to read
each effect.

[Just Crazy Jim]

Well, I noticed this:

Even though the file was never saved with Alcohol as a port demand, there are traces of it in every instance of a port demand/supply/exchange. This implies that the editor holds the entire GMP file in RAM and edits data on the fly. It also implies that cargo names in the ports/warehouses section are at specified points in the file as nul terminated values and the game & editor read to the nul value, then skips to the next fixed position. All values between the first instance of a nul and the next fixed point are ignored.

I also found that Milo's addresses did not correspond to the position of things in my GMP.

Beyond what PJay and Milo have in their note txts, what we know about GMP files is this:
There is an embedded JFIF file for display in the selection screen in the GMP
There is a section with height values for the map in the GMP
There is a section with rivers/lakes/ocean and associated exclusion areas in the GMP
There is a layer of the map that contains exclusion areas for cities in the GMP
There is a section for trees painted to the map in the GMP
There is a section for the terrain coloring painted to the map in the GMP

The lossy save fix for GMP files (cut & paste) includes data from some or all of them.

IIRC, in 2003, when RT3 and The RT3 Editor were made, 1 GB of system RAM (even 2 gigabytes of RAM) in home computers wasn't uncommon. 1 and 2 GB video cards were on the market, though not really priced for the average home user. 1024x768 was still king in the monitor market, but 1280x1024 was "the next big thing". I remember because my son came back from Afghanistan in mid-2003 and we built him a system that was "top of the line" for that half-hour and the first game we dropped onto the hard drive was RT3.

So, I am thinking that the PopTop devs did what every software dev did as the system resources grew beyond anything expected, they did nothing to prevent bloating and wrote sloppier code. Even Linux distros, famed for their slim installation, grew from single CD to multiple CD, then DVD, installs about that time. But that runs the opposite of what we know about the GMP lossy save. Why would they code it to be a bloated hog in part of the code and a disk space miser in another section?

[Gumboots]

Lack of time and money? There are some things that were obviously rushed in the game.

And it's possible the lossy save was simply a mistake. They might have had a handy wodge of code and just used it, not thinking it would cause any problems in practice.

Edit: This makes sense...

It also implies that cargo names in the ports/warehouses section are at specified points in the file as nul terminated values and the game & editor read to the nul value, then skips to the next fixed position. All values between the first instance of a nul and the next fixed point are ignored.

That's the usual way the game handles data in files. You can throw comments into .lco and .car files (and I often do) as long as you can fit them between the end of your actual data and the next fixed point.

[Cash on Wheels]

I was just looking through what MIlo wrote, and found this in his NOTES.txt file.

5c8de0 is the first of a long built-in list of buildings. I smell duct tape.
Referenced at 41029F and 41EC17, both times passed to 5A57CF, but swapping
places with the other operand. More evidence that 5A57CF is strcmp.
5c8df0 is the second in the list, referenced from 41027F and 421118... tell
me they aren't strcmping everything. Oh, god, this is awful, they're doing it
twice.

EDIT: This is interesting too. Apparently Milo extended 1.06's upper limit of locomotives by carving chunks out of the cargo limits and re-purposing them. This implies that if somebody wanted a greater range of cargoes, and was happy with being limited to 100 locomotives, it would be possible to create a hybrid 1.05/1.06 .exe that had up to 61 cargoes (as opposed to 1.05's 41 cargoes and 1.06's 51 cargoes).

Cargo types are limited to 61 on the conditions page and 71 on the effects
page. That almost certainly means I can carve the last 10 effects off and turn
them into loco-avail effects. I also have the empty space from 457-502,
which I can use if I can add an if statement to eval code. Should add
5+1+23 = space for 29 new locos, with 1 if statement required to read
each effect.

IIRC Milo said the hard limit for cargos in a economic cell was 52.5 cargos. Trainmaster has 52 cargos & the economic mod that I have shelved has 52. He tried everything to increase the number of cargoes.
So thats why he replaced the redundant cargo space with locos.
Old erased bits from the .GMP will be overwritten with time.

[Gumboots]

Ok, that makes sense.

[Just Crazy Jim]

52.5 Cargoes. That's still a lot of wiggle room within an economic model.

You could set things up so that as time passes, some cargoes become obsolete, as others become available. RT3 'nilla already does half of that, the adding new cargoes part (e.g., oil, rubber, uranium). Technically, it was impractical to ship live cattle until much later than it occurs in RT3. Oh, they tried, but the rate of loss due to stress was enough to make it financially unattractive. The Chisolm Trail marks the breakwater moment when shipping cattle by rail became economically rewarding. See this wiki.

It seems to me that the reverse would also be true. The way I see it, some cargoes simply become unprofitable to haul as the associated industry become superfluous. Take Meat Packing Plants as an example. Early game there may be a few Meat Packing Plants and the livestock may be far enough away from them to make a profit hauling them. But unless the map is specifically set to maintain that distance, more Meat Packing Plants appear and more Cattle Ranches, eventually the two reach an equilibrium and hauling ceases to be profitable. So, in this example, why not retire Livestock and Meat as cargoes at some point after year 2000 and replace them with a new cargo pair, such as rare earth metals and advanced electronics. (-2 cargoes from economic cell, +2 cargoes to economic cell, net sum zero).

Alternately, you could have a cargo change name, iron ore becomes "rare earth metals", the tool & die building becomes "Prison Labor Sweat Shop", and the finished goods could remain goods or be changed to any smart-sounding name you like.

Back on point: Having more than 52 cargoes doesn't mean you have to have every single one enabled on a map. Most of the scenarios and campaign maps have cargoes throttled to less than the full list of 'nilla RT3. Document the limit in the read_me. If the end user breaks the 52.5 cargoes per economic cell rule, it's not on you.

[Gumboots]

There's only one problem with your clever plan. You can't put an end date on industries. The game will only handle giving them a start date. Once you start making something, it keeps getting made until the Sun blows up.

Now you can do heavily customised scenarios, with special industry packs to go with them, if you want to. That way you could choose which industries you wanted in which era, up to a point. However, you can't have an industry ending when you cross over to another era during gameplay. That one is a no-go.

[RulerofRails]

I never tried an end date on industries (but trust Gumboots info), but in practical terms it would be useless because I'm sure that even if they stopped being seeded, the existing ones wouldn't be deleted off your map.

AFAIK, there is no end date on cty files for cargoes either. However, conversions in the bca file can have an end date. Alternatively, a new conversion may begin (start date set to the time of the switch). Trainmaster has Forts in the 19th century which afterwards are "converted" to tourist attractions, new conversions begin when the old ones end.

A simple workaround method to achieve this effect (without any changes to the files) would be to use an event (-99% on Livestock production) so Cattle Ranches stopped producing Livestock. For gameplay purposes the industry will then be dead. Not a pretty solution but it works. But, I agree absolutely, nobody found a way to "swap" cargo for a different cargo in the middle of a scenario. Only possibilities are hacking the gmp (See 1.06 release notes, quoted here, but CoW wasn't really trying this hack) or actual changes to the files.

I also agree that 52 cargoes is plenty of room for one scenario. With an industry pack as Gumboots suggested (at least custom bca and cty files), there's a lot of possibilities. Unless your scenario is very long (40+years), you will likely have many cargoes that can be re-purposed (especially in the 19th century when Uranium, Plastic, Aluminum, etc. are unheard of.

[Gumboots]

A simple workaround method to achieve this effect (without any changes to the files) would be to use an event (-99% on Livestock production) so Cattle Ranches stopped producing Livestock. For gameplay purposes the industry will then be dead. Not a pretty solution but it works.

I hadn't thought of that, but I can see it being useful sometimes. If an industry consistently makes a loss it will eventually be demolished by the game engine, and more of that type are unlikely to seed. That is a good analogy for how things happen in real life. The only catch is it would still be there on the industry list long after you wanted it, but in RT3 you can't have everything.

[Just Crazy Jim]

A simple workaround method to achieve this effect (without any changes to the files) would be to use an event (-99% on Livestock production) so Cattle Ranches stopped producing Livestock. For gameplay purposes the industry will then be dead. Not a pretty solution but it works.

I hadn't thought of that, but I can see it being useful sometimes. If an industry consistently makes a loss it will eventually be demolished by the game engine, and more of that type are unlikely to seed. That is a good analogy for how things happen in real life. The only catch is it would still be there on the industry list long after you wanted it, but in RT3 you can't have everything.

I have tried reducing base cargo production to nil before to see if the game engine would be delete unprofitable businesses from the map. Oh, I works, kind of, sort of. But doesn't stop new buildings for the dead industry from being seeded to the map. If anything, the game engine seems hypervigilant about maintaining the golden number of industries specified by the seeding formula for years from start and by percentage in the region setup. That's what brought me back here to post my latest Mad Genius Idea™ (patent pending).

A simple workaround method to achieve this effect (without any changes to the files) would be to use an event (-99% on Livestock production) so Cattle Ranches stopped producing Livestock. For gameplay purposes the industry will then be dead. Not a pretty solution but it works. But, I agree absolutely, nobody found a way to "swap" cargo for a different cargo in the middle of a scenario. Only possibilities are hacking the gmp (See 1.06 release notes, quoted here, but CoW wasn't really trying this hack) or actual changes to the files.

I also agree that 52 cargoes is plenty of room for one scenario. With an industry pack as Gumboots suggested (at least custom bca and cty files), there's a lot of possibilities. Unless your scenario is very long (40+years), you will likely have many cargoes that can be re-purposed (especially in the 19th century when Uranium, Plastic, Aluminum, etc. are unheard of.

That is worthy of a gold star for better idea. Hacking the GMP and replacing cargoes in the 2nd stage buildings sounds scads easier than decompiling and recompiling the EXE 400 times to get my idea to work. :D

[Cash on Wheels]

That is worthy of a gold star for better idea. Hacking the GMP and replacing cargoes in the 2nd stage buildings sounds scads easier than decompiling and recompiling the EXE 400 times to get my idea to work. :D

The 1.06 cargo types folder can only handle 52 cargo.cty files in the cargo types folder before crashing to desktop. Now I've stuffed up to 62+ ~1234cargos.cty in the 1.05 folder without any trouble. However the warehouse list only shows the first 52. If you try to use a cargo NOT in the warehouse list the game will C to D. Move Cargos around is alot of trouble.

[RulerofRails]

I never actually tried to see if a simple replacement of cargo names in the GMP file would do anything. I did try right now, with no luck. It seems that reference to a cargo type is only made one time (assuming there are no pre-placed buildings/ports supplying/producing that cargo). Which seems a dead-end, unless further discoveries are made.

But, there is a different way.

This is just the 1.06 Ingot cargo. I renamed the cargo to "Fish" by renaming "Ingots" in the Data\Language RT3.lng file. I changed the icon by unpacking the 106cargo pk4 from Data\PopTopExtraContent, picking out the files associated with Fish (fish_A.dds and ~4478Fish.imb) and simply renamed them as if they were Ingots (the files the game is calling): ingots_A.dds and ~4488Ingots.imb. Then I placed those two files loose into Data\PopTopExtraContent. (I'm not sure if the loose file over-write will work for all names).

I only tested this briefly, but seems to work except the Fish Icon didn't appear on the side of the boxcar. That may be a limitation.

PS.
If you want a Fish demand automatically at houses, Medicine might be a better cargo to hack (no particular reason I used Ingots). (IIRC, 1.06 with all cargoes enabled uses up all the slots available in the house that will display correctly, more will work just not display right.) If the price ends up being too high, reducing with an event is safe. It ships in a reefer too, which is probably the most sensible. Would need to hack the cty file to get it available in 1800.

[Gumboots]

Quite a few of the 1.06 cargoes don't call cargo icons or cargo models properly. I'd say your missing fish icon is a 1.06 bug.

However, you could make a custom reefer just for fish, and put all sorts of fishy stuff on the sides of it. If you wanted to.

[Just Crazy Jim]

So, what I'm picking up here is that it's a wee ickle bit more complicated than "editing the GMP".

Me? I just want to have ~4478Fish as a port/warehouse only item that is consumed by housing (and - IF I get it working - a retail building). End of story.

If I'm getting this lined up right, then to achieve that and only that, I need:

1. A CTY file (using reefer as the consist type)
2. A CAR file
3. A PK4 with the custom wagon files (3d models, textures, etc.)
4. Hack House.bca to add a demand for ~4478Fish (replacing something)
5. If I'm feeling froggy, repurpose the Customs House model as a Fish Monger's Shop (retail - although it should probably want Ice also. But first get ~4478Fish up and running.)
6. surely there is more, it can't be that simple

Truth is, we could hack fruit orchards to produce meat or salt or uranium, the wiggly part is getting a demand going so that the cargo goes onto the rails while not causing the Bong of Doom because the hacker (that would be me) gobbed up the files or exceeded the 52 cargo limit of 1.06 - 61(ish) in 1.05

My bent is for the early 19th century. So electronics, uranium, and a sluice of other 20th century cargoes are useless for my period of choice. Fact is, I could just rip Uranium out of the list of cargoes and replace it with ~4478Fish and replace the Nuclear Power Plant with a Fish Monger's Shop and the Uranium wagon with a Fish Reefer - in all, greatly more useful items than nuclear waste generation :D

[RulerofRails]

I never tried adding cargoes in 1.05, but I think you have the idea. Of course there's no right way. I would recommend that you consider how easy it is for a player to revert to the original. It will always be a bit fiddly, but simpler is best.

You cannot have a cargo produced only at a Port/Warehouse. A cargo will only be available if it is supplied by an active building (enabled on the industry list). But, I'm sure you don't want the player to build such an industry (at 0% in all regions/cities it should never be seeded). The good news is: there's a loophole. The building can have a start date far in the future, consequently being enabled on the industry list will not make it available for the player to build.

[Just Crazy Jim]

I never tried adding cargoes in 1.05, but I think you have the idea. Of course there's no right way. I would recommend that you consider how easy it is for a player to revert to the original. It will always be a bit fiddly, but simpler is best.

You cannot have a cargo produced only at a Port/Warehouse. A cargo will only be available if it is supplied by an active building (enabled on the industry list). But, I'm sure you don't want the player to build such an industry (at 0% in all regions/cities it should never be seeded). The good news is: there's a loophole. The building can have a start date far in the future, consequently being enabled on the industry list will not make it available for the player to build.

Thank you for that.

So, if I have it by the rights:

1. Make CTY file for ~####Cargo
2. Make CAR file for ~####Cargo, set start date to 1800
3. Make a building that produces ~####Cargo, set start date to some date outside reasonable expectation
4. Crash game 20 times, cussing optional
5. Come back to forum and bellyache
6. Make it work
7. Be happy for 15 minutes
8. Decide to make more stuff for game
9. Goto 1

[Gumboots]

That sounds about right, except that I would say cussing is essential. Or at least is bound to happen.

[Just Crazy Jim]

In the direction of my locomotive project, the first stage of the Grasshopper, the placeholder - get some feel for scale and ideas for how to proceed for the tender. The HopA model is a wee bit small....