FreeOrion

new tech tree files still produce this effect:

Are you actually using revision 3274, or are you just using my most recent content files?

_________________
Warning: Antarans in dimensional portal are closer than they appear.

i received 3274 exe's. thanks to Geoff.

another thing: here's my third turn

why i can see starlane marked with red X? IMO i cannot have any info about it yet...

fourth turn:

1. ship, marked red O is moving, but it's moving-dots path not showing.
2. why i cannot scroll the galaxy in direction marked with red arrow?

fifth turn:

1. moving-path for ship marked red O is now showing, but:
1.a where ship moving not showing
1.b ship doesn't move a bit during this turn
1.c ship's destination seems changed
2. star systems Phad and Penrose seems disappeared with all their starlanes

saves attached

It does not for me! Make sure you have indeed the latest content files.



Something else is odd there, you do know the system names, although you haven't been there. What buildungs did you produce?


This is not related to the current build, 0.3.13 behaves the same way.

Are you *sure* you're using it? 3274 shouldn't be able to produce that colonize button text, no matter what stringtable you're using...

Looking at your turn 4, I'm almost sure you're not using the 3274 exes I sent, since you've lost vision of the Dosis system that you could see on turn 3.

Edit: as for the scrolling, there are some weird limits to it that only seem to apply horizontally. I don't know why and haven't looked into it. /Edit

oops... seems like i unzipped exe's but not copy them to FO directory

i'm sorry. delete this topic

Now that it is here, I'd like to mention that fuel supply is not channeled through systems, that are not in range of a detector. It's obvious why this happens: the client doesn't know about the system, he just displays what he previously knew. Geoff is surely aware of this, though

ok. let it stays
here's my fifth turn, surely in 3274

marked ship moving to unknown system, but path-dots are not there
for me it's not very obvious. why client cannot transfer fuel while it can draw system and starlanes? what else knowledge it needs?

and by the way! marked ship doesn't really moving anywhere

saves attached

You might have noticed, that the implementation of stealth, detection and especially the "memory system" is new... and incomplete. Certain things, that worked previously need to be adjusted.


This has been mentioned by Geoff himself(at least in the mail, he sent to me).

There are a few related issues for fuel distribution. Whether the client displays fuel distribution dotted lines to a system and whether the server actually provides fuel to ships there are separate questions. What the client does doesn't determine what the server does, though usually the same code determines what both the client and server will do and display, respectively. In this case, fuel is probably being distributed according to visibility of systems, and possibly whether the system has been explored (ie. a ship has been in or passed through it), not whether starlanes connection systems are known. Likely the fuel distribution should be just based on knowledge of starlanes, although in the event we have starlanes that can be removed, knowing a starlane exists and thus seeing fuel being distributed to a system along that lane in the client won't actually cause that distribution to happen on the server.

As for the move lines, the issue mZhura is pointing out is one of his own ships, which should have the move lines be showing. The point in the email I mentioned to one or both of you was about enemy ships' move lines not showing ever.

mZhura might also be trying to say the ship doesn't move between turns, which would be a separate issue. It's not clear from the posted message, though. If that's the case, did you load a save that was created with the previous version to produce this problem, or did you start a new game? If you loaded an old version save (if that's even possible?) can you reproduce this with just the new version?

for now fuel distributed only through fully visible systemsso, there is no way to tell client that starlane ceased to exist?about enemy ships it seems logical not to know where it is heading to, just where it is in the current moment, but that is my ship, not enemy.exactly. ship just get stuck in the middle of starlanenew game of causesave in my previous post is of 3274-generated game

it seems to me that i found the cause of ship freezing in starlane. it seems like it always happen when ship enters starlane with less then 1 fuel and doesn't reaches the end of starlane on the next turn.

Fixed in SVN.


Right now, there's no way for starlanes to cease to exist... so it's a non issue.

If we added a way for starlanes to cease existing, as of now, the rest of the game would treat that like any other information known about objects that are destroyed that a player can't see when they're destroyed; the player would continue to think there's a starlane there, until the player got vision of a system attached to the starlane and thus became aware that it wasn't there anymore.

The game mechanics could be extended so that players would be informed if any starlane is removed that they'd previously observed but currently don't have vision of... but this doesn't seem like the appropriate thing to do given that I just spent / am still spending all this time adding a sensible visibility system...


How things current work, or are supposed to work, is that a player that sees an enemy fleet moving between systems is only told of the next system the fleet's moving to, and nothing about where it's headed after it reaches that system.


That was indeed relevant information. I'd recently replaced a bunch of functions that returned objects with functions that returned object IDs. If an object didn't exist, the functions used to return a null pointer, equal to zero, which is often checked for by something along the lines of

Which only executed the ... code if there was no returned system. GetSystem was replaced by SystemID, which returns a special object id, -1, when there is no object, so the resulting code said

which executes the ... code if any ID other than 0 is returned, which is the case when there is no object, since -1 is the returned sentinel value to indicate that there's no object.

Specifically what was happening was that the move path calculating function was checking if the fleet had no fuel and was in a system, and if so, was aborting early and returning an empty path, since the fleet can't leave the system without fuel. With the change, fleets that weren't in a system and which had no fuel matched this check, since their system ID was -1 to indicate not in a system, and so these fleets had an empty move path returned, and thus didn't move.

does it needs rebuilding of exe's?

Yes.

Given the lenthy code-discussion in my previous post, I'd have thought that was obvious...

yes, that was. my question appears because i didn't receive fixed exe's from you sorry to bother, but i need your help to continue testing