River's server

[Yona-TYT]

Try with region-4.6, I'm suspecting of an error with simutrans.

[River]

i'm sorry to say that i accidentally used 4.6 as the new version on the server instead of 4.7 got confused in the list of versions of the senario i got. but fix it now.

the error is in both 4.6 and 4.7.

[Yona-TYT]

That means it is a error with simutrans because before updating this did not happen.


You can return to the previous version while the problem is resolved.


I have reported here: http://forum.simutrans.com/index.php?topic=15388.msg155380#msg155380



Is fix  [size=78%].[/size]

[5045]

Is the server now using another game version or still 7919?
It would really be helpful if you could compile your server game with the version string activated, river

[River]

I'm still using 7919 for the region server and the stable build for the other server. I keep all this information up to date on the initial post of this thread.

My server is running the build provided by http://nightly.simutrans.com/ This is a problem in the nightly builds

[DrSuperGood]

The nightly build server must be building with the wrong compiler flags for some reason. One has to explicitly turn on version builds.

[River]

The region server has been updated to Nightlybuild rr7923.

[Yona-TYT]

I removed the code that gave problems. No need to remove bridges from other players.

[River]

The problem seems fixed.

[Yona-TYT]

Try this, I'm not sure if this solves the problem.

[River]

I was first gone wait till the bug in simutrans it self is fixed. then i will add this.

[River]

Both server will be down tonight from 17:00 UTC for 4-6 hours.

[AP]

Is there an easy download link for the pakset required for this server?

I tried using the Online Installer but the default version of pak64 it gives reports being mismatched with River's Server, and arrives in a folder just called "pak" so its hard to know exactly what version number it is.

[River]

The server has been reset. for the people that are interested here the last save game of the map can be found here: https://www.dropbox.com/s/2b9iqe4ai2li978/server13356-network%2011-12-2016.sve?dl=1

@AP your problem is propebly with the region server where we are using a nightly build of simutrans. the current version used is in the first post of this topic. at this moment the version is not available on the nightly servers. i will look at a fix soon!

[DrSuperGood]

If using nightly builds one must really restart the server at least once a week with the newest nightly. It kind of defeats the purpose to keep using an arbitrary old nightly.

[River]

The region server is gone be offline to make room for a private server.

[DrSuperGood]

Public server seems offline.

[River]

i was away last night so i cound't fix it before. the server crashed on "assertion failed" again. so i'm expecting more crashes to come. can't wait till the next official release. The server is back up again

[jamespetts]

The server crashing on failed assertions suggests that you are using a debug build; surely a server should be running a release build?

[River]

umm you think the problem was caused by using -debug in start up? would have been nice to have know in september when the problem was first showing up

[jamespetts]

umm you think the problem was caused by using -debug in start up? would have been nice to have know in september when the problem was first showing up

Assertions are a feature of the programming language: they will cause the program to terminate ("crash") when a certain condition is met in the code. This is used for testing/fixing ("debugging") code: the idea is that the conditions that cause the program to terminate should never be true, so that, if they are, the programmer knows that there is a fault and that it needs to be fixed.

Because of their function, assertions only work in debug mode. When debug mode is not defined, assertions are ignored. This may mean that some subtle error is occurring that cannot easily be traced (or not, depending on whether the assertion was aptly set), but the program will not crash for an assertion failure.

Also, depending on your makefiile settings, you may well find that a non-debug build is faster than a debug build - are you using optimisations, for example? For an actual game on a server (i.e. where the idea of the server is to play, not to test), I do recommend using release rather than debug builds.

[River]

i'm using the official release but i had -debug turned on cos of some crashes that happend, hoping to find the cause.

[jamespetts]

i'm using the official release but i had -debug turned on cos of some crashes that happend, hoping to find the cause.

The -debug runtime flag will not make a difference: what matters are the flags set at compile time. I thought that Simutrans-Standard had some years ago switched to release builds for the official releases - unless you are using the nightlies, which are probably still debug builds?

[River]

no this is the official release as you can download from the website.

[jamespetts]

That is a little odd. There is much to be said for making the official releases actual release builds with debugging disabled and optimisation enabled, and having only the nightly builds with debugging enabled.

Are you able to compile your own? If so, you can disable debugging (and therefore assertions) when you do so.

[DrSuperGood]

Was it the same assertion as before?

[River]

Was it the same assertion as before?

hard to say i was to tired to think to save the log.

[prissi]

Simutrans standard was released as debug build for years. Otherwise lots of debug log messages would be gone and the performance hit was not very big.

Also, a failure on assertion usually means the server will crash soon anyway, since something is not as it should be.

[Fifty]

Hey River, any chance you could move this server to JIT2 mode? We have a lot of issues on the map right now with factories hitting in transit limits due to long distances, slow routes, multiple players serving the same route, jams, etc. I moved it myself on a local copy and it seems to improve factory dynamics significantly.

[River]

Hey River, any chance you could move this server to JIT2 mode? We have a lot of issues on the map right now with factories hitting in transit limits due to long distances, slow routes, multiple players serving the same route, jams, etc. I moved it myself on a local copy and it seems to improve factory dynamics significantly.

i was under the impression that JIT2 is bugged. But i will consider moving it to it.

[Fifty]

There are some issues (and DSG Can tell you more about them) but it is in no way unplayable, which this game is starting to become with the in transit limits as they are.

[DrSuperGood]

JIT2 mechanics are not majorly bugged, at least within the pak64 dynamics. They are not optimum though and that is something I want to look into later this month for the next release. I want to make some pretty big revisions to it to fix some mechanical issues but I guess we will have to see.

They do suffer minor bugs with a few pak128 factories which involve huge quantities. Specifically some minor rounding errors causing visually unstable output (nothing to do with game stability). Something that I will finally fix this week (that patch has been outstanding far too long).

The major bug we encountered with JIT2 was power system stability and overflow. This occurred for a variety of reasons including wrong power logic (also affecting JIT1 which is what the servers are using) and me making a wrong assumption about power logic. However this does not break JIT2 directly other than a view visual anomalies or the already broke power logic.

I think the servers need rules such as "no over expanding". The pure green player literally has expanded to service everything and in doing so services nothing well. His passenger services are almost always over crowded. His industry services are hitting in transit limits due to bottlenecks. Despite these issues he then duplicates other player lines just because he can, breaking them in the process. He really should be running a company a quarter of his current size and even then he will probably struggle to keep things running well. This is not very multiplayer friendly behaviour. I have no problem with competition, as long as it really is competition and not someone messing around pretending to be competition but rather breaking everything.

[Wormer]

I've been playing this exactly a year ago on the previous New Year holidays and there been roughly the same discussion about JIT2... I was so lucky to play on an experimental server with JIT2 then and honestly was very impressed about it. I've came back here now in a hope to find a server that's running JIT2 to my big disappointment that there isn't such a server :( Please! Host something with JIT2 support already!

[River]

yesterday i switched the server over to JIT 2

[River]

i got another assertion failed for you guys.

https://s24.postimg.org/8adkuulad/assertion_error.png

the .logs final words where:

World destroyed.
World destroyed.
World destroyed.

lets hope there is an actual fix for this.