Stephenson-Siemens

[Mariculous]

It seems the pakset at stepehnson.siemens is not up-to-date. I just updated two devices using the nightly updater and it's showing a missmatch on both devices.

[Vladki]

Let's see tomorrow, it updates automatically every day. And I had to update today...

[Mariculous]

It still tells me a misssmatch so I am still force-connecting.

[Vladki]

That may be the reason why it crashes so often?

[Mariculous]

Maybe but Idon't think so. It should desync in case of differences, instead of crashing. Additionally, it crashed last night when I was not connected.
I have experienced frequent crashes when much route search happens in a short period of time, so I guess the stupid waypoint reversing could be the cause.

[Vladki]

I have updated my computer now from http://bridgewater-brunel.me.uk/downloads/nightly/pakset/ and it matches with server.
Probably you have to delete the whole pak128.britain-ex folder and unpack the new one.

[DrSuperGood]

I have updated my computer now from http://bridgewater-brunel.me.uk/downloads/nightly/pakset/ and it matches with server.Probably you have to delete the whole pak128.britain-ex folder and unpack the new one.

The nightly updater was designed to track both file additions and removals. There should be no need to do this unless some error I did not predict occurs.

If one deletes the pakset folder, deletes the "manifest.hash" file and then re-runs the updater it should automatically pull down the entire pakset again as it will detect the missing files and replace them.

[Mariculous]

Seems to be an issue with my laptops and funnys simutrans. My PC seems to be fine.

Edit: deleting the pakset dir and Simutrans Extended.hash, there was no manifest.hash, worked. So something went wrong at the update of ot both devices.

[Mariculous]

It seems the server currently suffers massive desync problems.
It will desync a few seconds after connecting to the server.

Edit: It doesn't desync anymore, it just crashed.


Could have been me causing both, I am currently trying to reproduce.
I just replaced my trams so they all went to the depot, for sure happy livelocking happened because I did not put oneway signs around the deopot. When in drive_by_sight, trams will try to reverse when they encounter a vehicle driving in the opposite direction and that rerversal semmed to have caused the desyncs but I can not yet confirm this for sure as I could not setup a reproduction case for this but I observed desyncs matched the reversal timer of my trams.

The crash was most likely because I was trying to place a one way sign underneath those livelocking trams.

I am currently trying to reproduce this on a localhost server and will update the existing crash bugreport and create a new one related to the desync as soon as I can confirm this.

[jamespetts]

Thank you for the report. I should be grateful if any reproduction case and bug report could be posted in its own thread in the development forum, one thread for the loss of synchronisation and another for the crash.

[Mariculous]

As mentioned, there already exists a report for the crash and I will update it with the reproduction case, I just reproduced it now but well, cliche Britains would say it's teatime, so it will take a while until I update that post.
I could not yet reproduce the desync but as soon as I got it reproduced, I will for sure file a bugreport. This post here is not meant to be a bugreport but moreover some information for anyone playing or administration the server.

[jamespetts]

Splendid, thank you for clarifying.

I should note that loss of synchronisation bugs are usually extremely hard to diagnose, sometimes taking many, many months of doing nothing else to fix.

[Vladki]

Stephenson-Siemens fails with this error:

Reading menu configuration ...
ERROR: create_general_tool():   cannot satisfy request for general_tool[40]!
For help with this error or to file a bug report please see the Simutrans forum:
http://forum.simutrans.com
Warning: create_simple_tool():  deprecated tool [27] requested
Reading object data from Pak128.Britain-Ex/...
Message: ground_desc_t::register_desc():        setting raster width to 1
FATAL ERROR: vehicle_reader_t::read_node() - Incompatible pak file version for Simutrans-Ex, number 5
Aborting program execution ...

For help with this error or to file a bug report please see the Simutrans forum at
http://forum.simutrans.com
FATAL ERROR: vehicle_reader_t::read_node() - Incompatible pak file version for Simutrans-Ex, number 5
Aborting program execution ...

For help with this error or to file a bug report please see the Simutrans forum at
http://forum.simutrans.com
Aborted (core dumped)


And the same error is for swedish game

[jamespetts]

What appears to be happening is that the command line version is failing to compile with the following error:

Code Select Expand


build/server/gui/components/gui_image_list.o: In function `gui_image_list_t::draw(scr_coord)':
gui_image_list.cc:(.text+0x317): undefined reference to `display_veh_form_wh_clip_rgb(short, short, short, unsigned short, bool, unsigned char, unsigned char, bool, signed char)'
gui_image_list.cc:(.text+0x5d0): undefined reference to `display_veh_form_wh_clip_rgb(short, short, short, unsigned short, bool, unsigned char, unsigned char, bool, signed char)'
build/server/gui/convoi_detail_t.o: In function `gui_convoy_formaion_t::draw(scr_coord)':
convoi_detail_t.cc:(.text+0x6d7): undefined reference to `display_veh_form_wh_clip_rgb(short, short, short, unsigned short, bool, unsigned char, unsigned char, bool, signed char)'
convoi_detail_t.cc:(.text+0x749): undefined reference to `display_veh_form_wh_clip_rgb(short, short, short, unsigned short, bool, unsigned char, unsigned char, bool, signed char)'
build/server/gui/convoi_detail_t.o: In function `gui_convoy_maintenance_info_t::draw(scr_coord)':
convoi_detail_t.cc:(.text+0xfdb): undefined reference to `display_veh_form_wh_clip_rgb(short, short, short, unsigned short, bool, unsigned char, unsigned char, bool, signed char)'
build/server/gui/convoi_detail_t.o:convoi_detail_t.cc:(.text+0x1078): more undefined references to `display_veh_form_wh_clip_rgb(short, short, short, unsigned short, bool, unsigned char, unsigned char, bool, signed char)' follow
collect2: error: ld returned 1 exit status
common.mk:21: recipe for target 'build/server/simutrans-extended' failed
make: *** [build/server/simutrans-extended] Error 1


This means that the old version of the server is running, which is incompatible with the latest pakset version. I am looking into this now.

[jamespetts]

Now fixed - emergency recompile underway.

[Mariculous]

The server seems to be still down

[Vladki]

I'm not at home so I can't update and start it. Please be patient

[Mariculous]

Okok, I am very patient. When will the server be online? 

[Vladki]

Probably before 20:00 CET

[Mariculous]

Don't hurry! I was joking!

[Vladki]

Can anyone connect now? I have updated server and my client but get pakset mismatch.

[Mariculous]

I got the same and just force-connected^^
Did you try deleting the Simutrans Extended.hash and running the updater again?

[Vladki]

I have my own updater - there's no *.hash file at all. I allways delete the whole pakset directory and unpack fresh.

[Mariculous]

strange...
Force connecting worked fine without any desyncs so far.

[Vladki]

@Ranran, please check here. I have double checked that the pakset at my client and server is exactly the same, but network connect window says there is packet mismatch.

[Ranran]

Sorry for the inconvenience. (´・ω・｀)
Can anyone verify that this fix is correct?
https://github.com/Ranran-the-JuicyPork/simutrans-extended/tree/fix-for-ex14-8

[jamespetts]

I have now merged this - thank you for the fix.

[freddyhayward]

How exactly does one force-connect to the server?

[Mariculous]

When it's online load the pseudo-savegame "net:<server_adress>"
In case of stephenson-siemens it is "net:list.extended.simutrans.org"

btw. you can even use -load "net:list.extended.simutrans.org" on the commandline. If you have a little start script, this is quite useful to automatically connect to the server whenever you start the game.

[Vladki]

Sorry for the inconvenience. (´・ω・｀)
Can anyone verify that this fix is correct?
https://github.com/Ranran-the-JuicyPork/simutrans-extended/tree/fix-for-ex14-8

I have updated the server and client, and still it show pakset mismatch

[Ranran]

Thank you for your report. I suppose this is due to an unintended rewrite of the formation picture reported by you in another thread.
I threw a pull request. It would be helpful if you could check this.
I sincerely apologize for the inconvenience I have caused.

[jamespetts]

Thank you for the fix: I have now merged this into the master branch.

[kittens45]

keep getting pack mismatched tried every redownloading methods

[Vladki]

please wait for "tonightly" build

[kittens45]

says I need 0.9.3 packset

[Mariculous]

It is a current bug. You need to force connect using the above described method.

[Ranran]

I may not have time to reply until next week due to my family's illness, but has the mismatch been resolved?

[Vladki]

but has the mismatch been resolved?

no

[Ranran]

I further investigated the pakset mismatch.

pak.sweden did not show much mismatch and were entirely vehicles but not related to any particular parameters.
Also, there is a case where a mismatch is determined and a case where a mismatch is not determined even for the same vehicle.

Judging from that, I speculated that it was a problem related to the timeline related to future vehicle information.
is_future has been expanded to two levels in future vehicle information.
It is helpful to confirm if the following fixes resolve the issue.
I apologize for any inconvenience this may cause.

[jamespetts]

I further investigated the pakset mismatch.

pak.sweden did not show much mismatch and were entirely vehicles but not related to any particular parameters.
Also, there is a case where a mismatch is determined and a case where a mismatch is not determined even for the same vehicle.

Judging from that, I speculated that it was a problem related to the timeline related to future vehicle information.
is_future has been expanded to two levels in future vehicle information.
It is helpful to confirm if the following fixes resolve the issue.
I apologize for any inconvenience this may cause.

Thank you for this - now incorporated.

[Ranran]

Sorry, the countermeasures did not seem to be effective. (´・ω・｀)
I am investigating the possibility of a matter when can_lead_from_rear = 0.

[Mariculous]

Listservers and stephenson-siemens down or is this some kind of DNS error here?

[Matthew]

I can't access Stephenson-Siemens either. I get a freeze when I click on the network icon. With Freahk's loading cheat the freeze is after the pakset has loaded.

GDB reports:

Code Select Expand

0x00007ffff7bc7777 in __libc_connect (fd=20, addr=..., len=16) at ../sysdeps/unix/sysv/linux/connect.c:26
26 ../sysdeps/unix/sysv/linux/connect.c: No such file or directory.

Backtrace:

Code Select Expand

#0  0x00007ffff7bc7777 in __libc_connect (fd=20, addr=..., len=16) at ../sysdeps/unix/sysv/linux/connect.c:26
#1  0x00000000005cd79f in ?? ()
#2  0x00000000005d8c56 in ?? ()
#3  0x0000000000773d53 in ?? ()
#4  0x0000000000711e4b in ?? ()
#5  0x0000000000725597 in ?? ()
#6  0x00007ffff6748b97 in __libc_start_main (main=0x410130 <main>, argc=7, argv=0x7fffffffdfb8, init=<optimised out>, fini=<optimised out>,
    rtld_fini=<optimised out>, stack_end=0x7fffffffdfa8) at ../csu/libc-start.c:310
#7  0x0000000000410199 in ?? ()

[Vladki]

Hello, the server is completely down. I have already asked Isaac what's going on. Sorry but I can't do anything about it at the moment.

[Isaac Eiland-Hall]

Yes, the server is down. I'm working on a replacement.

[Vladki]

Whoever has the last save, please share it.

[Mariculous]

I had a quick look at the server on 18th February at 13:00, it's April 2001 at roughly 5:38:46 ingame.
Anyone got a later save?

I cannot attach the save to the forums due to size restrictions but I might send it by mail.

[Matthew]

This save is from Tuesday 18 February at ~0430 CET. Earlier than Freahk's, but available now if that helps.

[Mariculous]

Any news?

[Isaac Eiland-Hall]

I have the replacement and had intended to give necessary access, but I started a new job and that has caused schedule delays. But I do have the next three days off, so I will be working on it. I do apologize for the delay.

Also, It is unknown if this server will be able to host a game. I hope it will be, but I'm unsure until it's tried.

[Mariculous]

Don't hurry and no need to apologize!
I greatly appreciate the effort and just wanted to know if there are some serious issues

It's all right. If the server cannot host the game, I might have a suitable server

[Ranran]

I proceeded to investigate the pakset mismatch issue.

I found a rail way vehicle that displays different formation pictures in network mode and offline mode.
This seems to happen on the side where bidirectional = 0 and no constraint is set.
From this I presumed that the related variables were not initialized correctly and fixed that.


I threw a pull request.
It would be helpful to confirm if this fix resolves the mismatch. (´・ω・｀)

In addition, the typo of help text has been modified in one place.

[Vladki]

The list server is up and running. I will try restoring the stephenson-siemens game when the nightly pakset build is fixed (i.e. when makeobj is fixed)

[Isaac Eiland-Hall]

Thank you all for your incredible patience with me. I am quite confident that the server situation will now be stable for the foreseeable.

[Mariculous]

I will try restoring the stephenson-siemens game when the nightly pakset build is fixed

Wait! That means it will be running on the new servers? I did not expect this tbh but that's awesome news!

Thank you all for your incredible patience with me.

Thank you for hosting all these services!

[Vladki]

Wait! That means it will be running on the new servers? I did not expect this tbh but that's awesome news!

Yes (if the hardware has enough power).

[Isaac Eiland-Hall]

I'm hopeful that it will be enough; I've just learned to try and keep my hopes in check these days. heh. :|

[Vladki]

Stephenson-Siemens is up again.

Thanks to Isaaac for providing the hardware, and to James for coding and fixing bugs...
I'll try to put some easter eggs there

[Mariculous]

The server does not seem to be up-to-date.

[Vladki]

Updated. I forgot to set up the automatic updater, now fixed.

[Mariculous]

purr-fect 😸

[Isaac Eiland-Hall]

Isaaac

Please don't take this as a complaint - I am rather amused. Most misspellings are "Issac". I've never seen "Isaaac" before. It helps make up for some of the "Issac" instances. XD

Okay, so I'm really really hoping this game does work. Sounds like it's not failing so far, at least!

[Vladki]

Not failing but quite a lot of desyncs, especially when someone is connecting, it kicks out the previous player.

[Vladki]

Stephenson-Siemens server is officially down.
- there were no regular players visiting
- the game was already in future, so no new vehicles, only obsolescence of older ones
- the nightly builds requires more recent glibc than available on the server
The last save is here: http://list.extended.simutrans.org/~simutrans/debug-saves/Stephenson-siemens.sve

If the library issue is solved I may run some other map in the future (e.g. wienna-krakow-budapest)

[freddyhayward]

Stephenson-Siemens server is officially down.
- there were no regular players visiting
- the game was already in future, so no new vehicles, only obsolescence of older ones
- the nightly builds requires more recent glibc than available on the server
The last save is here: http://list.extended.simutrans.org/~simutrans/debug-saves/Stephenson-siemens.sve

If the library issue is solved I may run some other map in the future (e.g. wienna-krakow-budapest)

I hope to see it again soon - it is good to have a smaller alternative to bridgewater-brunel, especially for new players and low-end PCs.

[jamespetts]

It would be good to see this up and running again with the library issue solved. Are you not able to set up the server to compile its own binary?

[Vladki]

It would be good to see this up and running again with the library issue solved. Are you not able to set up the server to compile its own binary?

The server is provided by Isaac (thank you). And has some other duties too, so I did not want to install compiler, and devel libraries, etc... But if Isaac would not mind, I can set it up to build it's own nightlies.

[Isaac Eiland-Hall]

I'm hosting a game? See how much I pay attention! lol.

The only thing I use that server for is server backups, so the resources can be pretty freely used otherwise. I know it's not a fast server, but if it's fast enough to run a game, awesome

[Vladki]

Thanks to Isaac for upgrading the server to ubuntu 20.04, so we can run the pre-compiled nightlies by James.  Stephenson-Siemes is up again, but I'm looking forward to replace it with a new map.

[jamespetts]

Thanks to Isaac for upgrading the server to ubuntu 20.04, so we can run the pre-compiled nightlies by James.  Stephenson-Siemes is up again, but I'm looking forward to replace it with a new map.

Excellent! A new map is probably a good idea so that people can start afresh. I suspect that you can afford to make the map a little bigger than the previous one.