Oolite trunk, strange stuff

[Screet]

Hi,

is this the official bugtracker for the trunk version?
http://developer.berlios.de/bugs/?group_id=3577

If so...it does list the "ships with turrets explode" bug for fuel stations, thargoid hives and such.

However, with my recently built trunk version, that bug is history.

Sadly, it's replaced by player ship turrets not working anymore which I cannot find on that list. I wonder if those two are related, as with 1.72.2 a Caduceus can be made "fuel station proof" by manually removing the turrets. With the trunk version, the turrets are still visible, but they do not rotate towards their targets anymore, nor do they fire.

Ohhh...and something with the view positions has changed: the side parts of the Caduceus have not been visible when viewing port/starboard sides in 1.72.2. With the trunk version, they do - and they also do this with very different looking parts of the ship, although I would expect that both sides would look exactly the same. Does anyone know what happened there and if it does stay this way? If, for example, the neolite ships are to be verified and fixed for view positions because of visible ship parts with the 1.72.2, that's useless if that leads to different results in 1.73...

Screet

[another_commander]

Screet, I just checked the turrets with a Griff Boa and they work perfectly to spec, like they always did. No idea why it does not work for you. Turrets on the trunk are fine.

[Screet]

Screet, I just checked the turrets with a Griff Boa and they work perfectly to spec, like they always did. No idea why it does not work for you. Turrets on the trunk are fine.

Mine simply don't move...they sit still and do nothing. Gotta find a test commander file with the boa and check that one, maybe the turrets are a little bit different between Caduceus and GriffBoa?

I've also verified that a cloak bug from 1.72.2 also exists in trunk version: If the cloak is enabled and the command then begins a new game or is shot down, the next launching commander has a cloaked ship without being equipped with a cloak. That "cloak" also does not drain energy.

Screet

[Screet]

Really, really strange:

Tested the Caduceus again - nothing from the turrets, not even moving.
Tested the GriffBoa: turrets working
Did load (without restarting the game) the Caduceus: turrets disabled
Did restart the game: Caduceus turrets suddenly working?!? They almost shot down a coriolis, but then there was "target lost" without a destroyed coriolis. Did retarget it, they began firing again and that time I was shot down *GRR*

Next launch of the trunk version:
Caduceus turrets not working
GriffBoa turrets working
Caduceus turrets not working

Could it be that the addition of max flight pitch/roll for the Caduceus turrets /most times/ (but not always) disable the turrets?

And, yes, I did have active targets all the time. Both stations and hostile (red) ships...and yesterday I tried it quite a few times, nothing from the turrets then.

Screet

[another_commander]

OK, I can confirm that there is a turret problem with the Caduceus, but not with the Griff Boa or the Falcon. This may have to do with subentity handling differences between 1.72.2 and 1.73. And yes, it needs fixing, but at least the actual turret code is OK and the problem could be either in something that needs to be changed in the OXP or something to do with changes in subentity positions calculations (at least that's what I think, but it's ages since I last looked at that part of the code).

Also, I note that the Caduceus OXP somehow messes up the shipyard. The ship lists contain lots of similar ships, while the Caduceus itself does not appear for sale even with a chance of 0.99 and techlevel set to 1. This happens in 1.73, but not in 1.72.2. Hmm...

[another_commander]

I've also verified that a cloak bug from 1.72.2 also exists in trunk version: If the cloak is enabled and the command then begins a new game or is shot down, the next launching commander has a cloaked ship without being equipped with a cloak. That "cloak" also does not drain energy.

This one is now fixed in SVN 2108.

[Kaks]

I'm having subentity problems with the seedy bar too.

Screet, this might not be what you had in mind originally, but if you could figure out why some turrets work & others don't it would be greatly appreciated!

[Screet]

I'm having subentity problems with the seedy bar too.

Screet, this might not be what you had in mind originally, but if you could figure out why some turrets work & others don't it would be greatly appreciated!

Well, bug hunting is pretty much what I was always good in, from a programmers point of view But I'm not familiar with the code yet, so I could use some hints where to look at...

I tried it the other way around, looking at the differences of the turret setup...but it wasn't the pitch/yaw/roll limits specified in the Caduceus file which are not present in the GriffBoa, it's something else.

Screet

[Kaks]

That part of the code is new to all of us: the new subentities code was submitted in revisions 2051, 2052, 2053 on the 11th of February by Ahruman, who has been busy for the past 2 weeks.

I haven't had a chance to look inside it, and I don't think a_c has either...

[Screet]

Hi,

looks like the turret part of my message did attract people to comment...but nothing on this issue here:

Ohhh...and something with the view positions has changed: the side parts of the Caduceus have not been visible when viewing port/starboard sides in 1.72.2. With the trunk version, they do - and they also do this with very different looking parts of the ship, although I would expect that both sides would look exactly the same. Does anyone know what happened there and if it does stay this way? If, for example, the neolite ships are to be verified and fixed for view positions because of visible ship parts with the 1.72.2, that's useless if that leads to different results in 1.73...

Since Simon asked for help and I thought it might be easy enough for me, I said that I'd try to check and fix the models view positions and exhausts. However, I have been to busy yet

What makes me worry with this problem here is that if I fix these things for 1.72.2 and then we update the game, I might have to fix them again...or if I fix the models for 1.73 then that won't properly fix things for 1.72.2.

Any ideas/suggestions? Pleaaaaase!

Screet

[another_commander]

As Ahruman has said in the past, worrying about compatibility with previous test versions should be a low priority concern. Once 1.73 gets out, nobody will care too much about being compatible with 1.72.2. That's why no official stable has been declared yet. Things move and evolve and changes do happen and will happen until we reach a point that we will be happy to call Final.

Having said that, the best fix would be the one that makes things work on both new and old versions.

[Screet]

Having said that, the best fix would be the one that makes things work on both new and old versions.

The problem is, that the viewpoints are somehow handled differently in both versions:

The Caduceus looks fine for port/starbord view in 1.72.2, but it did show massive bugs when playing 1.73 - and those were not even the same for port and starboard although one might expect that a symmetrical model would show the same faulty ship parts in both views.

Screet

[another_commander]

OK, I see what you mean. I actually just noticed the port/starboard views with the Caduceus. There is indeed something funny looking in 1.73, as if the model is being viewed from the inside. I wonder if the non-firing turrets of the Caduceus are related to that.

That would probably require further investigation.

[Screet]

OK, I see what you mean. I actually just noticed the port/starboard views with the Caduceus. There is indeed something funny looking in 1.73, as if the model is being viewed from the inside. I wonder if the non-firing turrets of the Caduceus are related to that.

That would probably require further investigation.

I've just begun to try tracking this down from the code side....miss my development environment with breakpoints and such. Is there any for this one?

Maybe I already found something. As I could not set breakpoints, I did add a log message here:
- (BOOL) setUpOneStandardSubentity:(NSDictionary *) subentDict asTurret:(BOOL)asTurret

in the activation for the turrets. The log was filled with entries for each turret of several ships - but not mine!

Now it looks like I need to track who's calling this, go there, log there, and so on...unless someone with more experience for this code or a better idea to track this can jump helpfully in.

However, this does not appear to be linked to the view point problems which came with v1.73 for the Caduceus. It appears to be something else...

Screet

[another_commander]

For debugger, there is gdb. Command line, rough and takes no prisoners. If you prefer a GUI for this, I would recommend Insight, which seems to utilize the functionality of gdb very well. However, I must say that I don't have much experience with it. I normally do it the hard way when debugging (using NSLog and fprintfs to stderr extensively).