The "Quit/Start/Load/Reload" Issue (FOTRSU)

[Front Runner]

For those interested, and, for those unfamiliar with this "issue," the "issue" is a failure to (serially) "re-load" a save game with Waypoints, Celestial Sphere, Navigation marks etc, accurate, correct and intact. I call it the "Quit/Start/Load/Re-load" test. The save games either "Pass" the test by Quitting the game, Starting the game, loading correctly and then re-loading correctly (Q/S/L/R) , or, "Fail" the test by Quitting the game, Starting the game, loading correctly once, but then failing to re-load correctly a second time (Q/S/L but NOT /R.) For example, waypoints missing, navigation marks missing, daylight when the Base time clock accurately indicates night time hours, etc.

My latest testing results reveal that it appears to only affect some versions of FOTRSUv* and only Campaign Careers starting Jan. 1, 1944 or later out of Pearl Harbor.

It is not a failure of the save game data. It is a failure to re-load the exact same data (a serial re-load) that was just previously correctly loaded.

I discovered this issue because I have a penchant for listening to Fred's Radio Stations while I am playing the game. While loading up a save game to start playing my session, the Radio came up already playing upon loading the game as it sometimes does. In times past, I could simply "re-load" that very same save game or a save game made just before (or after) with no problems. The "Radio already playing" bug is similar to the "Passing thermal layer" bug and sometimes requires a re-load of the save game file. When I re-loaded the save game, the Radio came up correctly, meaning that it was in the "Off" state, but, I had lost all of my navigation waypoints, and although I had saved at Sunset, the skies outside were now bright daylight with the Sun high in the sky, even though my Base time clock indicated it was "time" for sunset.

To discover this issue I had to be loading up a (FOTRSU) savegame from a Campaign Career, starting out of Pearl Harbor on Jan. 1, 1944 or later, and found it necessary to serially re-load an already successful (except for the Radio bug) save game. It helped that the savegame was made at Sunset, so after I had serially re-loaded the savegame, I went to the external view, and low and behold my beautiful sunset scenery was missing and it was broad daylight.

Many, many, many tests later and I believe that I have narrowed it down to "something" in the FOTRSU Campaign layer which affects only those Campaigns starting out of Pearl Harbor after Jan. 1, 1944. propbeanie is hard at work helping me find the culprit!

Anyone else who has experienced similar behavior and/or is simply interested in helping us track down this "bug" is more than welcome to chime in.

Thanks.

[Rockin Robbins]

That's a weird one. No mod can mess with the game save routines as they are hardcoded in the .exe file. So when we find out what's working it's very doubtful whether we can do anything about it. At least the savegame routines aren't as bad as SHIII. But they can be wonky, for sure.

[gutted]

Originally Posted by Rockin Robbins

. At least the savegame routines aren't as bad as SHIII.

Aye, i could never get a modded SH3 to load a save game without crashing. I always had to play missions in one go.

[BigWalleye]

Originally Posted by Rockin Robbins

That's a weird one. No mod can mess with the game save routines as they are hardcoded in the .exe file. So when we find out what's working it's very doubtful whether we can do anything about it. At least the savegame routines aren't as bad as SHIII. But they can be wonky, for sure.

There is a well-known aphorism in software development circles: "The changes I made had absolutely nothing to do withe THAT function and couldn't have caused it to stop working."

[Front Runner]

Originally Posted by Rockin Robbins

That's a weird one. No mod can mess with the game save routines as they are hardcoded in the .exe file. So when we find out what's working it's very doubtful whether we can do anything about it. At least the savegame routines aren't as bad as SHIII. But they can be wonky, for sure.

It appears to NOT be a "Save" (write) issue as the "Saves" all work on the first loading. So, when the first load is made (data is read), the saved game state is fine. It is when that game is "re-loaded" (second, or "serial" data read) without first quitting and starting the game, whatever happens, the game state loses waypoints, navigation marks, and celestial sphere sync. Also it appears that the "sky" defaults to the same state it was in at 1300 hours on the first day of the Campaign after leaving the Captain's Office. The "1300 Daylight state" is consistent no matter what "time" the save game was made, Sunset, Midnight, Sunrise, Submerged.) When that happens, a "Quit/Start/Load" brings the game state back to where it was when you made the save game and you can continue playing "in sync" from there. Further, it only affects Jan 1, 1944 and later Campaign starts from Pearl Harbor.

[EDIT] If hardcoded, it would affect SH4 Stock, TMO, RFB, TMO/RSRD, TMO/RSRD/OTC/ISP, etc. It doesn't.
Also, it doesn't affect FOTRSUv63. However, it does affect FOTRSUv71 and v80.

Propbeanie mentioned the simulated "International Date Line" as a possible culprit. The thing is that Midway Campaigns work OK and Midway is East of the IDL. So, IDK.

I'm wondering if there is a file in the "Documents/SH4/data/cfg" folder preventing the second load from reading into RAM correctly, in other words, portions of the RAM that have the correct information are being overwritten by Campaign start "default" information. No waypoints, no nav marks and 1300 Daylight sky. [EDIT] Perhaps "save.tmp" ?

[merc4ulfate]

Originally Posted by Front Runner

For those interested, and, for those unfamiliar with this "issue," the "issue" is a failure to (serially) "re-load" a save game with Waypoints, Celestial Sphere, Navigation marks etc, accurate, correct and intact. I call it the "Quit/Start/Load/Re-load" test. The save games either "Pass" the test by Quitting the game, Starting the game, loading correctly and then re-loading correctly (Q/S/L/R) , or, "Fail" the test by Quitting the game, Starting the game, loading correctly once, but then failing to re-load correctly a second time (Q/S/L but NOT /R.) For example, waypoints missing, navigation marks missing, daylight when the Base time clock accurately indicates night time hours, etc.

My latest testing results reveal that it appears to only affect some versions of FOTRSUv* and only Campaign Careers starting Jan. 1, 1944 or later out of Pearl Harbor.

It is not a failure of the save game data. It is a failure to re-load the exact same data (a serial re-load) that was just previously correctly loaded.

I discovered this issue because I have a penchant for listening to Fred's Radio Stations while I am playing the game. While loading up a save game to start playing my session, the Radio came up already playing upon loading the game as it sometimes does. In times past, I could simply "re-load" that very same save game or a save game made just before (or after) with no problems. The "Radio already playing" bug is similar to the "Passing thermal layer" bug and sometimes requires a re-load of the save game file. When I re-loaded the save game, the Radio came up correctly, meaning that it was in the "Off" state, but, I had lost all of my navigation waypoints, and although I had saved at Sunset, the skies outside were now bright daylight with the Sun high in the sky, even though my Base time clock indicated it was "time" for sunset.

To discover this issue I had to be loading up a (FOTRSU) savegame from a Campaign Career, starting out of Pearl Harbor on Jan. 1, 1944 or later, and found it necessary to serially re-load an already successful (except for the Radio bug) save game. It helped that the savegame was made at Sunset, so after I had serially re-loaded the savegame, I went to the external view, and low and behold my beautiful sunset scenery was missing and it was broad daylight.

Many, many, many tests later and I believe that I have narrowed it down to "something" in the FOTRSU Campaign layer which affects only those Campaigns starting out of Pearl Harbor after Jan. 1, 1944. propbeanie is hard at work helping me find the culprit!

Anyone else who has experienced similar behavior and/or is simply interested in helping us track down this "bug" is more than welcome to chime in.

Thanks.

I stopped doing things like this a long time ago. I never create a save game. That is what docking is for.

[Front Runner]

Simply put, a career started out of Pearl Harbor in January of 1944 using SH4 WOTP (Stock) and FOTRSUv63 "DO NOT" exhibit the issue. Both career starts begin at 1400 hours outside the harbor.

A career started out of Pearl Harbor in January of 1944 using FOTRSUv71 and FOTRSUv80 "DO" exhibit the issue. Both career starts begin at 1300 hours outside the harbor.

The genesis of the problem seems to coincide with moving the Pearl Harbor career start time from 1400 to 1300.

[KaleunMarco]

Originally Posted by Front Runner

Simply put, a career started out of Pearl Harbor in January of 1944 using SH4 WOTP (Stock) and FOTRSUv63 "DO NOT" exhibit the issue. Both career starts begin at 1400 hours outside the harbor.

A career started out of Pearl Harbor in January of 1944 using FOTRSUv71 and FOTRSUv80 "DO" exhibit the issue. Both career starts begin at 1300 hours outside the harbor.

The genesis of the problem seems to coincide with moving the Pearl Harbor career start time from 1400 to 1300.

that doesnt make sense.
im not calling you out but that doesnt make sense.

[Front Runner]

Originally Posted by KaleunMarco

that doesnt make sense.
im not calling you out but that doesnt make sense.

Think of the failure of the second re-load as almost a CTD.

[propbeanie]

It's an "internal data corruption", and I'd say it's similar to trying to switch from running The Museum and then going out on Patrol, or vice versa, where you try to run The Museum after having been on patrol. It usually does not "function as intended", apparently due to certain memory locations used by the game not getting "zeroed-out", or re-set if you will.

I have tried all sorts of different edits with several files in the mod that are different between v0.63 and v0.71. All that I can say conclusively is "don't load a Save after having loaded a Save." If you have to load a Save again, exit the game and try again. I have done several Starts in three versions (63, 71 & 80), and in each of them, upon loading a new career (emptied Save folder, new skipper name) I have gone to the NavMap on my boat, layed in a course to the southwest to the 20th parallel, and directly West from there, and then making a total of five waypoiints in the different versions. I then travel until sunset, stop the boat, take a picture through the Observation Periscope, Save, then exit. Start the game again, Load the Save and check the state of the game. In all three versions, everything is fine. Re-load the same Save, and in v0.63, everything is fine. In v0.71 & v0.80, it is now daylight-looking, still at about 1700 though, the sun is now at 295° instead of 335°, and elevated a good bit, just like it was at 1300 hours. The game has gotten confused.

Of note, the Save in v0.71 & v0.80 is 368kb, while the Save in v0.63 is 276kb. Same sub, same name, same location, same number of waypoints on a very similar route, nearly the same traffic around Pearl upon departure, but nothing showing on sonar, radar or visual several hundred miles south and west of Pearl. Differences between the mods are Scout Planes on ships, and torpedo-shooting AI subs and ships... That's about it. For the time being, that is where I'm laying the "blame", is at the feet of the mod's data load size... I'm still "testing" though.

[Front Runner]

Originally Posted by propbeanie

It's an "internal data corruption", and I'd say it's similar to trying to switch from running The Museum and then going out on Patrol, or vice versa, where you try to run The Museum after having been on patrol. It usually does not "function as intended", apparently due to certain memory locations used by the game not getting "zeroed-out", or re-set if you will.

......That's about it. For the time being, that is where I'm laying the "blame", is at the feet of the mod's data load size... I'm still "testing" though.

Roger that!
That re-set to "1300" daylight scene is either a "clue" or a "red herring."

[Jeff-Groves]

@ propbeanie,
Can you do a blank campaign test?
That would remove all the extra Units and such.

[Front Runner]

Originally Posted by Jeff-Groves

@ propbeanie,
Can you do a blank campaign test?
That would remove all the extra Units and such.

Explain to me how to set that up and I'll do it. PM me.

[Jeff-Groves]

I'm not sure how it's done for SH4 that's why I asked propbeanie.
I use to have a blank campaign for SH4 but that was a long time ago.

[propbeanie]

Yes. I've been doing some editing and deleting for such a move, with the plan to bring one thing at a time back in. I'll text you two a link when I get it set up.