New map features

[AlexG]

I don't know if I should post it here or elsewhere but it happens on the last trunk released with N.M.F. The error is bellow.
It happens when I've constructed a rail under a bridge (NuTracks with TBRS) and it wasn't the free grid to construct, but the bridge grid.

[cirdan]

Found another instance of the mass-replace missing a ReverseTrackdir (like with my trains in stations on page 4): Ships exit the wrong side of a shipyard. Fix: In src/pathfinder/yapf/yapf/yapf_ship.cpp, line 127 (last line of CheckShipReverse), insert a ReverseTrackdir around pos.td, making the line look like this:

Code: Select all

return best_trackdir == ReverseTrackdir(pos.td);

Right you are; fixed in the attached version.

I don't know if I should post it here or elsewhere but it happens on the last trunk released with N.M.F. The error is bellow.

If you are playing with my patch, then you should report bugs here, and I will take care of them.

It happens when I've constructed a rail under a bridge (NuTracks with TBRS) and it wasn't the free grid to construct, but the bridge grid.

I cannot reproduce this. Could you provide the exact sequence of steps that trigger the error, perhaps with a savegame?

This new version of the patch also fixes a crash with exit signals on a bridgehead.

[AlexG]

It happens when you try to place a rail over a road bridge and the rail crosses over a bridge "foot" or end. It should be an warning message there (as when you would try to lay a track over a building), right ?
I've attached a save game with the error posted earlier


And please, if is done, post it in a binary version...

[Supercheese]

nma-r25084.diff binary package attached.

[cirdan]

It happens when you try to place a rail over a road bridge and the rail crosses over a bridge "foot" or end. It should be an warning message there (as when you would try to lay a track over a building), right ?
I've attached a save game with the error posted earlier

This is precisely the kind of detailed steps that I wanted. Fixing a bug is easier for me if I can trigger it from scratch in a new game (so I can do 'git bisect'). To make sure we are talking about the same bug, it happens when you try to do what the picture below shows, right? If so, it is fixed in the attached patch.

And please, if is done, post it in a binary version...

As I already told ZxBiohazardZx, I can only provide linux binaries, since that is what I use, and they would be linked against very recent versions of all libraries involved, which I doubt many people with vanilla distros have (distros tend to be somewhat conservative in upgrading libraries). This means that my binaries would be of use only to someone who runs linux and actively compiles and installs system libraries on their own, and such people would be better off compiling my patch themselves anyway.

Besides, Supercheese and 3298 have been posting windows binaries for the last patches in a timely manner (better yet, they also report bugs ).

[AlexG]

Yes, that's what triggers the error, like in your posted picture
About binaries...I've asked because I'm not good at compiling, but now I see I've made a mistake trying to run binaries on Windows

Thanks and hope it did help somehow, my bug report

[oberhümer]

I can only provide linux binaries, since that is what I use

Using the MinGW development setup via Wine is quite straightforward, if a bit slow. Also see https://secure.openttd.org/wiki/Cross-c ... s_on_Linux.

[Supercheese]

Time for a nma-r25085.diff Win32 binary.

[3298]

I can only provide linux binaries, since that is what I use

Using the MinGW development setup via Wine is quite straightforward, if a bit slow. Also see https://secure.openttd.org/wiki/Cross-c ... s_on_Linux.

Yes, that's precisely what I used to produce the single binary I posted here. But "a bit slow" is an understatement, that setup is ssssslllllooooowwwww (one hour for OpenTTD only; also SVN somehow doesn't work in my setup). Lucky me, my Windows works again (sortof). That's probably the only place where it isn't Windoze.
I can't find any more bugs, so into trunk with it.

[oberhümer]

Well, my computer with a modern i5 does "./configure" in two and a half minutes and "make -j5" in four and a half under Wine/MinGW. Still 1/3 of the native speed, but not that bad.
For SVN, you'll need to use an original Windows msvcp60.dll in Wine (set a DLL override), as stated on the second of the linked wiki pages.

[AlexG]

Good job guys ! No more road bridge - rail problem

Thanks!

[3298]

While doing an "svn update" and compiling OpenTTD, GCC warned me:

Code: Select all

src/elrail.cpp: In Funktion »void DrawCatenaryRailway(const TileInfo*)«:
src/elrail.cpp:526:34: Warnung: »sprite_halftile« könnte in dieser Funktion uninitialisiert verwendet werden [-Wmaybe-uninitialized]

Roughly translated: "sprite_halftile" might be used uninitialized in the function DrawCatenaryRailway. I haven't seen any crashes or glitches from that and expect it to be a false positive, just reporting so you can silence the warning.

Well, my computer with a modern i5 does "./configure" in two and a half minutes and "make -j5" in four and a half under Wine/MinGW. Still 1/3 of the native speed, but not that bad.
For SVN, you'll need to use an original Windows msvcp60.dll in Wine (set a DLL override), as stated on the second of the linked wiki pages.

Mine is an old-ish EeePC with a single-core Atom, so slow compiling is somewhat expected. The really slow ./configure made me think that it still takes a very long time on better systems - it takes about 3 minutes on Wine for me, and on Linux outside the Wine environment it takes maybe 5 seconds.
About SVN: I just re-read the instructions, I downloaded the wrong version of the DLL. Thank you!

[cirdan]

Rebased against current trunk. Also, when you added a signal to a bridgehead and a train was already running on the bridge, the signal showed green initially, instead of red; this is now fixed.

[Supercheese]

Here's a binary made with the same process as the rest.

[AlexG]

Hi

Great job with the latest fixes

I just have a small problem while running a new game.
The problem I've posted here, at the bottom of the topic (page 30)
http://www.tt-forums.net/viewtopic.php? ... &start=580

Thanks

[cirdan]

I cannot imagine how my patch could cause this. Did you post here because you are using Supercheese's binary? If so, I guess it must be related to the other patch he used in the binary, which raises the maximum number of NewGRFs in a game, or perhaps there is a conflict between the NewGRFs in your game (and having more NewGRFs increases the likelihood of such a conflict).

[AlexG]

I've posted because some members said it could be the problem with the ottd version that I'm using, so I've tried my luck here too

[cirdan]

That is fine; if you are using my patch, or a binary published here, you should post your problems here, too. What I wanted to know is whether you compiled my patch yourself or you were using Supercheese's binary. I find it extremely unlikely that your bug is caused by my patch (which does not change anything related to industries); and, if you are using Supercheese's binary, I will let Supercheese comment on your bug, if he thinks that it may be related to the other patch he used in the binary.

My impression, however, is that you have some sort of NewGRF interaction. Have you tried to reproduce your bug in vanilla trunk?

[AlexG]

Sorry for the late reply...
I think it is solved
I'm using the binary and started a new game in "play heightmap" instead of "play scenario". I think that was the problem
I'll see these days what's happening...

[Supercheese]

Yeah, the only change I make in the binaries I post here other than using the simple .diffs posted is to change:

Code: Select all

	/** Last slot usable for (New)GRFs used during the game. */
	LAST_GRF_SLOT  = 63,
	/** Maximum number of slots. */
	MAX_FILE_SLOTS = 64

to

Code: Select all

	/** Last slot usable for (New)GRFs used during the game. */
	LAST_GRF_SLOT  = 127,
	/** Maximum number of slots. */
	MAX_FILE_SLOTS = 128

in fios.h.

The issue as described was due to changing industry grfs in a running game. There was a red warning somewhere that hinted to the potential for issues.