[New release] Simutrans-Experimental 7.0

Topic: [New release] Simutrans-Experimental 7.0 (Read 8943 times) previous topic - next topic

[New release] Simutrans-Experimental 7.0

November 07, 2009, 10:44:33 pm

It has been a long time in the coming - Simutrans-Experimental 7.0 is here, with some major enhancements (based on previous feedback) and bug fixes. See here for more information on how to download, install and set up the latest version. Note that this version comes with new configuration files and a new version of Makeobj. For those already familiar with the installation and setup, here are some quick links:

us (site down, do not visit) ]/simutrans-experimental.zip]Windows binaries
Configuration files (alternatively, us (site down, do not visit) ]/simutrans-experimental-config.zip]here)
(Linux binaries pending - see here for details).
us (site down, do not visit) ]/Pak128.Britain-Ex.zip]Pak128.Britain-Ex (updated for Simutrans-Experimental 7.0).

7.0 contains a number of significant enhancements, and is up to date with the latest stable version of Simutrans-Standard (later versions of the nightly builds incorporate precursors of the network code, which will need some work before they are merged in). Bernd Gabriel's new physics model is not incorporated for the time being, as there are still some issues to be addressed with that, but it may well be incorporated in a later version. A list of changes follows (for this version, a list of changes for Simutrans-Standard is omitted: only the changes made to the Simutrans-Experimental specific parts of the code are shown).

Changes (Simutrans-Experimental)

FIX: Crashes on industries closing
CHANGE: When industries close down, new industries spring up to replace them. The "chance" and "distributionweight" parameters are used to govern how many new industries open to replace closed industries.
FIX: Possible crash if "distributionweight" set to 0 for factories.
FIX: Scaling not properly set when loading games saved in Simutrans-Standard.
CODE: Cleaned up code for closing industries
CHANGE: Close down industries more slowly.
CHANGE: Chance of adding new industries to replace closed ones each month dependant on proportion of industries closed compared to those remaining.
CHANGE: Industry density is now based on the population level and 1/distributionweight of each industry. The simulation will attempt to keep a constant industry density throughout time, replacing closed down industries with new industries.
CHANGE: Notifications that industries have closed down are treated in the same way for UI purposes as notices that new industry chains have been created (i.e., they can both be controlled by the "industries" option in the message options dialogue).
CHANGE: Small change to translation text in the message options window.
CHANGE: Simutrans-Experimental specific introduction ticker text.
WORKAROUND: Disabled switching to distributed pathing, because distributed pathing does not work (crashes). Pending fix to or removal of distributed pathing. There should be no issues with using centralised pathing - it has been tested for some time, and is found to perform well and be stable.
CHANGE: Way constraints are now enabled for watercraft.
CHANGE: Any watercraft with any permissive way constraints set will not be able to travel on the open seas.
CHANGE: Industries when they close down can now be upgraded to another industry of the same dimensions and with the same input and output goods using upgrade[X] = (name of factory), where X is a number starting from 0 and incrementing by one for every possible upgrade, and "name of factory" is the name of the industry to which the present industry can be upgraded. If more than one is specified, then, if an upgrade occurs, it is chosen at random from the list, counting only those that are neither introduced in the future nor obsolete.
CHANGE: Stations/stops can have their: (1) price; (2) maintenance cost; and (3) capacity set independently of the "level". In default of setting any of the three values individually, the calculation from "level" is used in the conventional way. Keywords in the .dat file are "station_capacity", "station_maintenance" and "station_price". These are all scaled with "scale_factor".
CHANGE: The default introduction date, where it is not specified in a .dat file, is now 1492, not 1900.
FIX: Scaling the maintenance cost of ways checked whether the price was greater than zero before scaling, rather than checking the maintenance cost.
CHANGE: Removed unnecessary line break in the descriptive text for Makeobj.
ADD: Tractive effort ("tractive_effort") parameter to the vehicle .dat files. This is not currently used, but reserved for future enhancements to the physics engine.
CHANGE: Optimisations of the pathfinding routines [Knightly]

Feedback and testing

Thanks to all those who have tested Simutrans-Experimental and provided useful feedback. The major changes to the model for opening/closing industries in this version, the addition of way constraints for watercraft and a number of the bugs fixed would not have been possible were it not for the invaluable testing and feedback provided by members of this community. I am grateful to everyone for their patience in waiting for this new version, which has been delayed greatly whilst bugs are fixed and the new features implemented.

In order fully to test the industry upgrading, a new version of Pak128.Britain-Ex will be released soon (although may be delayed somewhat, since I am in the process of moving to a new flat). In the meantime, I have released the sources for the version of Pak128.Britain-Ex which I have used successfully to test the industry upgrading code, so anyone with the wherewithal to do so can test it by compiling their own version of Pak128.Britain-Ex with the latest version of Makeobj.

Any feedback from this version would be very much appreciated, including (1) bug reports; (2) information as to how well-balanced that the game is; (3) any difference in approach/strategy needed for Simutrans-Experimental as over Simutrans-Standard; and (4) any other useful feedback.

Thanks again to all those who have worked to test Simutrans-Experimental, and happy playing!

Re: [New release] Simutrans-Experimental 7.0

Reply #1 – November 10, 2009, 10:18:06 am

the time run very fast...

fyi: i've download all the files above and place them in each folders. also copying the font folder from the standard source...

Re: [New release] Simutrans-Experimental 7.0

Reply #2 – November 10, 2009, 01:16:01 pm

What do you mean that the time runs very fast - are you describing a problem with the program? If so, I should be grateful for more details.

Re: [New release] Simutrans-Experimental 7.0

Reply #3 – November 10, 2009, 03:05:54 pm

Sorry for my bad English words...

The game run too fast, i think there is no better words to describe this.
I've uploaded the screenshots below, see the ticker in the bottom of the screen.
The ticker moves normally, but look at the date and time, it changing every seconds.
And the season/climate also change as the changing of the time.

I'm running on Windows XP.

Re: [New release] Simutrans-Experimental 7.0

Reply #4 – November 10, 2009, 09:32:50 pm

I am getting a crash fairly frequently, and am not seeing any particular operation that causes it...

FATAL ERROR: vector_tpl<T>::[]
index out of bounds: 1 not in 0..0

Re: [New release] Simutrans-Experimental 7.0

Reply #5 – November 11, 2009, 01:44:18 am

Quote from: wlindley – on November 10, 2009, 09:32:50 pm

I am getting a crash fairly frequently, and am not seeing any particular operation that causes it...

FATAL ERROR: vector_tpl<T>::[]
index out of bounds: 1 not in 0..0

I have gotten that a couple of times, always when clicking on a depot.

Re: [New release] Simutrans-Experimental 7.0

Reply #6 – November 12, 2009, 03:58:14 pm

Am still on the 2009-11-02 version, as there are no later Linux binaries, but I have been seeing this for awhile. The train report lists separate cargos in the wrong place, even though the "consist" report shows the cars (carriages) properly loaded.

Also, why is "Prohibitive 0" listed as "May" while the "permissive" ones are listed as "Must" ? Isn't that backwards. Also "Prohibitive 0" is not terribly descriptive, (I realize that's probably a pak issue).

Re: [New release] Simutrans-Experimental 7.0

Reply #7 – November 12, 2009, 10:05:34 pm

Quote from: wlindley – on November 12, 2009, 03:58:14 pm

Am still on the 2009-11-02 version, as there are no later Linux binaries, but I have been seeing this for awhile. The train report lists separate cargos in the wrong place, even though the "consist" report shows the cars (carriages) properly loaded.

This bug should be fixed in r2576, which should be incorporated in exp...

Re: [New release] Simutrans-Experimental 7.0

Reply #8 – November 17, 2009, 03:06:32 pm

Still no 32-bit Linux builds. from http://www.43-1.org/~simutrans/simutrans-exp/i386/logs/20091108-0248.log

Code: [Select]

===> CXX gui/colors.cc
In file included from gui/colors.cc:13:
gui/../path_explorer.h:290: error: integer constant is too large for 'long' type
...
Build failed.

and subsequent builds just say:

Code: [Select]

Nothing to do: already built this version.

Re: [New release] Simutrans-Experimental 7.0

Reply #9 – November 17, 2009, 05:23:22 pm

Well, here: http://forum.simutrans.com/index.php?topic=3721.0 Knightly says he has fixed the problem, but it is not in jamesspetts' repository yet... Patience *smile*

Re: [New release] Simutrans-Experimental 7.0

Reply #10 – November 17, 2009, 05:39:58 pm

Thank you everyone for your feedback; it is very helpful. I have not been able to work on Simutrans in the last week or two because I have just moved into a new flat and don't have internet access there yet; I have internet access at work, but can't work on Simutrans here. I shall be able to be more active once I have got my internet set up.

Re: [New release] Simutrans-Experimental 7.0

Reply #11 – December 07, 2009, 10:16:12 am

I would like to report a finding, using Simutrans 102.2 Experimental 7.0 and Pak.Britain-Ex 0.3.

What I found is that the parameters: growthfactor_villages, growthfactor_cities and/or growthfactor_capitals appear to disable the parameter maintenance_building.

I have set maintenance_building to 100 credit cents (or 1 credit in the GUI), making the loading dock have a maintenance cost of 4.00c. However, when I tinker with the growthfactor settings (for instance increasing them by a factor of 2 or 10) and create a new map, I find that the loading dock has it's default cost of 32.00c back.

At the same time, the original parameter city_weightfactor appears to have no effect. Regardless of the setting (ranges I've tried range from 400 to 5000), with an internal p****engers service the town will grow by an average of 250-350 per year.

I hope this helps.

Re: [New release] Simutrans-Experimental 7.0

Reply #12 – December 07, 2009, 10:52:18 am

Dutchman on Rails,

thank you for your reports. As to the first issue, this would appear to be a potential bug with the code, but this code was simply imported from Simutrans-Standard - have you checked whether you get the same results with Simutrans-Standard, before I start trying to track it down?

As to the second point, the city weight factor has been replaced entirely with the growth factors from Standard, as they do the same thing in a more sophisticated way. I thought that I had documented that...?

Re: [New release] Simutrans-Experimental 7.0

Reply #13 – December 07, 2009, 06:39:02 pm

Hi James,

About the second part, though I've taken a brief look for documentation on this, I probably missed it. Sorry about that.

About the first part, if I have the time, I'll do a few tests with Simutrans standard and Pak.Britain 1.05.

Edit: I have don the test mentioned, but got confused. It seems with Simutrans standard 102.2 and Pak.Britain 1.05, the maintenance cost of 4 for a simple bus stop and 16 for a loading bay is the standard for maintenance_building = 100. It seems my capability to get it down to 1.00 and 4.00 is the anomaly...

I've uploaded a savegame that has the lower maintenance_building setting in effect. I hope it helps. Here is the link.

http://simutrans-germany.com/files/upload/DoR Lichfield 1750-01.sve

I hope this helps.

Re: [New release] Simutrans-Experimental 7.0

Reply #14 – December 08, 2009, 02:51:40 am

Any news on an 'offical' Linux binary?....

I downloaded the tarball of the latest version from the webpage (git annoyingly demands that you jump through hoops before using it the way one would use svn, so I still haven't figured out how) and am going to try to debug and compile it myself.

Re: [New release] Simutrans-Experimental 7.0

Reply #15 – December 08, 2009, 07:55:28 am

Neroden,

I am currently working on version 7.1, which should be available in both Windows and Linux versions. However, I am having difficulty in tracking down bugs iwth loading and saving, and also with convoy loading times, and will not be able to release it until those bugs are fixed. If you'd like to help in fixing those bugs, see this thread, which will also tell you how to access the latest development code. Thank you very much for your interest in Simutrans-Experimental, and apologies for the delay.