Hard Light Productions Forums

Hosted Projects - Standalone => Wing Commander Saga => Topic started by: ARSPR on November 07, 2007, 12:48:42 pm

Title: A possible bug related with autopilot
Post by: ARSPR on November 07, 2007, 12:48:42 pm
Well I've found a possible bug related with autopilot, but as I'm not sure, I prefer to post it here before mantising.

I've reinstalled WCS Demo (after a PC change). It is installed in its own folder outside FreeSpace2 so no possible conflicts here. As game exe (wcsaga.exe) is a pretty old bug, I've picked latest Launcher (5.5c) and Fsopen build 2007-10-28 Xt build (http://www.hard-light.net/forums/index.php/topic,50258.msg1015671.html#msg1015671), and I'm going to play play this MOD through these latest exes.

I've found the problem in the second mission. When you approach the TCS Aurora, and you initiate the autopilot jump with original wcsaga.exe your speed is kept matched at Aurora's one (22) so more or less you end the run in formation with the TCS.

But with 2007-10-28, your speed is set at 25 so you end the autopilot run quite far from Aurora.

I upload screenshots of both situations (although they are not really needed to see the problem).

I know that the autopilot code has received a lot of modifications since wcsaga.exe. So, I don't know either if we have a bug now or if there was a former solved bug which also needs an updated mission.



[attachment deleted by ninja]
Title: Re: A possible bug related with autopilot
Post by: Tolwyn on November 07, 2007, 01:38:49 pm
I am pretty sure it is the result of the recent autopilot modifications. The only thing you can do is following: extract all missions from the VP, open them in FRED and check "use autopilot sequences". I can't tell how well they will perform, but at least it is the only way to avoid the problem you described (and you save a bit of travelling time!)
Title: Re: A possible bug related with autopilot
Post by: ARSPR on November 07, 2007, 03:01:17 pm
I am pretty sure it is the result of the recent autopilot modifications. The only thing you can do is following: extract all missions from the VP, open them in FRED and check "use autopilot sequences". I can't tell how well they will perform, but at least it is the only way to avoid the problem you described (and you save a bit of travelling time!)

Thank's Tolwyn, you've confirmed what I thought but ...  :nervous: ... actually ...  :nervous: ... you haven't answered my question  :lol:

Which is the right option?:

Title: Re: A possible bug related with autopilot
Post by: Topgun on November 07, 2007, 03:12:43 pm
]It's a bug but it isn't worth fixing it. This way of generating autopilot travel is depreciated, (more or less like D3D). Even if it is buggy, nobody is going to fix it

I think.
Title: Re: A possible bug related with autopilot
Post by: Tolwyn on November 07, 2007, 03:14:21 pm
It is obviously a bug - why in havens would we want the transport to lag behind? :)

It is very unlikely that the patched version of the Prologue will use the old autopilot system, yet I would go ahead and put a report into mantis. As far as I can tell BtrL is using the autopilot system too.
Title: Re: A possible bug related with autopilot
Post by: ARSPR on November 08, 2007, 05:41:18 am
Ooops, I've taken a deeper look and I'm pretty sure it's not a build bug but a bad table setting. (Remember I'm not modder at all)

The Amadeus-Class seems to have incoherent table entries:

Code: [Select]
$Name:                        Amadeus-class
...
$Max Velocity:           0.0, 0.0, 25.0
...
$Max Oclk Speed:         23.0

So, reading what Wiki says, an Amadeus class transport goes faster at normal engine energy settings than at maximum settings  :wtf: ...

It seems that the new Autopilot code makes you travel at the slowest $Max Velocity of all set-nav-carry ships. But with the original ship.tbl, you are actually travelling faster than Aurora capabilities. (Maybe old autopilot code took the MIN between both $Max Velocity and $Max Oclk Speed, so it hid this table error).

I've swapped $Max Velocity (23.0) and $Max Oclk Speed (25.0) and then it runs fine. With these new settings, what happens in this mission is:
+ At the beginning, Aurora is flying at 15 because of the cap-waypoint-speed SEXP.
+ When Aurora recieves the cap-waypoint-speed = -1 SEXP, it accelerates till 25.0.
+ Then, you initiate the autopilot run and every single ship travels at 23.0.

I upload a patch table, but please fix it in the final WCS release.

[attachment deleted by ninja]
Title: Re: A possible bug related with autopilot
Post by: Tolwyn on November 08, 2007, 09:03:51 am
I don't have to fix it, as it was fixed months ago in my tables. ;) And, like I said, this version of autopilot is obsolete anyway.
Title: Re: A possible bug related with autopilot
Post by: ARSPR on November 14, 2007, 08:42:48 am
Another small issue that nearly makes me mad until discovered what happened  :mad:

In Demo-04-BG-Hermes.fs2 (The fourth mission) you've got this SEXP whenever you reach an autopilot location:

Code: [Select]
$Formula: ( when
   ( <
      ( distance
         "Alpha 2"
         "Waypoint Nav 1:1"
      )
      999
   )
   ( add-goal
      "Alpha"
      ( ai-stay-near-ship "Alpha 2" 89 )
   )
)

It is pretty simple: whenever you reach a nav, make Alpha wing stay near Alpha 2 (player ship). Well, if you play the WCS Demo with 2007-10-28 build, and I suppose because of the autopilot changes, this SEXP doesn't work, alpha wing doesn't get that order and it doesn't follow you. In order to make it work, you need to put 998 instead of 999 in the distance check. So it seems that the autopilot run, (while the ships doesn't seem to allow ai orders), is slightly longer now.

Nevertheless, if you use cinematic autopilot, then it works fine even with 999.


OTOH, please take a look at Mantis 1520 (http://scp.indiegames.us/mantis/view.php?id=1520) if you can. (I discovered it while trying to ascertain what was happening with the ai-stay-near-ship issue).
Title: Re: A possible bug related with autopilot
Post by: Tolwyn on November 14, 2007, 08:47:28 am
I usually use lt 5000 as distance check.
Title: Re: A possible bug related with autopilot
Post by: ARSPR on November 14, 2007, 10:03:35 am
I usually use lt 5000 as distance check.

That make things worse if you don't use cinematic autopilot. Giving orders just seem to work with a 998 or less distance check.

It's a minor bug but I've mantised it (1521) (http://scp.indiegames.us/mantis/view.php?id=1521).
Title: Re: A possible bug related with autopilot
Post by: Tolwyn on November 14, 2007, 10:21:03 am
I usually use lt 5000 as distance check.

That make things worse if you don't use cinematic autopilot. Giving orders just seem to work with a 998 or less distance check.

It's a minor bug but I've mantised it (1521) (http://scp.indiegames.us/mantis/view.php?id=1521).

nah, you just need to chain orders with one second delay to the "proceed to nav x" event ;)
Title: Re: A possible bug related with autopilot
Post by: karajorma on November 14, 2007, 03:59:25 pm
It is very unlikely that the patched version of the Prologue will use the old autopilot system, yet I would go ahead and put a report into mantis. As far as I can tell BtrL is using the autopilot system too.

We've not used the old autopilot system, only the navpoints from it thus far. Any new missions using autopilot will use the new system most likely.