Oolite Bulletins

Right, let’s see. *shuffles notes*

• Cube-mapped planets now have their poles in the right place. (I set out to fix this and found that it already worked. Don’t know when that happened.) Latlong textures are now wrong in a slightly different way than before, but I’m not very interested in fixing them; between the distortion, the misalignment, and the pinching, latlong textures have never looked very good. Use cube maps. (Bug #17243)
• Fixed the Mac sound bug where it would occasionally stick a knife into your ear and wiggle it around until your brain started running out of your nose. (Bug #17214)
• New ship script event shipTakingDamage(amount, fromEntity, damageType). Amount will be zero if damage is absorbed by shields. DamageType is the same as for shipDied(). The event is triggered after the damage is subtracted from the ship’s energy but before checking if it’s a kill, so energy may be <= 0 (and you can resurrect a ship by giving it energy in this situation). It should be called for any damage type except “removed”. (Feature request #5077)
• As far as scripts are concerned, the player can now only die once. (Previously, shipDied() would be called twice, once without parameters and once with.)
• When the player is killed by a q-bomb, the damageType parameter to shipDied() is now “cascade weapon” (not “energy damage”) and the player triggers a new cascade.
• Sounds that are listed in customsounds.plist are now preloaded to avoid loading delays. (The code was there before, but was never invoked.) Sounds that are played through JavaScript by specifying a file name cannot be preloaded; if you’re doing this, and your sound isn’t particularly rare, I suggest going through a customsounds.plist entry.
• New Entity method dumpState(). Same type of dump as pressing 0 while paused, or the super-secret shift-H key (I’d completely forgotten about that one).
• The JavaScript global and console objects are now replaced when resetting the game, and can no longer be used to smuggle information across sessions.
• Reduced memory usage of various special effects and JavaScript objects.
• Windows builds now support the JavaScript trace and profile commands (see upthread; only native functions at this time), and also the ability to turn Objective-C exceptions into JavaScript exceptions instead of crashing horribly.
• Updated to FireFox 4b10 version of SpiderMonkey, and dropped support for SpiderMonkey 1.7. Boring for you, fiddly for me, but I got to check in -121970 lines of code in one go. Nice big drop in the LoC graph.
• Bugs #17202, #17214,#17215, #17243, #17391,#17454, #17819, #17888, feature requests #5058,#5077,#5095 and task #3373 closed. Feature request #5031 deferred until after MNSR.

You had a quiet week then, admiral.

Well, not just me. Those other fellahs helped around the place and whatnot, dontchewknow.

Ahruman wrote:

Right, let’s see. *shuffles notes*

• Fixed the Mac sound bug where it would occasionally stick a knife into your ear and wiggle it around until your brain started running out of your nose. (Bug #17214)

I love you!

During this last week, instead of hanging around on the boards, I've started actually playing the game again, and I decided to do it with sound on (which I usually have turned off). And I really can't tell you just how often my brain started running out of my eardrums! At times it happened once a minute! There was one instant when I was seriously afraid that my laptop's speakers could have been actually damaged, because after one particularly vicious sound attack all sounds were muffled. I had to restart the computer to be able to listen to music again.

Even if there were nothing else at all, Oolite Mac 1.75 is going to be a major breakthrough just for fixing this particular bug!

I love you!

Awwwwww

TGHC wrote:

I love you!

Awwwwww

well, we are the friendliest Board etc, don't you know!

Ahruman wrote:

[*] Fixed the Mac sound bug where it would occasionally stick a knife into your ear and wiggle it around until your brain started running out of your nose.

Oh hooray! I remember playing Oolite with in-ears once, and well, you can imagine what happened.
Great job!

Ahruman wrote:

Cube-mapped planets now have their poles in the right place. (I set out to fix this and found that it already worked. Don’t know when that happened.)

Mystery solved: Eric fixed it in November, when I wasn’t paying attention. A bit of an anticlimax to find that my last major task was already done, but that’s a disappointment I can live with. ;-)

Ahruman wrote:

Mystery solved:

Yep, I only wrote something about changed rotation axis in the comment because the mapping itself was always correct for the cube maps. The sun now keeps shining above the main planets equator. Nothing should keep texture makers to add planets with ice caps now

-----
Update for thargoid containing oxps

The npc missile launches always had an interesting 'feature' that when the missile_launch_position was defined inside the bounding box, the missile got 10 m/sec extra speed for every meter it was from the outside of the bounding box. This resulted in launch positions close to the ships center having a very strong extra speed boost. Examples are the thargons from the thargoids that were thrown out at high speed, for the internal thargoids, while ox's that used a more correct position, like the Griff ones, didn't have that extra speed.

Somehow that looked wrong. And now it seems that even the direction was wrong: not away from the ship, but always along the same system vector, no matter of the position of the ship. The only thing preventing missiles from crashing was additional code that made missiles crash-proof during the first 0.25 seconds of their existence.

Anyhow, this code is now removed, resulting in a new problem: the missile_launch_position of thargoids was defined exactly behind the ship at the same position as the aft_eject_position. This will cause thargons to immediately crash into their mother. I examined many ox's with thargoids and all added the thargons wrong. The only exception are Griff his thargoids.

Trunk will now check for such wrong launch positions at [0,0,-x] and replace it with a default position. It will also write an error to the log to inform that the position was wrong, so that oxp makers know that they have to update their work.

Gold star for Griff!! *smug face*

Griff wrote:

Gold star for Griff!! *smug face*

Have five team points as well.

Custom docking ring models are no longer supported. The implementation of this feature was unsafe, and relied on old model loading code that is no longer used for anything else.

As a partial replacement, the new implementation uses a distorted regular polygon defined by three station properties: tunnel_corners (ranges from 4 to 128), tunnel_start_angle (in degrees; 0 means the first corner is straight up) and tunnel_aspect_ratio, which is the proportion of the tunnel’s width to its height.

The default values are 4, 45 and 2.67, which produces the same rectangular appearance as in 1.74. The rock hermit uses 6, 0, 1 for a hexagon.

Ahruman wrote:

As a partial replacement, the new implementation uses a distorted regular polygon defined by three station properties: tunnel_corners (ranges from 4 to 128), tunnel_start_angle (in degrees; 0 means the first corner is straight up) and tunnel_aspect_ratio, which is the proportion of the tunnel’s width to its height.

No triangular docks, then?

If you really, really want them, I can lower the limit. :-)

Edit: Actually, I lowered it anyway because purely arbitrary limits are silly.

Be like a giant shape-sorter for Boa Class Cruisers