NewStations v0.6

[Drakous79]

Blame curious engineers.

Ok, thank you for the effort Michael!

[michael blunck]

So far I have a save (attaching), where the crash happens on 24th of April.

Mmh, I don´t get any crash, neither on april 24th, nor later. OTOH, I get some strange graphical glitches.

xxx_may.png (71.24 KiB) Viewed 7130 times

xxx_october.png (39.79 KiB) Viewed 7130 times

regards
Michael

[Drakous79]

Ah, and what is your OS and game setup?

I have:
* Windows 7 64-bit
* Running the game in Windows XP (SP3) compatibility mode
* Playing in window - resolutionwidth 1670 / resolutionheight 970
* Haven't copied dplayx.dll (and renamed to dplay.dll) TTD root folder
* TTDPatch 2.6 alpha 0 r2366 (Windows)

I don't see those white glitches near Findingwell. Is it monorail track in it? What happens, if you lower the land there?

---
Edit: I tried to set various settings in Windows compatibility tab and TTconfig.exe, but nothing helps. Then I double checked my NewStations version and it was 0.51 (reverted back to check waypoints behavior in 0.51 and forgot to switch back ) Now it is back to version 0.6:

Sprite number 1041 crash with "xxx" save (TRT32) on 24th April no longer happens. No white glitches near Findingwell in October.
Sprite number 0008 crash with "xxx" save (TRT32) on 22nd June 2040 happens always.
Sprite number 0008 crash with "- Crashtest" (TRT31 posted here) happens randomly in 1967. In this year I also had sprite number 12DA crash.

[michael blunck]

I don't see those white glitches near Findingwell. Is it monorail track in it? What happens, if you lower the land there?

See screenshots. I can both lower the land and/or destroy those. The right glitch now had part of a normal track in it, BTW. I think this is spurious overwriting memory somewhere, but I can´t check out the savegame for any sort of corruption myself.

xxx_glitch1.png (32.43 KiB) Viewed 7066 times

xxx_glitch2.png (28.84 KiB) Viewed 7066 times

BTW, as you´re already in year 2039, when did you start that game?

Sprite number 1041 crash with "xxx" save (TRT32) on 24th April no longer happens. No white glitches near Findingwell in October.

This crash is still a mystery to me, since it´s in the middle of a (real) sprite block with no connection to any AI action.

Sprite number 0008 crash with "xxx" save (TRT32) on 22nd June 2040 happens always.
Sprite number 0008 crash with "- Crashtest" (TRT31 posted here) happens randomly in 1967.

As well as this crash is a mystery, because it´s in a totally different part of the code, but also not directly related to AI action, but usually referenced by player action, since it´s a simple action0A used to establish a recolour sprite (original gray glass) which is frequently used in your game.

In this year I also had sprite number 12DA crash.

And this is a sprite for some of the modern freight sheds. The AI should never try to build one.


All in all, I think your game is corrupted in some way, so the sprite errors and crashes are somewhat spurious. I fear that I can´t help you with this, except suggesting to start a new game and check whether those crashes persist. Sorry about that.

regards
Michael

[Drakous79]

I wonder, why I don't see those glitches

...but I can´t check out the savegame for any sort of corruption myself.

BTW, as you´re already in year 2039, when did you start that game?

How can be the savegame checked for corruption?

In despair, I tried to rename the savegame to *.sv0, loaded it in scenario editor, saved it, renamed it back to *.sv1 and the crash (0008) happened lately in 1969 instead of 1967.

The savegame has some history Started it in 2005 real time with TTDPatch 2.0 rev 1 and the first save I have is from 1940 TTD time. 2039 TTD time is cheated Cht: Year 2039 so I could check waypoints' behavior with maglev.

Sprite number 0008 crash with "- Crashtest" (TRT31 posted here) happens randomly in 1967.

As well as this crash is a mystery, because it´s in a totally different part of the code, but also not directly related to AI action, but usually referenced by player action, since it´s a simple action0A used to establish a recolour sprite (original gray glass) which is frequently used in your game.

I let the time flow without doing anything. Using fast forward time, sitting back and watching.

All in all, I think your game is corrupted in some way, so the sprite errors and crashes are somewhat spurious. I fear that I can´t help you with this, except suggesting to start a new game and check whether those crashes persist. Sorry about that.

Good suggestion.

I'd try OpenTTD, but the savegame I'd like to finish first doesn't work in it. OpenTTD seems more actively developed so it looks like a way to go for future play.

There is another option - to turn NewStations off and finish the game, but it would feel like not playing fully, because of absence of your wonderful work Losing Bamberg? Nooo. NewStations really add a lot to the experience.

Or I can save offten, and turn off the set only for critical periods of time.

Hope I didn't waste your time or derailed the thread too far with this crazy 9 years old game's problems.

[michael blunck]

I wonder, why I don't see those glitches

Memory allocation seems to be different for both of us. And that´s a hint that there´s something fishy with your savegame.

The savegame has some history Started it in 2005 real time with TTDPatch 2.0 rev 1 and the first save I have is from 1940 TTD time. 2039 TTD time is cheated Cht: Year 2039 so I could check waypoints' behavior with maglev.

Well, TTDPatch is quite stable with all that newGRF disabling and re-enabling, time cheats, and starting it with a rather old TTDPatch version. It´s for sure that you couldn´t have done the same with OTTD.

As well as this crash is a mystery, because it´s in a totally different part of the code, but also not directly related to AI action, but usually referenced by player action, since it´s a simple action0A used to establish a recolour sprite (original gray glass) which is frequently used in your game.

I let the time flow without doing anything. Using fast forward time, sitting back and watching.

Well, I meant you were already using that gray glass, resp the recolouring routine, for the existing stations, and it seemd to work w/o a crash.

I'd try OpenTTD, but the savegame I'd like to finish first doesn't work in it. OpenTTD seems more actively developed so it looks like a way to go for future play.

OTTD has a bit more newGRF features, but not so for stations, so NewStations 0.6 should work for both programs. O/c, there´s that problem with adding sprite offsets, corrupting those waypoints (and other "multiple groundtile" feature usage), either in TTDPatch or in OTTD.

How can be the savegame checked for corruption?

We could ask Lakie in the "Problems with TTDPatch" sub-forum.

In despair, I tried to rename the savegame to *.sv0, loaded it in scenario editor, saved it, renamed it back to *.sv1 and the crash (0008) happened lately in 1969 instead of 1967.

That´s a good approach, I´ve also tried it successfully many times, mostly for foreign corrupted games. Unfortunately, it doesn´t work all the time.

Hope I didn't waste your time or derailed the thread too far with this crazy 9 years old game's problems.

I should also test the set with one of my 10 year old savegames.

regards
Michael

[Drakous79]

O/c, there´s that problem with adding sprite offsets, corrupting those waypoints (and other "multiple groundtile" feature usage), either in TTDPatch or in OTTD.

I don't understand those additional sprite offsets. I mean stations are multiple tiles too and they look ok when built, but waypoints, technically stations too, are affected.

I should also test the set with one of my 10 year old savegames.

That'd be very extensive testing

---
Writing again, because sprite number 0008 seems to be related to AI actions after all! Good news, but it depends on point of view.

In danger of losing my favorite station, I loaded the very first save. There is nothing built in 1940 except towns.

Resaved in 1942. AI started construction this year and the crash happened later. Reloaded 1942 save and set "Difficulty settings / Maximum no. competitors: 0" ... no problems since.

I went further and checked ttdpatch.cfg setting:

Code: Select all

aichoosechances off

The AI started on 8th April 1942 (Wrintford Transport) ... crashes always on 26th June 1942 (0008)

Code: Select all

aichoosechances on
aibuildrailchance 0

The AI started on 8th April 1942 (Pedtown Transport) ... no problems for 10 years

Loaded 1967 save, bought all AI companies and set no. competitors:0 ... no problems for 38 years

---
Hurray, I can play on even with fishy AI landscape architects, that are great for dirty work (like landscaping, city buildings removal, tree cutting) in or around cities. Considering me, 0008 is solved, even if I know there is something wrong.

[michael blunck]

O/c, there´s that problem with adding sprite offsets, corrupting those waypoints (and other "multiple groundtile" feature usage), either in TTDPatch or in OTTD.

I don't understand those additional sprite offsets. I mean stations are multiple tiles too and they look ok when built, but waypoints, technically stations too, are affected.

As I wrote here, it´s a technical problem with arranging sprites for that "multiple ground sprites" feature, both available in TTDPatch and OTTD. I´m using this feature to implement those footpaths over track for the waypoints, but also for other minor features in NewStations (buffer stops, separating track, ...). Obviously, TTDPatch and OTTD have a different understanding with counting sprites, so TTDPatch in that case adds a second waypoint building instead of the footpath ground sprite.


W/r to those crashes, I´m going to run a longer (fresh) test game with the AI in TTDPatch these days.

regards
Michael

[Thun]

Can someone explain me how the tow different subways in v0.6 work? I wasn't able to find any information in the Readme. Thanks!

[michael blunck]

Can someone explain me how the tow different subways in v0.6 work? I wasn't able to find any information in the Readme. Thanks!

Please, see here.

regards
Michael

[Transportman]

I've found a bug with the Small Station Hall, the non-track tile is not always set to non-track when using multiple platforms. See the attached screenshots, after removing the station, the tiles with tracks on it are the tiles not set to a non-track status.

[michael blunck]

I've found a bug with the Small Station Hall, the non-track tile is not always set to non-track when using multiple platforms.

Thanks for the report. Silly c&p bug, the non-track large hall was OK.

regards
Michael

[keibi]

When you will add edge and island platforms to newgrf (with passengers graphics in platform)?

PS: I like your newgrf and passengers graphic in stations
PS2: Sorry for my english...

[S-Transport]

I do not want to demand anything, but that would be really nice.

[Dave]

Don't wish to speak for Michael but I understand he isn't a fan of how it looks (it does look a bit clunky...)

[Nagyzee]

I've decided to create and play a scenario in TTDPatch (r2366) for nostalgic reasons. That's how I've run into this possible bug with NewStations v0.6

I'm playing in temperate with variable snowline. And quite a lot of NewStations' station tile graphics fall apart when I build the station above snowline. On the other hand they display correctly if under the snowline. See attached picture. I've checked and the problem is not present in v0.44.

If needed I can compile a list of the station tiles that are affected. Thanks in advance for looking into it.

[michael blunck]

I've decided to create and play a scenario in TTDPatch (r2366) for nostalgic reasons. That's how I've run into this possible bug with NewStations v0.6

Thanks for the report.

AFAIR, v0.6 has never been tested in TTDPatch ...

I'm playing in temperate with variable snowline.

... let alone with "temperate and variable snowline". Tests have usually been done with AlpineClimate, which is essential "sub-arctic".

And quite a lot of NewStations' station tile graphics fall apart when I build the station above snowline. On the other hand they display correctly if under the snowline. See attached picture. I've checked and the problem is not present in v0.44.

I presume that "partially snowed" is handled differently in TTDPatch and OTTD. At least, decision which station sprites (snow or non-snow) to be used is done on the same variable (var42 - terrain, on or above snowline) in both programs. So, "Bendinghattan West" could be OK.

OTOH, "Bendinghattan South" looks indeed broken.

If needed I can compile a list of the station tiles that are affected. Thanks in advance for looking into it.

That would be interesting and probably helpful.


Anyway, next version of NewStations (v0.7) won´t be playable in TTDPatch at all, but will need a recent version of OTTD.

edge and island platforms

It´ll make available those platform styles, filling the "gaps" with fake track when being used together with the DBRails track set to be released together with v0.7. See screenshots on the German tt-forums:

http://www.tt-ms.de/forum/attachment.php?aid=6809
http://www.tt-ms.de/forum/attachment.php?aid=6790
http://www.tt-ms.de/forum/attachment.php?aid=6791

regards
Michael

[wallyweb]

I presume that "partially snowed" is handled differently in TTDPatch and OTTD.

There is a difference indeed. Some time ago I analyzed the issue and this is the result:

snowlineComparison.png (251.06 KiB) Viewed 774 times

[Nagyzee]

Here are the details:

v0.44 shows correct snowy graphics in arctic climate above the snowline. In temperate it shows graphics without snow even if above the snowline but there are no glitches otherwise.

v0.6 also shows correct snowy graphics in arctic above the snowline. In temperate the following is the situation:

- From the Platform class Roofs, Platform, Benches, Station Hall (small), Station Hall (large), Empty , Parking lot (front), Parking lot (back), Flat Roof, Glass Roof and Stairs fall apart graphically, Overpass and Underpass show non-snowy but otherwise correct graphics and Void shows correct snowy graphics.

- From Station buildings (small) class all fall apart graphically.

- From Station buildings (medium) class Freiburg (back view) and Freiburg (front view) fall apart but Avetri, Bamberg and Bamberg (extension) show correct snowy graphics.

- In the Non-track tile class all have the graphical glitch.

- In the two Freight station classes and the Locomotive depot class all station tiles show correct snowy graphics. (The roundhouses don't have snow but that's mentioned in the readme as well.)

- In the Signal box class all show correct graphics but without snow.


As far as I can see the station tiles newly added in the v0.5 and v0.6 releases tend to not have the glitch and to handle snow in temperate perfectly while older station tiles usually either glitch or show correct graphics but without snow. Those tiles that have the correct snowy graphics handle not only snow in temperate but also the variable snowline part fine; they change graphics exactly when they should. Tiles with glitchy graphics alternate between the glitchy graphics (when above snow) and correct non-snowy graphics (when below snow) also at the right time. So the glitchy graphics aren't "locked in" at creation of the tile and the variable part of the snow feature doesn't seem to cause any problems for your grf in TTDP.

Attached in a zip file are pictures of the glitches.

[michael blunck]

@Nagyzee

Thanks for the detailed report. I´ll have a serious look into this.

regards
Michael