Page 20 of 138
Re: Chill's patchpack v9_3
Posted: 10 Sep 2010 15:06
by ChillCore
Programmable signals:
Can you please be more specific? It works over here.
You mean that the window does not open or you mean that you do not know how to program the signal?
Tunnel signals:
Yes it is the version as found in the first post but the syntax is modified to work with the programmable signals patch.
In the patchpack however you will be searching for 4 instead of 6. (Just search for "100 tiles" to find the sweet spot

)
Re: Chill's patchpack v9_3
Posted: 10 Sep 2010 15:14
by Wasila
Programmable signals:
I open the Signal Selection window.
I then lay down a signal.
I press the 'cog' (program signal) button.
I click the signal.
Is there something wrong there?
Also I am afraid to report a crash. According to my own investigations (I had the same problem in my own custom build), I believe that it is a result of the extra-large-maps patch. .log and .png attached.
Re: Chill's patchpack v9_3
Posted: 10 Sep 2010 15:54
by ChillCore
Programmable signals:
Are you sure you are clicking a programmable signal? If you click a pbs signal after the program button noting will happen indeed.
You can not program other signals types and the signals to depend upon can only be exit signals. (combo signals are exit signals too.)
Crash:
Please tell me how to reproduce. I can remove tramtracks without problems.
As you modified the source it would be helpful if I knew what exactly you changed.
If you PM me the patch from your source I can more easily compare it to mine to see if you have maybe forgotten to do something while modifying/bumping. No promises that I can fix it though and please make sure that the new files are included also.
There is a reason I do not always post my patch pack against the current trunk or ask not to bump. Read: if there are codechanges in trunk those changes have to be made in the patches also to keep them working correctly, if they are affected by the codechange that is.
Sometimes the change is simple, sometimes I need to go and dig the logfiles if I missed it when the change was made to trunk.
Re: Chill's patchpack v9_3
Posted: 10 Sep 2010 16:12
by Wasila
Programmable signals:
Are you sure you are clicking a programmable signal? If you click a pbs signal after the program button noting will happen indeed.
You can not program other signals types and the signals to depend upon can only be exit signals. (combo signals are exit signals too.)
Ooohhhh! I just noticed that there was a 'programmable signal'.
Please tell me how to reproduce. I can remove tramtracks without problems.
As you modified the source it would be helpful if I knew what exactly you changed.
If you PM me the patch from your source I can more easily compare it to mine to see if you have maybe forgotten to do something while modifying/bumping. No promises that I can fix it though and please make sure that the new files are included also.
There is a reason I do not always post my patch pack against the current trunk or ask not to bump. Read: if there are codechanges in trunk those changes have to be made in the patches also to keep them working correctly, if they are affected by the codechange that is.
Sometimes the change is simple, sometimes I need to go and dig the logfiles if I missed it when the change was made to trunk.
I believe that the problem only occurs with a large map - starting one and then messing around with trams a bit (two trams with orders usually) seems do the trick, although it has crashed at other times. The patch is compiled with the same revision as is stated in the .diff's file name.
Changed:
Code: Select all
+/* Maximum bridge size is 100 tiles, 16 bit available for signals in m2 =>
+ * 50 blocks of 2 tiles available between ramps */
+static const uint8 bridge_signal_block_size = TILE_SIZE * 2;
Code: Select all
+ while (bridge_signal_position <= bridge_section) {
+ bridge_signal_position += 2; // 2 tiles is block size, max 48 signals, 50 blocks * 2 = 102 on bridge of max length e.g. 100 tiles
I'll PM the diff to you.
There was, while I was testing the game for the previous crash, another crash! Attached is the usual. Again, this was caused when setting up a few trams and after giving them orders to a station.
Re: Chill's patchpack v9_3
Posted: 10 Sep 2010 18:47
by ChillCore
I still can not reproduce the two crashes you had Wasila.
Can you post savegames please if it happens again so I can run them through a debug build. (Maybe you still have them?)
The tunnel signal gap difference between your code and mine should not be a problem if there are no vehicles in them in your savegame.
ps: do not post 8k *8k savegames as my pc can not handle them because of the memory usage with such big maps.
Re: Chill's patchpack v9_3
Posted: 10 Sep 2010 18:53
by Rubidium
ChillCore wrote:ps: do not post 8k *8k savegames as my pc can not handle them because of the memory usage with such big maps.
Then you can't reproduce at least the first of the two crashes. Well, with any map having 16 777 216 or more tiles would allow you to reproduce it; so 4k*8k should make it possible as well. Also make sure that you've got the NewGRF debugging tools enabled and you're using the land query tool on a tile with index >= 16 777 216.
Re: Chill's patchpack v9_3
Posted: 10 Sep 2010 19:05
by Wasila
The map is 4k*8k so it should be alright. I've just noticed (speaking of the first crash) that the crash always takes place on the 12th of January - it appears to have nothing to do with trams.
http://rapidshare.com/files/418292630/GlobalTrans.sav
Attached are the newGRFs I'm using, in case it's something to do with that.
Re: Chill's patchpack v9_3
Posted: 10 Sep 2010 19:30
by ChillCore
Rubidium wrote:
Also make sure that you've got the NewGRF debugging tools enabled and you're using the land query tool on a tile with index >= 16 777 216.
That did the trick. Thank you very much. Some (down)tuning is to be done it seems.
I will also change the patch I posted in the larger maps thread once I have the correct maximum numbers to insert.
Wasila wrote:
The map is 4k*8k so it should be alright. I've just noticed (speaking of the first crash) that the crash always takes place on the 12th of January - it appears to have nothing to do with trams.
http://rapidshare.com/files/418292630/GlobalTrans.sav
Attached are the newGRFs I'm using, in case it's something to do with that.
Thank you for the savegame, that will make testing easier.
I will have a looksie to see if the issue is not caused also because of too many tiles. I am not sure if I can or should fix handling such big maps.
Also, talking about GRFs, making a habit of selecting them BEFORE you start the game will save you a lot of potential problems.
Re: Chill's patchpack v9_3
Posted: 10 Sep 2010 19:36
by Wasila
I will have a looksie to see if the issue is not caused also because of too many tiles.
I should hope not, that's the point of the patch!
Thanks for having a look at this - it would be great if you could PM me if/when you figure it out

.
Re: Chill's patchpack v9_3
Posted: 10 Sep 2010 20:54
by ChillCore
Man that Rapidshare is rapid ... Yes It took until now to download those 20MB.
For testing I changed newgrf_debug_gui.cpp. 24 to 26 5 times.
Your savegame was still running up to februari. Also I was able to query the last tiles.
Would you mind recompiling with these changes and doing some more testing Wasila before I post a new patch.
As for the "official" bigger maps patch I'm not sure what to do, downsize the maximum mapsize (in number of tiles) or make the same adjustments as mentioned above?
Re: Chill's patchpack v9_3
Posted: 10 Sep 2010 21:35
by Wasila
Man that Rapidshare is rapid ... Yes It took until now to download those 20MB.
Was it really that slow?

Perhaps I should use a new hoster... any suggestions?
For testing I changed newgrf_debug_gui.cpp. 24 to 26 5 times.
I'm not entirely sure what you mean. Could you please post exactly which bit of code was changed?
Re: Chill's patchpack v9_3
Posted: 10 Sep 2010 21:48
by ChillCore
Wasila wrote:Perhaps I should use a new hoster... any suggestions?
I was able to download it that is the most important.
No suggestions sorry.
Could you please post exactly which bit of code was changed?
Change line 47 in newgrf_debug_gui.cpp
from
to
and do the same for line 56, 60, 215 and 419. (read: replacing "24" by "26")
I'm not sure if that fixes the crash you had in rail_map.h also though. I guess not. Feel free to post a savegame if it happens again.
Re: Chill's patchpack v9_3
Posted: 10 Sep 2010 22:07
by Wasila
Oh, I understand now. I quickly ran the newly compiled game and it hasn't crashed yet; I'll test it properly tomorrow. Thanks

.
I don't know if you know this, but there were a few other things that looked like error messages which popped up while compiling.
Wasila
Re: Chill's patchpack v9_3
Posted: 10 Sep 2010 22:28
by ChillCore
I did not know about the compiling errors in terraform_cmd.cpp, the file compiles cleanly on my pc.
A quick look shows that all of them refer to debugging messages, they should not affect gameplay.
Thank you for mentioning them.
If it is not too much trouble and you find some time, would you mind compiling the more height levels patch alone to see if they show up in there also ... that would be much appreciated.
If they are there also I could report them to ic111.
Re: Chill's patchpack v9_3
Posted: 11 Sep 2010 07:17
by Wasila
Very curious... I compiled the moreheightlevels patch in the thread's OP with the corrent version (r20761) and not a single hunk succeeded...
Re: Chill's patchpack v9_3
Posted: 11 Sep 2010 07:55
by Muzzly
Few words about compiling experience.
For 3-4 week I was using minGW. I do not remember what version it was but I am sure it was not the latest one, I tried to uses the version, that wiki suggest ( if it was still available).
After a while I realized that minGW compiled version was crashing very often and Visual Studio compiled version not.
From then I uninstalled minGW and now I am using only Visual Studio. I cannot say if this will happen if you take the latest minGW version.
Because Wasila is using minGW, I would suggest to reproduce crash with VC binary, just to be sure

Re: Chill's patchpack v9_3
Posted: 11 Sep 2010 08:06
by Terkhen
Wasila wrote:Oh, I understand now. I quickly ran the newly compiled game and it hasn't crashed yet; I'll test it properly tomorrow. Thanks

.
I don't know if you know this, but there were a few other things that looked like error messages which popped up while compiling.
Wasila
The second screenshot might be related to
http://mingw.org/wiki/FAQ (check the first two points of Porting Issues).
Wasila wrote:Very curious... I compiled the moreheightlevels patch in the thread's OP with the corrent version (r20761) and not a single hunk succeeded...
http://www.tt-forums.net/viewtopic.php?p=902415#p902415
Re: Chill's patchpack v9_3
Posted: 11 Sep 2010 08:58
by Wasila
I'm not really willing to update my minGW after spending so much time setting it up in the first place...
How exactly do I convert the file to EOS?
Re: Chill's patchpack v9_3
Posted: 11 Sep 2010 09:37
by Nickel_Plate
I have not run program yet but i had same compiling errors as Wasila using MinGW.
Re: Chill's patchpack v9_3
Posted: 11 Sep 2010 16:11
by Nickel_Plate
Patched and compiled r20768.
Copy/Paste crashes everytime try to use it. NO CRASH LOG, dumps straight to windows.
Mini map only goes to top corner.