Unable To Launch

posted on September 16th, 2006, 8:18 pm

The steps:
I did not have ST:Armada2 installed.
I ran the 3.0 installer while the cd is in the drive, it seemingly installed fine.
Ran the 3.0 patch 1 installer, seemingly installed fine.
If I run the shortcut that was installed, or the exe directly, or what-not, it loads, brings the icon on top of all, fading it, using about about 7% cpu while fading in, then no cpu usage, no cd drive spin-up, nothing.
After it sits for about 11-12 seconds, it then disappears, the process is removed from the task manager, it just completely disappears.

Same thing happens if I use your launcher and click the launch button, icon comes up, nothing happens, then it disappears.

Thoughts?


System:
AMD Athlon 3400+
1GB Ram
4 ATA Controllers
2 SATA Controllers
1 SCSI Controller
~1.3 TB HD space total, spread over 6 hard drives, free space is ~4GB
1 CD-recorder
1 dead DVD-Reader
ATI Radeon 9700 Pro
SB Audigy 1 Platinum
And a couple more un-mounted drives (Tape drive, an old scsi HD, etc..., un-mounted meaning they are installed and usable in the system, just not mounted in Windows as drives, I'm out of letters...)

[Cpt Ryan]

posted on September 16th, 2006, 9:36 pm

Last edited by Cpt Ryan on September 16th, 2006, 9:38 pm, edited 1 time in total.

your using the actual CD for the launch, ie its in the drive when you try it?
is the CD your using a copy or genuine copy?

EDIT: the drive you are using is an actual drive mounted to the system or one of your other external ones?

[OvermindDL1]

posted on September 16th, 2006, 10:06 pm

Last edited by OvermindDL1 on September 16th, 2006, 10:07 pm, edited 1 time in total.

Actual CD, bought soon thereafter ST:Armada 2 came out.
It is genuine copy, printed cd, instruction book, probably still have the box around here as well, CD is in perfect condition.
All of my drives are internal to the computer (the motherboard has all of the above mentioned controllers embedded on it, they are not expansion cards, yes it was an expensive board, but cheaper then buying all the controllers seperatly). The CD drive is on ATA Controller 0, Secondary Master.

posted on September 16th, 2006, 10:19 pm

try uninstalling and reinstalling, so what happens

posted on September 17th, 2006, 3:17 am

that issue should have been gone with patch 1.1. could you please attach the FODataTempLogLoader.txt here?
thx

DOCa Cola

[DOCa Cola]

posted on September 17th, 2006, 3:27 am

Last edited by DOCa Cola on November 17th, 2006, 11:13 pm, edited 1 time in total.

The log is below, I already did uninstall and reinstall once before, will try again if you wish. I may have a little time this evening, I will trace the executable and see if I notice anything overly odd if I can. Out of curiosity, what does your executable do exactly? I notice it links winsock, does it handle mini-dump uploading or something?

Code: Select all

[09/16/06 - 21:23:50] Fleet Operations Loader Log - initialized
[09/16/06 - 21:23:50] (c) 2003-2006 Fleet Operations Development Team - http://www.fleetops.net
[09/16/06 - 21:23:50] Pre Launch - Full Parameters list:
[09/16/06 - 21:23:50] Pre Launch - Initializing cursors stage 1 (non zero value expected): -1
[09/16/06 - 21:23:50] Pre Launch - Initializing cursors stage 2 (non zero value expected): -1
[09/16/06 - 21:23:50] Pre Launch - Initializing cursors stage 3 (non zero value expected): -1
[09/16/06 - 21:23:50] Pre Launch - Initializing cursors stage 4 (non zero value expected): -1
[09/16/06 - 21:23:50] Pre Launch - EditorMode False
[09/16/06 - 21:23:50] Pre Launch - Chosen Splash FODataSplashSplash1.bmp
[09/16/06 - 21:23:50] Pre Launch - Splash effect enabled
[09/16/06 - 21:23:51] Process launch - Version number refreshed
[09/16/06 - 21:23:51] Process launch - Path to executable G:Program FilesActivisionStar Trek Armada II Fleet Operations
[09/16/06 - 21:23:51] Process launch - Executable Command line "G:Program FilesActivisionStar Trek Armada II Fleet Operationsarmada2.tgt"
[09/16/06 - 21:23:51] Process launch - Path to executable exists -1
[09/16/06 - 21:23:51] Process launch - Reported Process ID 1568
[09/16/06 - 21:23:51] Post launch - 7315536/1568/292/-1
...
[09/16/06 - 21:23:51] Post launch - 7543696/1568/292/-1
[09/16/06 - 21:23:51] Post launch - 7543724/1568/292/-1
[09/16/06 - 21:24:02] Process launch - Opening Helper file stream handle@ reported ID 292
[09/16/06 - 21:24:02] Loader end processes - Closing stream handle reported (non zero value expected): -1
[09/16/06 - 21:24:02] Loader end processes - Resetting setcursors reported (non zero value expected): -1
[09/16/06 - 21:24:02] Loader end processes - Executing self termination


EDIT 1: Okay, this is funky wierd, I *know* I see traces of Visual Studio (7.0, maybe 6.0, but I think 7.0, really should upgrade to 7.1 at least, *far* more complient), yet I see so much of delphi layout as well, causing a lot of hopping around (hits the cache rather hard, should not do it, although the program does not do much so not really necessary to worry about). There seems to be a *massive* amount of unused code path's and resources, bloating the code far larger then what is necessary. Although some of these might be accessed through function pointers rather then direct links, but that is quite rare. Still have not yet run it, may not have time...

EDIT 2: Traced a bit, seems to churn something 5000 times, once done, performs more things (seems to call destructors of class-like structures to clean up things, like the window) and quits, digging more into that loop (don't suppose you could say what it is doing, maybe checking for armada2 startup or something?)...

EDIT 3: (Mostly my own note taking):
In the loop, which exists inside the exe and not a linked library:
It calls a function in user32.dll (to be figured out, 0x4AF818).
Then makes sure the return value is not zero (which it always is thus far, would jump away if it was not zero at this point).
Then calls a function inside the application itself (seemingly to make up for the missing library function), this function then jumps into kernel32.dll (0x7c802451)
After the kernel function returns it then jumps back and repeats the loop, no test after calling the kernel function, the kernel function probably sleeps or something (to be checked, now checked, SleepEx called with a value of 1).

EDIT 4: Kernel function checked, it is SleepEx, set above, checking the user32 function now...

EDIT 5: Have not yet figured out what the user32 function is that is called, but the purpose of this is that I notice that the log is in the middle of its "Post launch" sections when I terminate the trace inside the loop...

EDIT 6: Got more time to look, the function in the linker map for user32.dll is at location 0x1E581, which corresponds perfectly to the ASCII version of FindWindow (FindWindowA), and the params passed match as well (pass a null pointer for the class name and pass a pointer to a string {"Armada2"} for the window name param). So, the 5 second long loop tries to access the window named Armada2, if not, it sleeps for a millisecond (which will not actually sleep, just toss the context switch around to give other processes cpu time if need-be), and then loops. If not found within '5' seconds (in quotes because, as stated, the context switch will just be thrown around, possibly taking more then 1 millisecond, sleep is highly unreliable when used as timing on a granularity smaller then ~24ms, so it usually takes between 10-12 seconds in reality), it dumps out, cleans up everything, and dies. I wager I will find CreateProcess sometime before the loop is called, havn't checked, probably where the problem is however, will check later, I only had a minute to spare here...

EDIT 7: Before I need to go, just remembered something... Why is the window handle being looked for, and in such an active mannor as well? Only reason to need the window is to inject events directly into to. If you are just going to watch for when the application closes then you can get the process handle from CreateProcess when you launch the app, then just put a lock on it so the launcher freezes until the app closes, then when the app closes the launcher continues on as if it never stopped, just now instantly knowing the app is gone (handle is now invalid). Using that handle, before you put a lock on it or what-not, you can do a wide variety of functions on it, far more power then just a window handle, even injecting events into the process itself (rarely a reason to inject directly into a specific window and not an entire process). I have to admit that I am a tad bit confused as to why FindWindow is being called at all? Either way, need to go, talk later.

[DOCa Cola]

posted on September 17th, 2006, 10:50 am

Last edited by DOCa Cola on September 17th, 2006, 11:01 am, edited 1 time in total.

the findwindow call is there for various reasons. the pr 3.0 patch 1.1 executable consists more of a bit unfinished code as it includes some code meant for the final 3.0 release.
findwindow is used to wait for the main armada 2 window to know when the original armada 2 splash screen has closed. also to change the armada 2 main window icon and description. so actually the findwindow loop should stop after the main armada 2 has shown up.
yes, i am using delphi 2005. the bloated code may be related to vcl i used to build the loader.
also the loader includes a (not self written) exception handler using winsock to mail an error report to us.

oh and please feel free to contact me via instant messenger

[OvermindDL1]

posted on September 17th, 2006, 9:36 pm

Last edited by OvermindDL1 on September 17th, 2006, 9:36 pm, edited 1 time in total.

Ah, that makes sense, although I would just listen to the windows being created and got a handle on it then (this is an async operation so you can still have the main part of the program block on the armada2 process until it closes and still do everything to the window that you want so as to not have a comparatively expensive loop). I added you to msn, I basically never use it, my main is GTalk, but it is up currently.

[DOCa Cola]

posted on November 17th, 2006, 3:10 am

Last edited by DOCa Cola on November 17th, 2006, 11:12 pm, edited 1 time in total.

Hello,

I didn't want to create another thread because I am also unable to launch the game.
I installed:

fosetup3pr.exe
FleetOps3PRPatch1.exe

Then I loaded the game. I see the starting logo and then my computer totally freezes. I think things are still working in the background, but I cannot switch to them. I cannot access the task manager either.

My specs:

Core 2 Duo E6600 - no overclocking
DFI Infinity 975 X/G
2GB of RAM
EVGA Nvidia GeForce 7600 GT - dual monitor (also tried single, so doesn't seem to be the problem)
2x WD Raptor 74G RAID 0
AMCC-3ware 9590SE Raid 5 card w/ 8xWD 400G


here is my logs:

[11/16/06 - 21:55:56] Fleet Operations Loader Log - initialized
[11/16/06 - 21:55:56] © 2003-2006 Fleet Operations Development Team - http://www.fleetops.net
[11/16/06 - 21:55:56] Pre Launch - Full Parameters list:
[11/16/06 - 21:55:56] Pre Launch - Initializing cursors stage 1 (non zero value expected): -1
[11/16/06 - 21:55:56] Pre Launch - Initializing cursors stage 2 (non zero value expected): -1
[11/16/06 - 21:55:56] Pre Launch - Initializing cursors stage 3 (non zero value expected): -1
[11/16/06 - 21:55:56] Pre Launch - Initializing cursors stage 4 (non zero value expected): -1
[11/16/06 - 21:55:56] Pre Launch - EditorMode False
[11/16/06 - 21:55:56] Pre Launch - Chosen Splash FODataSplashSplash6.bmp
[11/16/06 - 21:55:56] Pre Launch - Splash effect enabled
[11/16/06 - 21:55:57] Process launch - Version number refreshed
[11/16/06 - 21:55:57] Process launch - Path to executable C:Program FilesActivisionFO
[11/16/06 - 21:55:57] Process launch - Executable Command line "C:Program FilesActivisionFOarmada2.tgt"
[11/16/06 - 21:55:57] Process launch - Path to executable exists -1
[11/16/06 - 21:55:57] Process launch - Reported Process ID 2036
[11/16/06 - 21:55:58] Post launch - 7315536/2036/504/-1
...
[11/16/06 - 21:55:58] Post launch - 7543724/2036/504/-1
[11/16/06 - 21:57:16] Process launch - Opening Helper f


That is exactly how the log looks. The Process launch line is cut off.

Also, One time I rebooted and my network driver was messed and windows tried to reinstall it. Might that be pertinent?

Thanks,
Janosek.

posted on November 17th, 2006, 7:14 am

strange, the log seems all fine to me until that cut off line..

you have to reboot your computer after starting fleet operations?

posted on November 17th, 2006, 2:02 pm

Correct. My computer becomes completely unusable. I see disc activity in the background, but I cannot multitask to kill the process. I tried a few more times to get it to work, but the same effect happens. I am starting to think it has something to do with networking because a couple times when I reboot, windows searches for a driver to my network card.

At first, I had trouble installing it, but I found out that was due to Lavasoft's Ad-watch blocking the install.

I then uninstalled my Anti-virus, seeing if that might affect it, but it didn't.

Are there any other logs that might help you? The original Armada 2 with patch 1.1 works fine on my computer at 1600x1200 resolution and multi-display.

[DOCa Cola]

posted on November 17th, 2006, 8:59 pm

Last edited by DOCa Cola on November 17th, 2006, 11:12 pm, edited 1 time in total.

uninstalled and reinstalled... still no go

[11/17/06 - 15:54:41] Fleet Operations Loader Log - initialized
[11/17/06 - 15:54:41] © 2003-2006 Fleet Operations Development Team - http://www.fleetops.net
[11/17/06 - 15:54:41] Pre Launch - Full Parameters list:
[11/17/06 - 15:54:41] Pre Launch - Initializing cursors stage 1 (non zero value expected): -1
[11/17/06 - 15:54:41] Pre Launch - Initializing cursors stage 2 (non zero value expected): -1
[11/17/06 - 15:54:41] Pre Launch - Initializing cursors stage 3 (non zero value expected): -1
[11/17/06 - 15:54:41] Pre Launch - Initializing cursors stage 4 (non zero value expected): -1
[11/17/06 - 15:54:41] Pre Launch - EditorMode False
[11/17/06 - 15:54:41] Pre Launch - Chosen Splash FODataSplashSplash1.bmp
[11/17/06 - 15:54:43] Pre Launch - Splash effect enabled
[11/17/06 - 15:54:45] Process launch - Version number refreshed
[11/17/06 - 15:54:45] Process launch - Path to executable C:Program FilesActivisionFO
[11/17/06 - 15:54:45] Process launch - Executable Command line "C:Program FilesActivisionFOarmada2.tgt"
[11/17/06 - 15:54:45] Process launch - Path to executable exists -1
[11/17/06 - 15:54:45] Process launch - Reported Process ID 1364
[11/17/06 - 15:54:45] Post launch - 7315536/1364/504/-1
...
[11/17/06 - 15:54:45] Post launch - 7543676/1364/504/-1
[11/17/06 - 15:54:45] Post launch - 75436

cut off as you see it...

[DOCa Cola]

posted on November 17th, 2006, 11:09 pm

Last edited by DOCa Cola on November 17th, 2006, 11:14 pm, edited 1 time in total.

no limited user account (no admin).
yes, the cutoff occours when the loader crashes or get crashed i fixed that in the newer builds of the loader.
anything in the armada 2 log (not the loader log)?
you could try to create a fresh user account and see if fleet operations works from there. i assume you are using windows xp?

i cut down the logs of the posts a bit so this thread can be easier read

DOCa Cola

posted on November 17th, 2006, 11:36 pm

HA! I created a new administrator account and it worked!

I suppose the question is what is interfering with the game in my primary account?

[DOCa Cola]

posted on November 18th, 2006, 8:58 am

Last edited by DOCa Cola on November 18th, 2006, 9:00 am, edited 1 time in total.

that's a good question, i currently have a very similar problem with any armada 1 and armada 2 (modded and unmodded) just popping up and closing right away with my current account. so i start armada through 'run as...' and it works without any problems..
i hope this is enough for you as temporary fix much of the loader code of v3 final was rewritten and runs much smoother than the v3pr version.

DOCa Cola