RE: Another Leader Bug savegame file (Full Version)

All Forums >> [Current Games From Matrix.] >> [World War II] >> War In The Pacific - Struggle Against Japan 1941 - 1945 >> Tech Support



Message

latosusi -> RE: Another Leader Bug savegame file (11/18/2004 10:43:36 PM)

My torpedo boat TF commander changed to jap one. I don't think i allocated a leader for it.



Tankerace -> RE: Another Leader Bug savegame file (11/18/2004 11:04:59 PM)

Pry, I just got you a repeatable save, scenario 15. I made 2 TFs at Pearl, and manually changed their TF commanders (12/7/41). When I ran the turn, the captain of the USS Shaw (docked at Pearl) changed to a wing commander (WCDR). Sending in a few.



Tankerace -> RE: Another Leader Bug savegame file (11/18/2004 11:18:40 PM)

Sent



Brausepaul -> RE: Another Leader Bug savegame file (11/19/2004 1:02:19 AM)

Just for the record:

my 1.3 only AI vs. AI game has proceeded until 5/17/42 and I found two Japanese commanders on Allied subs. That seems to mean that the AI is affected, too. I didn't check this match every day so I can't deliver a save game, I just wanted to find out if this bug is player only or not.

I know that new save games have been sent to Pry and I know that Matrix is doing anything to squeeze bugs, but this annoying bug really needs to be fixed. I waited for 1.3 to start a game, but I think I'll wait for another patch.



dtravel -> RE: Another Leader Bug savegame file (11/19/2004 4:39:53 AM)

<begins to very patiently wait for Pry to vivesect Tankerace's game> [:D]



pry -> RE: Another Leader Bug savegame file (11/19/2004 2:00:42 PM)

Justin, the save looks like a good test subject I will give this a hard look Saturday (my real job leaves me little time during the week to spend on this) and I will forward it up to Mike Wood so he can look at it.

I appreciate the help here guys and can use a few more saves as samples if you all can isolate any more, so that we can try to isloate all the conditions that might contribute to this.

LCU commander changing (non Russian units, there are other contributing causes that have to be corrected first)

Commander (Ship or TF that changes from default to another of the same side but different branch IE: Ship captain that changes from navy to airforce commander)

Ship or TF commander that changes from one side to the other side (Allied ship commander that changes to Japanese (any branch) commander)

You all get the idea and your help is greatly appreciated in nailing this little pain in the butt critter,

You can send additional saves to me at pry-witp@houston.rr.com



PeteG662 -> RE: Another Leader Bug savegame file (11/19/2004 4:27:53 PM)

When I see one again you will get it for sure.



dtravel -> RE: Another Leader Bug savegame file (11/19/2004 11:19:13 PM)

And based on other earlier statements by Pry, I believe that all saves sent to him need to be from games that were started after patching to 1.3.

Brausepaul, could you send your save to Pry also? It suddenly occurred to me to wonder if the leader replacement you saw in your AI vs. AI game occured during the game before you switched it over to look at it or when you switched it over. But the only way to find out is to use something other than the game executables to examine the save, which hopefully Pry or someone else at 2by3 has the tools to do.



Brausepaul -> RE: Another Leader Bug savegame file (11/19/2004 11:25:10 PM)

Ok, will do so :) I hope that pry can see this from the save



Ron Saueracker -> RE: Another Leader Bug savegame file (11/20/2004 2:55:21 AM)

Can't the auto assignment of leaders just be nixed as a feature?



pry -> RE: Another Leader Bug savegame file (11/20/2004 3:20:47 AM)

quote:

ORIGINAL: dtravel

And based on other earlier statements by Pry, I believe that all saves sent to him need to be from games that were started after patching to 1.3.


To clairify I need saves from games started in 1.3 anything from a game started eariler and patched up to 1.3 is no good because we simply can not be sure what kind of problems got carriered forward.



latosusi -> RE: Another Leader Bug savegame file (11/20/2004 4:51:29 PM)

Done!



michaelm75au -> Thought on leader bug issue (11/21/2004 6:54:43 AM)

Testers/programmer
There are about 6 empty slots in the Leader database. These slots show as rank WO with zero values including leadership and inspiration. At least one of these entries has a delay date of 9999, the rest a value of zero.

IIRC, one of the leader bugs is that units are controlled by WOs.

Now, the program knows how to skip these empty entries in the leader array (at least I am assuming it does[;)]).

The question is: When a leader is killed (eg goes down with his ship), he is removed from the leader array by ...blanking out the entires in the leader array as if he never existed?[&:] or is his attributes (leadership, aggressive,etc) still there and somehow he is being selected based on them even though he is DEAD.

This is one of those times when showing the leader unit's ID would be useful to see if the KILLED leader is coming back to life as a WO.

[These are just my thoughts and have nothing to do with the way the game actually is programmed.[:-]]

Michael



dtravel -> RE: Thought on leader bug issue (11/21/2004 9:20:22 AM)

Hey, at this point personally I'm considering this possibility (listen carefully, its a bit involved):

Grigsby, in an attempt to simulate the historical instances of leaders falling sick or dying in car crashes, has secretly added a routine to randomly change leaders at intervals. The other programmers, not aware of this, structured the databases with some empty slots for expansion later, filling them with 0/0 WO's and did other programming work related to the changing and loss of leaders which are incompatible with Grigsby's Gotcha. The interaction of these incompatible routines is progressively trashing the game, with the most obvious symptom being the Leader Leak bug.

Comments? [:D]



Tankerace -> RE: Thought on leader bug issue (11/21/2004 9:25:14 AM)

While I hope that isn't the case, I must say it does fit...

But, if that IS the case, then apparently all my good commanders are drunk drivers.

Personally, I think it is just one misspelled word in an IF THEN statement that is causing this. Something so trivial it was never even noticed.



dtravel -> RE: Thought on leader bug issue (11/21/2004 10:12:03 AM)

Misspelling a word in a program is actually pretty hard to do. The compiler tends to choke. But what does happen is you either A) forget a period where you need one, or B) put a period where you don't want one. Now, those typos are really hard to find.



Tankerace -> RE: Thought on leader bug issue (11/21/2004 10:39:33 AM)

Yeah, but something has to have done it. With as hard as the testers tested it, and they never saw it, yet on day 1 the community finds it, Then its declared fixed in 1.3 and again in day 1 we find it, I dunno....

Something is rotten in the State of Denmark.....



dtravel -> RE: Thought on leader bug issue (11/21/2004 10:12:30 PM)

Eennnhhh... <wags hand back and forth> I've done debugging and tech support. Some things you literally can't see when you are approaching it as a programmer or tester. You already "know" enough that you don't try some things that "obviously" no one will ever do because they shouldn't be done. But the first user who comes along who doesn't have the programmers' outlook does three keystrokes and *boing*, the whole thing crashes. You can also develop a kind of tunnel vision, where you are just looking at the code without looking at or considering what the user sees. (And it is hard to get around this. Testers tend to get hired because of their compatibility with the programmers, having similar outlooks. If their outlook is different enough to get around this, they don't get hired/signed up because they don't work well with the programmers. This isn't limited to just Matrix/2by3.)

The impression that I am getting is that specific things were tested in WiTP as individual features. But no one took a few steps back and just played the game as an entire package. Hmmm, not sure that gets across what I'm trying to say but I can't think of a better way of saying it.

But, yea, as much as I hate the implied and unintended slight, it does seem like QA was lacking. (Sorry guys. [:(])



fbastos -> RE: Another Leader Bug savegame file (11/21/2004 10:29:49 PM)

Got a leader bug as early as on the 4th turn of scenario 16 (that is, Dec 11, 1941). Can't repeat it, though. Keep trying.

F.



Tankerace -> RE: Thought on leader bug issue (11/21/2004 10:45:10 PM)

I know what you are saying.

Though I hate it, I am technically a COBOL programmer (not bad for a 19 year old, eh?). I remember I made a program, and it worked fine except for one thing screwed it up. I read the code over and over for about a week before I realized I had 1 period in the wrong place. Course when I fixed that it screwed up a lot of other stuff.

I'm kinda betting what happened is like what you said. They made all these features, and either 1) tested them, but not sit back and see what they do, or 2) looked at the code, thought it was flawless, and expected it to work.

OTOH, look at this. Pry says he has been trying his darnedest to recreate the bug, and he can't. Yet in 4 turns we can count about 10 instances of it. It makes me wonder if the testers were told not to do something (Such as don't manually assign TF commanders), yet that information was not passed on to us, the players.

All I can say is I am hoping....no,no, I am praying this gets fixed. With Kid already hinting that 1.4 may be the last shot, if this doesn't gets fixed then we are doomed.



dtravel -> RE: Thought on leader bug issue (11/21/2004 11:03:52 PM)

My last job was COBOL. Usually you don't find the little typos with big consequences until you're running the program thru an animator and see the execution suddenly jump to a section that has nothing to do with what is supposed to be happening or not jump past a chunk of code.

I think the biggest problem with finding the source of the Leader Leak bug is that it is random. We may be finding saves where the probabilities of it repeating are high, but I don't think it is possible to get a true "repeatable" save. If its something corrupting what's in memory or a broken pointer or link, then the only way to find a cause is to run the code thru the animator over and over and over and over and over and .... Even then, that only finds one cause. I don't think there is only one source. Either multiple things are breaking or the same piece of buggy code was copied to multiple places. Just a hunch though.



Tankerace -> RE: Thought on leader bug issue (11/21/2004 11:14:40 PM)

Well, if a lot of IF THENs were used (and if I understood Mike Wood correctly a lot were) then it is quite possible.

I just hope it gets fixed, because its hard to take a game seriously that so much $$$ was paid for, yet WOs command flattops.



michaelm75au -> RE: Thought on leader bug issue (11/22/2004 12:34:12 AM)

I finally got a Japanese leader assigned to my PT boat.
However the leader name is NOT one that is in the Leader database.

I am at work so don't have details. But the surname exists, but not the initial for that surname.

I will re-examine when I get home tonight.

Michael



Tankerace -> RE: Thought on leader bug issue (11/22/2004 12:44:38 AM)

It creates leaders randomly. I think Frag said this was done to make the game different each time. Who knows, maybe this is what is causing it?



pasternakski -> RE: Thought on leader bug issue (11/22/2004 12:46:32 AM)

The fact that a Japanese infantry major whose name doesn't appear in the database can pop up in command of a PT boat with the subscript "best suited to command an air combat task force" ought to tell a troubleshooter something.

Particularly when you click on the commander's name and get a list of three other Japanese infantry majors (all pre-qualified as air combat TF commanders) who are ready and willing to be plugged into RN MTB command.

I think this is so far broken it will never be fixed. I believe this to be one of many reasons everybody just decided to fold their hands behind their backs, whistle a happy tune, and walk away from the UV/WitP system.



Tankerace -> RE: Thought on leader bug issue (11/22/2004 1:15:22 AM)

Though I too am afraid of that, I hope to God that isn't the case, not after 80 bucks.



dtravel -> RE: Thought on leader bug issue (11/22/2004 5:51:43 AM)

While I've been one of the people predicting that this will not be a quick fix, lets not get too "doom and gloom" people. Give Pry and Mike some time to work on this.



Tankerace -> RE: Thought on leader bug issue (11/22/2004 6:00:29 AM)

Mike is an awsome programmer, and if I learned one thing from talking to him on the phone, its that he knows his business. We just have to give this time, and the bug will be killed.



pasternakski -> RE: Thought on leader bug issue (11/22/2004 9:58:22 AM)

quote:

ORIGINAL: Tankerace

We just have to give this time, and the bug will be killed.


That's what the first Sky Marshal in the Starship Troopers movie thought. Next thing you know, it was 100,000 killed (humans) in one hour. Then, the Joycelyn Elders clone Sky Marshal announced, "To defeat the bug, you must understand the bug."

Let's just remember - none of the "substitute" commanders actually exists in the database. They are all coming from the "random generation" routines and are being unintentionally superimposed at some point in the turn sequence. Moreover, their "specialty" is completely unrelated to the command they undertake (you would not expect a Japanese WO to be especially suited to command of an aircraft carrier, for example). This did not happen in UV. Something was added in WitP. It's not a bug. It's wacky design.

Isn't this enough of an indication of where the problem lies in order to track it down and correct it? What is the need for "repeatable saves?"

Get 'er done.



Tankerace -> RE: Thought on leader bug issue (11/22/2004 10:10:31 AM)

I do have to agree with that. It was something that was added to WitP that caused it. The question is, what was it?



Page: <<   < prev  1 [2] 3 4 5   next >   >>

Valid CSS!




Forum Software © ASPPlayground.NET Advanced Edition 2.4.5 ANSI
0.6875